Insight Editions · No. 266
Measure What Matters
How Google, Bono, and the Gates Foundation Rock the World with OKRs
John Doerr
John Doerr's case for Objectives and Key Results, the goal system Andy Grove built at Intel, read as a manual for focus, alignment, tracking and stretch, with the stories of Google, Intel and the Gates Foundation.
A study guide to Measure What Matters by John Doerr. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
The Book on One Page
Venture capitalist John Doerr argues that ideas are easy; execution is everything, and that the best tool for execution is Objectives and Key Results (OKRs), the goal system Andy Grove built at Intel and Doerr carried to Google in 1999. An objective says what you will achieve; three to five measurable key results say how you will know. Used with transparency, a quarterly rhythm and honest grading, a handful of OKRs gives any organization four superpowers: focus, alignment, tracking and stretch.
- Origins. Grove replaced Drucker's annual, private, pay-linked management by objectives with goals that are quarterly, public, partly bottom-up, separated from compensation, and judged by verifiable output, not activity. Doerr learned it as a 1975 Intel intern; Operation Crush (1980) used it to rally a near-billion-dollar company against Motorola in weeks, chasing 2,000 design wins.
- The evidence and the warning. Locke's 1968 finding that hard, specific goals beat easy or vague ones has survived 1,000+ studies; the "Goals Gone Wild" critique (Pinto, Sears, Enron) is answered by pairing quantity key results with quality ones.
- Superpower 1: focus and commit. Choose three to five objectives, each with five or fewer key results that are specific, time-bound and measurable; leaders must commit publicly or adoption fails (Remind's one metric, Nuna's false start and relaunch).
- Superpower 2: align and connect. Public OKRs, visible up to the CEO, link work to the mission, expose hidden dependencies and let peers connect sideways; rote cascading costs agility, so about half of OKRs should come from the bottom up (MyFitnessPal, Intuit).
- Superpower 3: track. OKRs live on a shared platform, get weekly check-ins, can be continued, updated, started or stopped mid-cycle, and end with a 0.0-1.0 score, self-assessment and reflection (Gates Foundation: a mission is not an objective).
- Superpower 4: stretch. Separate committed goals (expected 1.0) from aspirational ones (average 0.7); 10x targets force reinvention, as with Gmail's gigabyte, Chrome's two missed user targets before 111 million, and YouTube's billion daily hours, which first required switching from views to watch time.
- CFRs. OKRs need a human counterpart: Conversations, Feedback and Recognition replace the annual review (Adobe's Check-in saved 80,000 manager hours), and pay is divorced from OKR scores to prevent sandbagging; at Zume, OKRs and non-work one-on-ones trained managers early.
- Culture. OKRs and culture reinforce each other (Aristotle's five questions, catalysts and nourishers, Seidman's trust index); where trust and accountability are missing, fix culture first (Lumeris), or let a founder drive a pivot through OKRs (Bono's ONE, toward Africa-led work).
The book closes with a tribute to Bill Campbell, the coach who guarded Google's committed goals, and four resources: Google's OKR playbook with its six traps, a typical OKR cycle, a bank of performance-conversation questions and the In Sum checklist.
Doerr is OKRs' leading evangelist and an investor in many of the featured companies (Google, Zume, BetterWorks among them). The cases are success stories told by insiders, so the book shows how OKRs work when they work, not how often they do.
Doerr names one factor above every rule and tool: conviction and buy-in from the organization's leaders. Where executives treat OKRs as HR paperwork (Nuna's first try, Lumeris's "OKRs on paper"), the system fails.
Chapter map
| Ch | Title | One-line payload |
|---|---|---|
| 1 | Google, Meet OKRs | Doerr brings Grove's system to Google in 1999; goals beat no goals |
| 2 | The Father of OKRs | Grove turns Drucker's MBO into output-based OKRs; seven hygiene rules |
| 3 | Operation Crush | One objective, nine KRs: Intel routs Motorola in 1980 |
| 4 | Superpower #1: Focus | Three to five OKRs; pair key results (Pinto, Wells Fargo) |
| 5 | Focus: Remind | One big metric (active teachers) lets a start-up say no |
| 6 | Commit: Nuna | OKRs fail without public leadership commitment |
| 7 | Superpower #2: Align | Public goals; cascading's four costs; half bottom-up |
| 8 | Align: MyFitnessPal | Hidden dependencies, not bad goals, wreck execution |
| 9 | Connect: Intuit | Transparent OKRs link peers, not just bosses |
| 10 | Superpower #3: Track | Check in weekly; score 0-1; self-assess; reflect |
| 11 | Track: Gates Foundation | Tracking moves money; a mission is not an objective |
| 12 | Superpower #4: Stretch | Committed vs aspirational; 0.7 is success; Gmail's 10x |
| 13 | Stretch: Chrome | Missed 20M and 50M, then hit 111M in 2010 |
| 14 | Stretch: YouTube | Views to watch time, then 1 billion hours a day |
| 15 | OKRs and CFRs | Replace annual reviews; divorce pay from OKRs |
| 16 | Adobe | Check-in replaces 80,000 manager hours of reviews |
| 17 | Zume Pizza | Seven internal benefits; OKRs train managers early |
| 18 | Culture | Culture is the medium; OKRs catalyze, CFRs nourish |
| 19 | Lumeris | Fix culture first, relaunch OKRs, sell your reds |
| 20 | Bono's ONE | OKRs structure passion and a pivot to Africa-led work |
| 21 | The Goals to Come | No dogma: adapt the tool and execute |
| D | Bill Campbell | The coach who made "better every day" real |
| R1 | Google's Playbook | Google's rules, six traps, litmus tests |
| R2-4 | Cycle, Conversations, In Sum | Quarterly timetable, question bank, checklist |
Google, Meet OKRs
In 1999 John Doerr brought young Google a goal-setting method he had learned from Andy Grove at Intel: Objectives and Key Results (OKRs). His mantra frames the whole book: ideas are easy, execution is everything. OKRs are a collaborative goal-setting protocol that makes an organization focus its effort on the same few important issues, measured by results that either happen or do not. The chapter tells how Google adopted the system, answers the case against goals, and lays out the book's plan.
Foreword: Larry Page
Larry Page, Alphabet CEO and Google cofounder, says he wishes he had owned the book nineteen years earlier, when Google was founded, or even when he was managing nobody but himself. He dislikes process, yet holds that good ideas plus great execution are what produce magic, and that is the slot OKRs fill.
- The 1999 lecture. Doerr turned up and explained how Google should be run on OKRs, drawing on Intel. The founders knew Intel was well run and the talk made intuitive sense, so they tried it; Page's dry verdict is that it worked out pretty well.
- Adapt it. OKRs are a simple process that moves very different organizations forward, and Google has changed how it uses them over the years. Page tells readers to treat the method as a blueprint and make it their own.
- Visibility and pushback. For leaders, OKRs show what the whole organization is doing and give a productive way to challenge priorities: why can't a YouTube video load almost instantly, and isn't that more important than the goal you planned for next quarter?
- Bill Campbell. Page joins the book's tribute to the late coach, a warm man who was nearly always right about people, told anyone when they were talking nonsense, and was still liked afterwards. Page misses his weekly haranguing.
- The payoff. He writes forewords rarely but wanted to credit the gift: OKRs helped Google to 10x growth, many times over, made the mission of organizing the world's information seem achievable, and kept him and the company on time and on track when it mattered most.
The bet and the founders
On a fall day in 1999 Doerr arrived at Google's new two-story, L-shaped building off the 101 freeway, leased two months earlier after the company outgrew rooms above an ice-cream parlor in downtown Palo Alto. Two months before that he had made the largest bet of his nineteen years as a venture capitalist at Kleiner Perkins: $11.8 million for 12 percent of a start-up founded by two Stanford grad-school dropouts, plus a board seat. Barely a year after incorporating, Google had already declared its mission: to organize the world's information and make it universally accessible and useful.
- Sergey Brin. Exuberant, mercurial and strongly opinionated, a Soviet-born immigrant who could leap intellectual gaps in one bound; a canny, creative negotiator and principled leader, so restless he might drop for push-ups mid-meeting. He handled the commerce of the technology.
- Larry Page. An engineer's engineer and son of a computer-science pioneer; soft-spoken, nonconformist, with a 10x cause: making the internet exponentially more relevant. He worked on the product and imagined the impossible while keeping his feet on the ground.
The pitch had been thin on business. The deck ran seventeen slides, only two with numbers, padded with three cartoons. Google had a small deal with the Washington Post but had not yet unlocked keyword-targeted ads, and it was the eighteenth search engine on the web, a late arrival that in tech is usually fatal. None of this stopped Page lecturing Doerr on how poor existing search was and how much bigger it would become. The founders had no business plan, but their PageRank algorithm was clearly better than the competition even in beta.
Doerr asked the founders how large Google might become, having privately estimated that if everything went right it could reach a $1 billion market cap. Page said $10 billion. Doerr checked he meant market cap; Page said no, he meant revenue. At a normal growth rate for a profitable tech firm, that implied a $100 billion market capitalization, the territory of Microsoft, IBM and Intel, rarer than a unicorn. Doerr found no bragging in it, only calm judgment, and did not argue.
Long before Gmail, Android or Chrome, Google was full of big ideas. What the founders lacked was management experience. To have real impact, or even to get off the ground, they would need to make tough choices, keep the team on track and, given their appetite for risk, pull the plug on losers quickly. Above all they needed timely, relevant data to track progress: to measure what mattered.
The gift: a tool for execution
Doerr's present was a tool he had first used in the 1970s as an engineer at Intel, where Grove, whom he calls the greatest manager of any era, ran the best-run company Doerr had seen. Since joining Kleiner Perkins he had preached Grove's method to fifty companies or more. He reveres entrepreneurs and innovation, but had watched too many start-ups stumble over growth, scale and getting the right things done.
In the early 1980s Doerr took a fourteen-month sabbatical from Kleiner to run the desktop division at Sun Microsystems, suddenly responsible for hundreds of people and frightened by it. Grove's system was his anchor in every meeting: it empowered his executive team and rallied the division. There were mistakes, but also wins, including a new RISC microprocessor architecture that secured Sun's lead in workstations. That experience was his personal evidence for what he was bringing to Google.
He defines OKRs as a collaborative goal-setting protocol for companies, teams and individuals, and is careful about limits: OKRs are not a silver bullet and cannot replace sound judgment, strong leadership or a creative culture. With those fundamentals in place, they can take an organization to the top.
The ping-pong table lecture
About thirty Googlers, essentially the whole company, including Page, Brin, Marissa Mayer, Susan Wojcicki and Salar Kamangar, stood around the ping-pong table that doubled as a boardroom table or lounged in beanbags. Doerr's first slide defined OKRs as a management methodology for making sure the company focuses efforts on the same important issues throughout the organization.
- Objective = WHAT is to be achieved, no more and no less. By definition significant, concrete, action-oriented and ideally inspirational; designed well, a vaccine against fuzzy thinking and fuzzy execution. An objective can be long-lived, rolled over for a year or more.
- Key results = HOW you get there. They benchmark and monitor progress and must be specific, time-bound, aggressive yet realistic, and above all measurable and verifiable. Mayer's rule: it is not a key result unless it has a number.
- Binary at period end. At the end of the period, usually a quarter, each KR is declared met or not met, with no gray area. KRs evolve as the work progresses.
- The completeness test. Once all key results are complete, the objective is necessarily achieved; if it is not, the OKR was badly designed.
Objective: build a planning model for Google.
- KR 1. Finish the presentation on time.
- KR 2. Create a sample set of quarterly Google OKRs.
- KR 3. Win management agreement for a three-month OKR trial.
He illustrated with two scenarios: a fictional football team whose general manager cascades a top objective down the franchise's org chart (chapter 7), and Operation Crush, Intel's campaign to win back the microprocessor market (chapter 3). His closing value proposition: OKRs surface primary goals, channel effort and coordination, and link diverse operations, giving the whole organization purpose and unity. He stopped at the ninety-minute mark, on time.
The case against goals, and the evidence for them
In 2009 a Harvard Business School paper, "Goals Gone Wild", opened with a catalogue of destructive goal pursuit and argued that the dark side of goal setting could swamp its benefits. The paper is still widely cited.
- The examples. Exploding Ford Pinto fuel tanks, wholesale gouging at Sears auto-repair centers, Enron's reckless sales targets, and the 1996 Mount Everest disaster that killed eight climbers.
- The metaphor. Goals as a prescription-strength medicine needing careful dosing and close supervision.
- The label. Goals may cause systematic problems through narrowed focus, unethical behavior, increased risk taking, decreased cooperation and decreased motivation; apply with care.
- Doerr's reply. The caveat has merit; any management system can be run well or badly, and the book aims to show the good way. But for anyone seeking high performance, goals are necessary, and even the paper's authors conceded they can inspire people and lift performance.
- 1968, the year Intel opened, University of Maryland psychologist Edwin Locke proposed that hard goals beat easy goals and that specific hard goals produce more output than vague ones, an idea Doerr thinks influenced Grove.
- 1,000+ studies in the following half century confirmed it as one of the most tested ideas in management theory; 90 percent of field experiments find that well-defined, challenging goals raise productivity.
- Under one third of U.S. workers are engaged, per Gallup, which speaks of a worldwide engagement crisis; more than half of the disengaged would leave for a raise of 20 percent or less, and in tech two in three employees think they could find a better job within two months.
- No. 2: Deloitte finds retention and engagement second only to building global leadership among leaders' worries; more engaged groups earn more profit and lose fewer people.
- Top factor: a two-year Deloitte study found nothing raises engagement more than clearly defined goals that are written down and shared freely, which create alignment, clarity and job satisfaction.
Goal setting has failure modes: conflicting priorities or unclear, meaningless or arbitrarily shifting goals leave people frustrated, cynical and demotivated. Doerr's answer is a proper goal management system, which he equates with OKRs. It ties goals to the team's broader mission, respects targets and deadlines while adapting to circumstances, promotes feedback, celebrates wins large and small and, most important, stretches people toward what seems out of reach.
Google, meet OKRs: a perfect fit
When Doerr opened the floor, the audience seemed intrigued. Brin said Google needed some organizing principle, had none, and this might as well be it. Doerr insists the match was not random: it was an impedance match, a seamless transcription into Google's genetic code.
- Data. OKRs were an elastic, data-driven apparatus for a freewheeling, data-worshipping company.
- Openness. They promised transparency to a team that defaulted to open: open source, open systems, open web.
- Daring. They rewarded "good fails" and boldness in two of the boldest thinkers of their time.
- Writing it down. The founders, with few preconceptions about business, knew that writing goals down makes them real, and liked putting what mattered most on one or two pages made public to everyone.
With Eric Schmidt, who became CEO two years later, the founders were tenacious, insistent, even confrontational about using OKRs. Schmidt told Steven Levy that Google aimed to be the systematic innovator of scale: new things, done in big, reproducible ways. The three supplied the decisive ingredient for OKR success, conviction and buy-in at the top.
Who uses OKRs, and why
Doerr calls himself long on OKRs. As Google and Intel alumni move on, hundreds of firms of every size are adopting structured goals; he calls them Swiss Army knives. Adoption is broadest in tech, where agility and teamwork are imperatives: AOL, Dropbox, LinkedIn, Oracle, Slack, Spotify, Twitter. Beyond Silicon Valley: Anheuser-Busch, BMW, Disney, Exxon, Samsung.
| Setting | What OKRs do there |
|---|---|
| Start-ups | Survival tool: everyone pulls one way; backers get a yardstick before cash runs out (built the product, proved the market with 25 customers, here is what they will pay) |
| Scaling firms | Shared execution language: what must get done fast and who owns it; alignment vertically and horizontally |
| Large enterprises | Neon road signs: break silos, connect distant contributors, enable frontline autonomy, keep winners stretching |
| Nonprofits, government | Real-time data: the $20 billion Gates Foundation fights malaria, polio, HIV; Sylvia Mathews Burwell took OKRs to the OMB and HHS, where they helped fight Ebola |
Twenty years of OKRs at Google
Doerr argues no organization, not even Intel, has scaled OKRs better than Google. The regimen is simple in concept but demands rigor, commitment, clear thinking and intentional communication; it builds a goal muscle, with some pain for real gain. Schmidt and Jonathan Rosenberg wrote in How Google Works that OKRs became the simple tool that institutionalized the founders' "think big" ethos.
- Founder scrutiny. In the early years Page spent two days a quarter reviewing the OKRs of every software engineer, finding coherence in many moving parts (Doerr sat in on some). Later he opened each quarter with a marathon debate on his leadership team's objectives.
- Billion-user products. OKRs were the scaffolding for seven products with a billion-plus users each: Search, Chrome, Android, Maps, YouTube, Google Play, Gmail.
- Code Yellow. In 2008 a company-wide OKR rallied everyone to fight latency, the lag in fetching data from the cloud.
- Bottom-up plus 20 percent time. Grassroots OKRs work alongside "20 percent time", which frees engineers for promising side projects.
- Flat structure. Many firms cap managers at seven direct reports; Google sometimes makes seven the minimum (Rosenberg, running product, had up to twenty). A flatter chart means less oversight, more frontline autonomy and more room for breakthroughs, and OKRs make that workable.
October 2018 The CEO leads the whole company in grading progress on top-level OKRs, for the seventy-fifth consecutive quarter.
November-December Each team and product area plans the coming year and distills it into OKRs.
January CEO Sundar Pichai presents the high-level strategy and the year's OKRs to the company; the executive team also grades the prior year's OKRs, dissecting failures openly.
Following months Thousands of Googlers draft, discuss, revise and grade team and individual OKRs, free to browse the intranet and see how every other team measures success.
Page's projection now looks conservative: Alphabet's market cap exceeds $700 billion, the world's second-most valuable company, and in 2017 Google was No. 1 on Fortune's Best Companies to Work For for the sixth year running. Doerr roots this in stable leadership, technical resources and a values-based culture of transparency, teamwork and innovation, but says OKRs were essential too; neither he nor the founders can picture Google without them. Schmidt credits them with changing the course of the company forever.
Google's success rests on PageRank, search advertising and talent; Doerr concedes OKRs are no silver bullet, but the stories that follow come from his own portfolio and admirers, and he is an investor in several. Treat the causal credit given to OKRs as testimony, not a controlled finding.
Why this book, and its plan
For decades Doerr was the Johnny Appleseed of OKRs, spreading Grove's method with twenty slides, but felt he was only skimming the surface. The book, with its companion site whatmatters.com, is his attempt to go deep. He has taken OKRs to the world's most ambitious nonprofit and to an Irish rock star, and watched individuals become more disciplined, clearer and more purposeful. If the book were an OKR, its aspirational objective would be to make your life more fulfilling.
Grove was ahead of his time: acute focus, open sharing, exacting measurement and licence to shoot for the moon. Where OKRs take root, merit beats seniority, managers become coaches, mentors and architects, and actions and data count for more than words.
| Part | Chapters | Content |
|---|---|---|
| Origins | 2-3 | Grove's Intel and Operation Crush |
| Superpower 1 | 4-6 | Focus and commit: choose what matters and what doesn't |
| Superpower 2 | 7-9 | Align and connect: transparent goals, cross-team links, bottom-up OKRs |
| Superpower 3 | 10-11 | Track: data, check-ins, grading, no-blame accountability |
| Superpower 4 | 12-14 | Stretch: beyond the possible, freedom to fail |
| CFRs | 15-16 | Conversation, Feedback, Recognition replace annual reviews |
| Improvement | 17 | Zume Pizza, OKRs in every function |
| Culture | 18-20 | How OKRs speed culture change |
The tour visits about a dozen organizations, from Bono's ONE Campaign in Africa to YouTube's pursuit of 10x growth.
The Father of OKRs
OKRs were invented at Intel by Andy Grove, who took Peter Drucker's management by objectives and rebuilt it around measurable output. Doerr learned the system as a 1975 summer intern and saw it run a company with no margin for error. Grove's core idea: what you know matters far less than what you actually deliver, and a goal is only useful if it comes with a key result you can verify without argument. The chapter closes with Grove's seven rules of OKR hygiene.
An intern meets Andy Grove
The story starts with a romance. In summer 1975, between terms at Harvard Business School, Doerr drove through Yosemite into Silicon Valley with no job and no home, hoping to win back an ex-girlfriend, Ann, who had left him and was working somewhere in the Valley. He could program: while earning a master's in electrical engineering at Rice University he had cofounded a company writing graphics software for Burroughs, one of the "Seven Dwarfs" fighting IBM.
- Every VC said no. He wanted a venture-capital internship; all the firms turned him down, and one suggested a chip company it had funded in Santa Clara, Intel.
- One cold call. He phoned the most senior Intel person he could reach, Bill Davidow, head of the microcomputer division, who invited him in once he heard Doerr could write benchmarks. Davidow passed him to marketing manager Jim Lally, who passed him down the line; by five o'clock he had an internship.
- Ann was there too. She worked down the corridor and was not pleased to see him, but by Labor Day they were back together.
- Who runs this place. Mid-orientation Davidow told him there was one man running operations: Andy Grove. Grove was executive vice president and would wait twelve more years to succeed Gordon Moore as CEO, but everyone knew he was in charge: communicator, operator and taskmaster-in-chief.
- Gordon Moore. Shy, revered deep thinker behind Moore's law, which underpins the exponential scaling of technology (Doerr's gloss: processing power doubles every two years).
- Robert Noyce. Co-inventor of the integrated circuit; the charismatic Mr. Outside, as comfortable at a congressional hearing as buying drinks at the Wagon Wheel for a hard-partying chip crowd.
- Andy Grove (András István Gróf). The least likely by pedigree: a Hungarian refugee who narrowly escaped the Nazis and reached the U.S. at twenty with no money, little English and severe hearing loss; compact, curly-haired, maniacally driven. In his eleven years as CEO Intel returned more than 40 percent a year to shareholders, matching the arc of Moore's law.
"Intel delivers": the iOPEC seminar
Intel was Grove's laboratory for management innovation, and he loved to teach. Days after joining, Doerr was invited to iOPEC (Intel Organization, Philosophy, and Economics), a strategy and operations seminar taught by "Dr. Andy Grove." In an hour Grove traced the company's history year by year and listed its aims: a profit margin twice the industry norm, market leadership in every product line it entered, and challenging jobs and growth opportunities for staff. Doerr had heard similar talk at business school.
What stuck was Grove's contrast with Fairchild, his previous employer, where he had met Noyce and Moore and pioneered silicon wafer research. Fairchild was the industry's gold standard but lacked achievement orientation: people were hired and promoted for expertise, and whether they turned it into results was shrugged off. Intel was the reverse. It almost did not matter what you knew; what counted was what you could accomplish with what you knew or could learn. Hence the slogan "Intel delivers." Doerr found this real-world preference for accomplishment over credentials thrilling, something Harvard would not teach.
In his closing minutes Grove outlined a goal system he had begun installing in 1971, when Intel was three years old.
- Objective (the direction). Intel wants to dominate the mid-range microcomputer component business.
- Key result (a milestone) for this quarter. Win ten new designs for the 8085.
- The test. A key result must be measurable, so at the end anyone can check, without argument, whether it was done: yes or no, no judgment involved. Whether Intel dominates the market is for debate over years; whether it won ten designs is known within a quarter.
He called it a "very, very simple system," knowing simplicity appealed to engineers. Doerr was mesmerized.
Our MBO ancestors
Grove's system was fresh against the stale orthodoxy of the day, but it had a precursor in Peter Drucker, the Vienna-born gadfly Doerr calls the first great modern management thinker.
- Taylor and Ford. Frederick Winslow Taylor and Henry Ford first measured output systematically and studied how to raise it. They held that the most efficient organization was authoritarian: Taylor defined scientific management as knowing exactly what you want men to do and seeing that they do it the best and cheapest way. Grove described the result as crisp and hierarchical, order-givers and order-takers.
- Drucker's humanism. Half a century later Drucker demolished that model with an ideal that was results-driven yet humanistic: a corporation as a community built on trust and respect for workers, not just a profit machine. Subordinates should be consulted on company goals; crisis management should give way to balanced long- and short-range planning, informed by data and regular conversation among colleagues.
- The founding truth. Drucker sought a principle that gives full scope to individual strength while creating common direction and teamwork. He saw that people who help choose a course of action are likelier to see it through. In 1954, in The Practice of Management, he named it "management by objectives and self-control", Grove's foundation and the seed of the OKR.
- MBOs spread. By the 1960s forward-looking firms used MBOs, most visibly Hewlett-Packard as part of the "H-P Way." A meta-analysis of seventy studies found productivity gains of 56 percent where commitment to MBOs was high versus 6 percent where it was low.
- MBOs decay. Goals were often centrally planned and trickled slowly down the hierarchy; elsewhere they went stale without updating, got trapped in silos, or shrank into KPIs, numbers without soul or context. Deadliest of all, they were tied to salaries and bonuses, so why risk anything? By the 1990s MBOs were out of fashion and Drucker himself called them just another tool, not a cure for management inefficiency.
Measuring output
Grove's quantum leap was to apply manufacturing production principles to the soft professions: administrative, professional and managerial work. He wanted an environment that values output and avoids what Drucker called the activity trap: emphasizing output raises productivity, while chasing activity can lower it. Output is easy to see on an assembly line and hard to see when people are paid to think, so Grove worked on two riddles: how to define and measure knowledge workers' output, and how to increase it.
- A scientific manager. He read everything in behavioral science and cognitive psychology. The new theories offered nicer ways to get people working than Ford's, but controlled university experiments showed no leadership style was reliably better; he concluded no optimal style existed.
- Aggressive introverts. He hired people in his own image who solved problems quickly, objectively, systematically and permanently, confronting the problem without attacking the person and setting politics aside for faster, sounder, more collective decisions.
- A new name for a new thing. Intel ran on systems everywhere. Grove called his goal system iMBOs (Intel Management by Objectives) in tribute to Drucker, but rarely mentioned objectives without key results, a term he seems to have coined. Doerr assembled the acronym OKR from Grove's vocabulary. In almost every respect, the new method reversed the old.
| Dimension | Classic MBOs | Intel OKRs |
|---|---|---|
| Content | "What" | "What" and "How" |
| Cadence | Annual | Quarterly or monthly |
| Visibility | Private and siloed | Public and transparent |
| Direction | Top-down | Bottom-up or sideways (~50%) |
| Pay link | Tied to compensation | Mostly divorced from compensation |
| Risk | Risk averse | Aggressive and aspirational |
Doerr's first OKR
By 1975 the system was in full swing: every knowledge worker wrote monthly individual OKRs, and days after iOPEC Doerr's supervisor told him to write his. He was writing benchmarks for the 8080, Intel's latest 8-bit microprocessor, which ruled its market; the job was to show it was faster and beat the rivals.
Objective: demonstrate the 8080's superior performance compared with the Motorola 6800.
Key results (as measured by):
- Benchmarks. Deliver five benchmarks.
- Demo. Develop a demo.
- Training. Develop sales training materials for the field force.
- Proof. Call on three customers to prove the material works.
Intel's lifeblood
Doerr typed the OKR on an IBM Selectric (the first commercial laser printer was a year away) and posted it on his carrel for passers-by. He had never worked anywhere that wrote goals down, let alone let you see everyone's, up to the CEO. It was a beacon of focus and also liberating: when colleagues asked mid-quarter for new data sheets, he could say no without fear, because his OKRs spelled out his priorities publicly.
Under Grove, OKRs were central at weekly one-on-ones, biweekly staff meetings, and monthly and quarterly divisional reviews. That rigor let Intel manage tens of thousands of people etching a million lines of silicon or copper to within a millionth of a meter. In chipmaking, without rigor yields collapse and chips fail; OKRs kept reminding teams what to do and told them exactly what they were achieving or missing.
Doerr also trained Intel's domestic sales team. Grove learned that the person who knew the 8080 best was a twenty-four-year-old intern and told him to come to Europe. With Grove and his wife Eva, Doerr went to Paris, London and Munich, trained the European sales force, called on three big prospects and won two, and dined at Michelin-starred restaurants where Grove knew the wine list. Back home Grove had Davidow write a letter promising Doerr a job the next year. Doerr nearly quit Harvard; instead he worked part time on Intel's Digital Equipment Corporation account, dragging DEC into the microprocessor era, finished his last semester, and then stayed at Intel four years.
Andy Grove, OKR incarnate
The mid-1970s were the birth of the personal computer. As a first-year product manager, Doerr took Grove to the first West Coast Computer Faire at the San Francisco Civic Auditorium, where a former Intel executive was demonstrating the Apple II. Doerr argued that Intel already had the operating system, the microchip, the compilers and a BASIC licence, so it should build a PC. Walking past vendors selling plastic bags of parts, Grove judged them hobbyists and refused. Intel never entered the PC market.
- Compassion. Not demonstrative, Grove still tried to find a failing manager another role, perhaps lower down, where the person could succeed and regain respect. A problem solver who seemed to know exactly what he wanted and how to get it: a walking OKR.
- Andy's Late List. In the Free Speech and Haight-Ashbury era, punctuality was out of fashion. Grove put a sign-in sheet at the front desk logging anyone arriving after 8:05 and collected it at 9:00 sharp (Doerr, when late, waited in the car park until 9:05). Nobody was docked, but the list signaled that self-discipline mattered in a business with no margin for error.
- Creative confrontation. Hard on everyone, himself most; proudly self-made and at times arrogant; intolerant of fools, rambling meetings and half-baked proposals. He kept rubber stamps on his desk, one reading BULLSHIT, and believed problems are solved by facing people bluntly, directly and unapologetically.
- Leave your stripes outside. For all his temper he was approachable and open to good ideas: managers left rank at the door, and every big decision began with an egalitarian free-discussion stage. The way to earn his respect was to disagree, hold your ground and be proven right.
Sales, management and leaving
After eighteen months as a product manager, Lally, by then head of systems marketing, told Doerr that to become a good general manager he had to sell, get rejected and meet a quota: expertise aside, you succeed or fail on whether your team makes its numbers.
- Chicago, 1978. After marrying Ann, Doerr became a technical sales rep in the Midwest, his favourite job ever, helping customers build better dialysis machines and traffic-light controllers. His father Lou Doerr, an engineer who loved selling, was his model. First-year quota: $1 million, beaten.
- Marketing manager. Back in Santa Clara he had to hire a team, guide its work and measure it. With an Intel manager coaching him, he learned OKRs at a second, deeper level: more discipline, more constancy, clearer communication. None of it came naturally.
- Kleiner Perkins, 1980. Grove, by then president and devoted to Intel above all but perhaps his grandchildren, could not see why Doerr would leave. He offered him the software division, a business that did not yet exist, and a real P&L, and added that venture capital was not a real job, more like being a real estate agent.
Grove's legacy
Grove died at seventy-nine after years of stoic suffering from Parkinson's; The New York Times ranked him among the most acclaimed figures of the computer and internet era. He was not an immortal theorist like Moore, an icon like Noyce, or prolific enough to sit beside Drucker, yet he changed how we live. In 1997 Time named him Man of the Year as the person most responsible for the growth in microchips' power. Doerr calls him a rare hybrid: a supreme technologist and the greatest chief executive of his day.
A healthy OKR culture, Doerr says, came from Grove's character: ruthless intellectual honesty, disregard for self-interest and deep loyalty to the team. But it was his engineer's nuts-and-bolts approach that made the system work, and OKRs are his most lasting management practice.
Lessons from Grove and from Jim Lally, his OKR disciple and Doerr's mentor:
- Less is more. A few well-chosen objectives make clear what you say yes and no to. Limit each cycle to three to five OKRs, each objective with five or fewer key results (chapter 4).
- Set goals from the bottom up. Teams and individuals should write roughly half their own OKRs, in consultation with managers; all-top-down goals corrode motivation (chapter 7).
- No dictating. OKRs are a cooperative social contract setting priorities and how progress is measured. Even after company objectives are closed to debate, their key results stay negotiable; collective agreement maximizes achievement (chapter 7).
- Stay flexible. If the climate changes and an objective no longer looks practical or relevant, key results can be modified or discarded mid-cycle (chapter 10).
- Dare to fail. Output rises when everyone aims beyond their immediate grasp. Some operational objectives must be met in full, but aspirational OKRs should be uncomfortable and possibly unattainable: Grove's "stretched goals" (chapter 12).
- A tool, not a weapon. OKRs put a stopwatch in a person's own hand to pace themselves; they are not a legal document for a performance review. To encourage risk and prevent sandbagging, keep OKRs and bonuses separate (chapter 15).
- Be patient; be resolute. Intel stumbled many times before it understood the system's purpose. Expect four or five quarterly cycles to embrace OKRs, and longer to build mature goal muscle.
Moore's law, as Moore stated it (1965, revised 1975), concerns the number of transistors on a chip doubling roughly every two years, not processing power directly. Doerr's picture of Grove is affectionate memoir; the PC misjudgment he recounts is a reminder that OKRs execute strategy but do not choose it.
Operation Crush: An Intel Story
In late 1979 Intel's 8086 microprocessor was losing to Motorola's 68000, and Intel's top management reset the company's priorities in four weeks. Told largely by Bill Davidow, who led the campaign, Operation Crush shows all four OKR superpowers at once: focus, alignment, tracking and stretch. A cascaded set of OKRs let a nearly billion-dollar company turn on a dime: one objective, nine key results, thousands of people pointed the same way.
The threat
Doerr frames Crush as Intel's fight for survival and the book's first extended OKR story, showing how the system moves many departments and thousands of people toward one goal.
- The lineage. Intel engineer Ted Hoff invented the microprocessor, the "computer on a chip," in 1971. In 1975 Bill Gates and Paul Allen programmed the third-generation 8080 and launched the PC revolution. In 1978 Intel shipped the 8086, the first high-performance 16-bit microprocessor.
- The squeeze. The 8086 was soon being beaten badly by two chips that were faster and easier to program: Motorola's 68000 and upstart Zilog's Z8000.
- The alarm. In late November 1979 district sales manager Don Buckout sent a desperate eight-page telex; his boss Casey Powell forwarded it to Andy Grove, then president and COO. Within a week the executive staff met; a week later a blue-ribbon task force convened. Zilog was judged no real threat; Motorola, an industry giant with an international brand, was.
- The battle cry. Jim Lally set the tone in profane terms: there was only one competitor, Motorola, and Intel had to crush it so thoroughly it would not come back. The name stuck.
Nov 1979 Buckout's telex reaches Grove via Powell.
Dec 4, 1979 The task force meets for three days running; Grove "volunteers" Davidow to lead.
Dec 7, 1979 By Friday a plan to mobilize the company exists.
Dec 11, 1979 A nine-part program is approved, including Intel's first multimillion-dollar ad spend; within another week the sales force has it, all before Christmas.
Jan 1980 Noyce and Grove kick off Crush at the San Jose Hyatt House; Crush teams armed with Grove videos fan out to field offices worldwide.
Q2 1980 The new strategy is fully deployed by sales.
Q3 1980 On track for 2,000 design wins, deals for customers to build the 8086 into their products.
End 1980 Motorola routed.
1986 The 8086 family holds 85 percent of the 16-bit market.
Changing the terms, not the product
Not one Intel product was modified for Crush. Grove's team changed the terms of engagement instead: marketing rebuilt around Intel's strengths, customers steered toward long-term systems and services value over short-term ease of use, and selling aimed at CEOs instead of programmers.
Grove "volunteered" Bill Davidow, head of the microcomputer systems division (later a venture investor and author), to run it. Doerr singles out one of his contributions: Davidow grafted the phrase "as measured by" (a.m.b.) into Intel's company OKRs, as in "we will achieve this objective as measured by these key results." It made the link explicit for everyone. At a 2013 Computer History Museum panel, Crush veterans recalled OKRs being used down into the trenches; Doerr calls the Crush OKRs classics, time-bound and unambiguous, with every what and how in place.
Lally admits he was an OKR skeptic until Grove explained them. Send people toward the centre of Europe and some march to France, some to Germany, some to Italy; vectors pointing different ways add up to zero. Point everyone at Switzerland and the result is maximized. Grove then told him he had to teach it.
Davidow: the key result system as teaching
Davidow describes the system as Grove's way to mold behavior. Grove was single-mindedly committed to making Intel great, discouraged people from sitting on outside boards (Intel was meant to be your life), and used OKRs to consolidate that commitment.
- A philosophy, not only a mechanism. Senior managers teach, and OKRs were Intel's seminal teaching system; everyone learned that what gets measured gets better.
- The cascade in practice. Top goals were written with Grove around the executive staff table. As a division manager, Davidow adopted any relevant company key results as his objectives and spent the next week with his team deciding what to do that quarter. When Grove said what the corporation would do, everybody went all out; they were on a winning team and wanted to stay there.
- Rose bush and lawn. At lower levels OKRs might cover nearly 100 percent of someone's output, but managers also had day-to-day duties. If the objective is a beautiful rose bush, it goes without saying the lawn stays green. Davidow never had a key result about walking around to check morale: you write down what needs special emphasis.
Intel's urgency
In December 1979 Davidow went into Grove's staff meeting complaining that the microprocessor group should be winning more 8086 designs and needed goading. Grove told him to solve the problem himself.
- Pull-through ×10. The 8086 earned little alone, but each design win opened doors for EPROM (invented at Intel in 1971), peripheral and controller chips, worth up to ten times the original sale.
- Davidow's own stake. His division sold design aids (software development systems) for Intel-based products; growing fast, it depended on customers choosing Intel chips. Lose the 8086, lose the systems business.
- A company under siege. Intel had just lost the lead in DRAM to a start-up, and Japanese rivals were attacking its lucrative EPROM market. Microprocessors were its best hope.
- The stated purpose. An early slide: create a sense of urgency and set in motion company-wide decisions and plans to meet a life-threatening competitive challenge.
The task force met for three long days from Tuesday December 4, treating it as a huge puzzle. With no time to redesign the chip, it worked out what Intel actually had to sell. Davidow's answer was a new narrative: the microprocessor a customer picked now would be its most important decision of the next decade. Motorola could claim a cleaner instruction set, but not Intel's broad product family, system-level performance, superb technical support or low cost of ownership; Intel peripherals got products to market faster and cheaper, and Intel design aids made engineers more efficient. Motorola made everything from two-way radios to pocket televisions; Intel stuck to memory, microprocessors and supporting systems. Which would a customer rather call when something broke, or trust to stay ahead?
Lally wrote the ideas on the whiteboard, including publish a future products catalog and build a sales pitch for fifty seminars, with attendees getting the catalog. Speed was the difference: hit between the eyes by Powell's message, Intel responded within two weeks; Motorola, very well run but with less urgency, could not move as fast. One Motorola manager said he could not get a Chicago-to-Arizona plane ticket approved in the time Intel took to launch.
Intel excelled at turning broad generalizations into coordinated programs, and each of the nine projects became a company key result.
Objective: establish the 8086 as the highest-performance 16-bit microprocessor family, as measured by:
- Benchmarks. Develop and publish five benchmarks showing superior 8086 family performance (Applications).
- Repackaging. Repackage the entire 8086 family of products (Marketing).
- Speed. Get the 8MHz part into production (Engineering, Manufacturing).
- Coprocessor. Sample the arithmetic coprocessor no later than June 15 (Engineering).
Objective: deliver 500 8MHz 8086 parts to CGW by May 30.
- Artwork. Develop final art to photo plot by April 5.
- Masks. Deliver Rev 2.3 masks to fab on April 9.
- Tests. Test tapes completed by May 15.
- Fab start. Fab red-tag start no later than May 1.
The engineering objective is one corporate key result (the 8MHz part) made concrete with dated milestones: the cascade in miniature.
Turning on a dime
Just after New Year Bob Noyce and Grove held the kickoff at the San Jose Hyatt House for close to 100 managers: Intel would win in 16-bit microprocessors, was committed, and should treat this as the priority until done. The message reached two management levels at once and a third within 24 hours. Intel was nearly a billion-dollar company and turned on a dime; Davidow has never seen anything like it.
- Enthusiasm needs a system. Davidow has watched countless people leave meetings vowing to conquer the world, with nothing done three months later: whipped up, but not knowing what to do. Without the key result system Grove could not have launched all of Crush at once. It gave management a tool for rapid implementation and black-and-white criteria for judging reports.
- Top-down with input from below. Crush was thoroughly cascaded and driven from the top, but neither Grove nor Davidow could know every mechanic of winning. You can tell people to clean up a mess without telling them which broom to use: while the top said crush Motorola, someone at the bottom could decide the benchmarks were lousy and write better ones.
The greater good
Intel stayed on a war footing for six months. Davidow held a staff role with no line authority, yet got whatever he needed because everyone knew it mattered to Grove; division key results came back with virtually no dissent, resources moved on the fly, and he does not think he even had a budget.
- Scope. Top management, the entire sales force, four marketing departments and three locations pulled as one.
- Apolitical. Managers gave up fiefdoms. If the microprocessor division's futures catalog revealed a missing peripheral, that rippled into the peripherals division's engineering plans; sales ran the seminars but leaned on applications engineers, marketing and Davidow's division; corporate communications gathered trade-press articles from across the company.
- Culture counts. Grove wanted problems brought to management. A field engineer could tell a general manager that the "turkeys" at headquarters did not understand the market, and within two weeks the company was realigned top to bottom. It mattered that Buckout and Powell could speak without retribution; without that there is no Crush.
Bad companies are destroyed by crisis. Good companies survive them. Great companies are improved by them.
The outcome
By 1986, when Intel left its founding memory-chip business to go all-in on microprocessors, the 8086 had recaptured 85 percent of the 16-bit market. A cheaper variant, the 8088, went inside the first IBM PC, which standardized the personal computer platform. Today tens of billions of microcontrollers, in computers, cars, smart thermostats and blood-bank centrifuges, run on Intel architecture. Doerr's conclusion: none of it would have happened without OKRs.
The 8088's selection by IBM in 1980 owed much to price, availability and Intel's support ecosystem; OKRs organized the campaign, but the counterfactual is unprovable. "Tens of billions of microcontrollers on Intel architecture" also stretches the term: much of today's embedded market runs on ARM and other designs.
Superpower #1: Focus and Commit to Priorities
The first OKR superpower is focus: deciding what matters most for the next three, six or twelve months, deferring the rest, and committing to those choices in word and deed. Leaders who stand behind a few top-line OKRs give their teams a compass and a baseline; wrong decisions can be corrected once results arrive, but nondecisions, or hastily abandoned ones, teach nothing. If you try to focus on everything, you focus on nothing: three to five well-chosen OKRs, each with three to five key results, is the discipline.
The question that starts it
Measuring what matters begins with one question: what is most important for the coming period, and where should people put their effort? Doerr's answer is that good goal setting starts with disciplined thinking at the top, by leaders willing to spend time and energy choosing what counts. Paring the list is always hard but worth it: no person or company can do it all, and a select set of OKRs highlights the few things that must get done as planned and on time.
In the beginning: where top-line OKRs come from
For organization-level OKRs, the buck stops with senior leadership, who must personally commit. Google started from its mission statement, organizing the world's information and making it universally accessible and useful. Android, Google Earth, Chrome, the improved YouTube search engine and dozens more share that lineage: in each case the push came from the founders and executives, who made their focus plain through OKRs.
But good ideas are not bound by hierarchy, and the most energizing OKRs often start at the front line.
Rick Klau, a YouTube product manager, ran the site's homepage, the third most visited in the world. Only a small fraction of users logged in, so hundreds of millions missed features such as saving videos and channel subscriptions, and YouTube lost valuable data. His team proposed a six-month OKR to improve the login experience and pitched it to YouTube CEO Salar Kamangar, who consulted Larry Page. Page raised it to a Google company-wide OKR, with one change: three months, not six.
Klau says that once your OKR is top-line, all eyes in the company are on your team; nobody knew how to do it in three months, but owning a company-level goal showed their work came first. Page's emphasis also told other teams what mattered, and, as in Operation Crush, they rallied to help. The team finished on time, though it shipped one week late.
Leaders need goals of their own
However top-line goals are chosen, leaders must also set personal OKRs, publicly. Values cannot be passed on by memo, and structured goal setting will not take root by decree; Nuna's Jini Kim learned this the hard way (chapter 6). When a CEO says "all my goals are team goals", Doerr sees a red flag: talking a good OKR game is not enough.
You've got to say "This is what we're doing," and then you have to model it.
Campbell's point continues: if the leader does not model it, nobody else will do it.
Communicate with clarity
Top-line goals only help decisions, morale and performance if they are understood throughout the organization, and most firms fail at this.
- Two in three companies admit they do not communicate their goals consistently.
- 11,000 senior executives and managers surveyed: a majority could not name their company's top priorities.
- Only half could name even one.
Leaders must explain the why as well as the what. People want more than milestones; they want meaning, to see how their goals relate to the mission. And unveiling OKRs at a quarterly all-hands is not enough. LinkedIn CEO Jeff Weiner holds that by the time you are tired of saying it, people are starting to hear it.
Key results: care and feeding
Objectives and key results are the yin and yang of goal setting: principle and practice, vision and execution. Objectives are about inspiration and far horizons; key results are earthbound and metric-driven, usually hard numbers for gauges such as revenue, growth, active users, quality, safety, market share and customer engagement. Drucker's point applies: a manager must be able to measure performance and results against the goal.
- Levers. Key results are the levers you pull and the marks you hit to reach the objective.
- Three to five. For a well-framed objective, three to five KRs usually suffice; more dilute focus and hide progress.
- Each one a challenge. If you are sure you will hit a KR, you are probably not pushing hard enough.
What, how, when: choosing a cadence
OKRs shock the established order, so easing in can make sense. Some firms start on an annual cycle while moving from private to public goals or from top-down to collaborative ones. The best practice may be a dual cadence: short-horizon OKRs for the here and now supporting annual OKRs and longer strategy. But it is the shorter-term goals that drive the actual work and keep annual plans honest and executed.
- Deadlines focus. Clear time frames intensify focus and commitment; nothing moves people like a deadline, and markets demand nimbleness.
- Quarterly as default. Doerr finds a quarterly cadence best for fast-changing markets: three months curbs procrastination and yields real gains.
- Grove's rule. In High Output Management Grove says feedback works only if it arrives soon after the activity it measures, so with yearly plans, OKRs should run at least quarterly, perhaps monthly.
- No religion. An engineering team might use six-week cycles to match development sprints; an early-stage company still seeking product-market fit might use monthly cycles. The right cadence fits the business's context and culture.
Pairing key results
One-dimensional OKRs are dangerous, and the more ambitious the OKR, the greater the risk of overlooking an essential criterion.
In 1971, losing share to fuel-efficient Japanese and German cars, Ford launched the budget Pinto subcompact. To meet CEO Lee Iacocca's aggressive demands, product managers skipped safety checks; the gas tank sat six inches in front of a flimsy rear bumper. Engineers knew it was a firetrap, but the heavily marketed target, under 2,000 pounds and under $2,000, was enforced with an iron hand: when a crash test showed a one-pound, one-dollar plastic part stopped the tank being punctured, it was dropped as extra cost and weight.
The car's internal green book set three product objectives: True Subcompact (size, weight); Low Cost of Ownership (price, fuel use, reliability, serviceability); Clear Product Superiority (appearance, comfort, features, ride and handling, performance). Safety appeared nowhere. Hundreds died in rear-end collisions and thousands were badly injured. In 1978 Ford recalled 1.5 million Pintos and sister Mercury Bobcats, then the largest recall in automotive history. The diagnosis: specific, challenging goals (speed to market, fuel efficiency, cost) were met at the expense of unspecified ones (safety, ethics, reputation).
A more recent banking scandal grew from ruthless, one-dimensional sales targets. Branch managers felt pushed to open millions of fraudulent accounts customers neither wanted nor needed; in one case a manager's teenage daughter had 24 accounts and her husband 21. More than 5,000 bankers were fired, the credit card and checking businesses fell by half or more, and Doerr thinks the brand may be damaged beyond repair.
Grove's fix in High Output Management: pair each output key result with a counterpart that stresses the quality of the work. In accounts payable, vouchers processed pair with errors found by auditors or suppliers; a custodial group's square feet cleaned pair with a quality rating from a senior manager whose office is in that building.
| Quantity goal | Quality goal | Result |
|---|---|---|
| Three new features | Fewer than five bugs per feature in QA testing | Developers write cleaner code |
| $50M in Q1 sales | $10M in Q1 maintenance contracts | Sustained sales attention raises customer success and satisfaction |
| Ten sales calls | Two new orders | Lead quality improves to meet the order threshold |
The perfect and the good
Sundar Pichai told Doerr his team often agonizes over goal setting: a single OKR line can take an hour and a half of thought to make sure it does something better for the user. That comes with the territory, but Doerr paraphrases Voltaire: do not let the perfect be the enemy of the good. An OKR can be modified or scrapped at any point in its cycle, and the right key results sometimes surface weeks or months in. OKRs are works in progress, not commandments in stone.
- Succinct, specific, measurable. Every key result.
- Mix outputs and inputs. Results achieved plus the activities that produce them.
- Completeness. Finishing all key results must attain the objective; if not, it is not an OKR.
The quality continuum below shows the rules in action: weak KRs lack numbers, average ones add numbers, strong ones add measurable inputs (wind tunnel tests, daily practice) and a quality pair (pit stop errors).
| Grade | Key results under the objective "Win the Indy 500" |
|---|---|
| Weak | Increase lap speed. Reduce pit stop time. |
| Average | Increase average lap speed by 2 percent. Reduce average pit stop time by one second. |
| Strong | Increase average lap speed by 2 percent. Test at wind tunnel ten times. Reduce average pit stop time by one second. Reduce pit stop errors by 50 percent. Practice pit stops one hour per day. |
Less is more
Steve Jobs said innovation means saying no to a thousand things. The ideal number of quarterly OKRs is usually three to five; letting more past the velvet rope blurs focus or sends people chasing the next shiny thing. At MyFitnessPal, CEO Mike Lee found his team was putting down too much, so priorities were unclear; they deliberately set fewer OKRs and made sure those were the ones that really mattered.
For individuals, as Doerr found at Intel, selective goal setting is the first defense against overload. Once people have agreed their quarter's OKRs with their managers, any add-on goal must fit the existing agenda.
In a working OKR system, top-down orders to "just do more" are obsolete; orders give way to questions, above all what matters most?
The one thing an OKR system must deliver above all is focus, which requires few objectives. Every commitment forfeits another, as with any finite resource, so planners need the guts, honesty and discipline to drop projects as well as start them, to say no as well as yes.
Top-line objectives must also be significant: not a catch-all wish list, nor the sum of a team's mundane tasks, but stringently curated goals that deserve special attention, move people forward now and link to the larger purpose. Grove defined the art of management as picking, from many activities of apparently similar weight, the one, two or three that give far more leverage than the rest, and concentrating on them. Larry Page's version is to put more wood behind fewer arrows, which Doerr calls the essence of the first superpower.
Both the Pinto and Wells Fargo involved leaders who ignored known risks; pairing key results cannot fix a culture that punishes bad news or rewards fraud. Some analysts (notably Gary Schwartz's 1991 study) also argue the Pinto's overall fire death rate was not far out of line with comparable cars, though the design and the recall are not disputed.
Focus: The Remind Story
Brett Kopf, cofounder of Remind, a safe texting platform linking teachers, students and parents, learned focus first as a struggling student and then as a founder scaling too fast. OKRs let Remind say no to good ideas that would not move its one big metric, active teachers, because you can only do one big thing at a time really well. The story also shows the classic start-up lesson that preceded OKRs: talk to your users before you build.
The problem Remind attacks
A Brown University study suggested a fix for American schooling: better teacher-family communication. When summer-school teachers made daily calls and sent texts or notes home, their sixth-graders completed 42 percent more homework and class participation rose by nearly half.
For decades firms tried to lift achievement by putting technology into schools, and it did not work. Meanwhile tens of millions of children walked into class with a smartphone, and texting became teenagers' main way of communicating. Remind's opening was to make texting a secure, practical channel for principals, teachers, students and parents.
Doerr's introduction stresses that focus is how you separate OKR wheat from chaff. He remembers Kopf's passion for customers and a start-up exquisitely focused on teachers: in the bathroom of the tiny loft office, the company objectives were taped to the mirror above the toilet. In 2012 Brett and his brother David made Forbes' "30 Under 30" in Education. Scale demanded more focus, and OKRs formalized a process they had already begun.
Growing up with ADHD
Kopf grew up in Skokie, Illinois, unable to sit still; a forty-minute math lesson felt endless and he was the spitball kid. Tested in fifth grade, he was diagnosed with attention deficit hyperactivity disorder and dyslexia: words were hard, numbers harder. His entrepreneur parents were at work by 5 a.m.; he worked hard too, but his grades and confidence sank, and in high school on Chicago's North Side he believed the kids who called him stupid.
In his junior year teacher Denise Whitefield began working with him one-on-one. Each day she asked what he had to do; he listed a history worksheet, an English essay, a math test. Then she told him to pick one and talk it through, and he finished it. She reassured him she had all day. His panic eased and he began to believe he could cope. His mother spoke to her weekly and came to school at least monthly: "Team Brett" would not let him fail. The ACT (some six hundred questions over four hours without moving) was a horror for someone with ADHD, but he got into Michigan State, his first big win.
The lesson he drew: reformers start with curriculum or "accountability" (test scores), and what gets lost is human connection, which is what Remind is about.
Twitter for education
Remind began, like many ventures, with one person's problem. At college Kopf was hopeless with deadlines that professors changed on a whim; cut off from his support system, he failed three majors before settling on agricultural economics. He still had five syllabi a semester, each with up to thirty-five assignments, quizzes and tests, and college success is time management, a matter of dynamic goal setting.
- The trigger. In junior year he slaved over an essay, got a mediocre grade, and then had to dig for it on a clunky web system. Friends texted in real time on BlackBerrys; why not school data, and teachers reaching students on their phones anywhere?
- The 24-hour offer. He gave his older brother David, in web-services security at a big Chicago insurer, twenty-four hours to decide whether to start a company with him. David called back five minutes later: in.
- Two years of fumbling. Neither knew technology, product or operations (Brett's work experience was an internship at Kraft Foods, mostly stocking cookies). Students shared syllabi and Brett loaded them into David's Excel macros, which texted alerts such as a reminder about an 8 a.m. history quiz. Archaic and unscalable, but it worked for a few hundred users, including Brett, who graduated.
- The arrogant failure. In early 2011 he moved to Chicago full time with $30,000 from friends and family (pasta every night). They failed because they spent their time meeting investors and drawing elaborate website schematics, and none learning teachers' problems.
Down to a few hundred dollars, they survived by getting into Imagine K12, the Silicon Valley accelerator for education. Their mission line: Remind101, a safe way for teachers to message students and parents, using SMS as the hook, "Twitter for education." With Demo Day ninety days away, David quit his job and they moved to the Valley.
- Solve a problem.
- Build a simple product.
- Talk to your users.
While David locked himself away to teach himself to code, Brett set a single ten-week goal: interview 200 teachers across the U.S. and Canada. He contacted 500 on Twitter and held 250 one-on-ones, beating the target.
- The pain point. Communication outside school ranked high. At the final bell, teachers stuck "homework due tomorrow" sticky notes on students' shoulders.
- The constraints. Phone trees and permission slips were labor-intensive and unreliable; texting between a thirty-year-old teacher and a twelve-year-old carried liability. Teachers needed a secure platform with no personal data attached, accessible yet private, and less work, not more.
- The paper beta. By Day 15 Brett had a crude prototype on printer paper: hand-drawn phone and email icons, a line saying students could receive your messages, and three buttons, Invite, Print, Share. On Skype he held it to the camera and explained that teachers could text the class and students would never see their phone number or social profile. Teachers nearly fell off their chairs, every time. That was when the brothers knew they were on track.
Scaling on a shoestring
By Day 70 the software worked: teachers signed up on the web, formed a virtual class, and gave students and parents a dedicated number to text. Remind had the hockey-stick chart every start-up wants. On Demo Day Brett pitched for two minutes to a room of eleven other start-ups and a hundred investors, then mingled for two hours and handed out at least forty cards.
| Date | Milestone |
|---|---|
| Launch + 3 weeks | 130,000 messages sent |
| Early 2012 | $10,000 in debt; rescued by $30,000 seed from Ulu Ventures (Miriam Rivera, Clint Korver) |
| 2012 | More from Maneesh Arora, Google PM, later founder of MightyText, Brett's mentor |
| 2012-13 | Adding 80,000 users a day with five staff, two engineers, zero marketing |
| Fall 2013 | 6 million users; Series A from Chamath Palihapitiya and Social+Capital |
| Feb 2014 | Series B led by Kleiner Perkins; Doerr pitches OKRs |
| Aug 2014 | 300,000+ student and parent downloads a day; No. 3 in Apple's App Store |
| End fall 2014 | Past 1 billion messages |
| 3 quarters after OKRs | $40 million Series C |
Growth felt like the sorcerer's apprentice, fast and out of control. Brett asked teachers for feedback and each spread the word to fifty colleagues; because the service was free, no district approval was needed.
Measuring what mattered
Goals stayed strictly qualitative until fall 2013. Arora had nudged them to back decisions with data; Palihapitiya showed them how to tell the story on one page of data and to ignore inessentials such as registered users: nobody cared how many teachers signed up if they never came back.
By the time Doerr saw the list above the toilet it named three numbers: Weekly Active Teachers (WAT), Monthly Active Teachers (MAT) and retention, plus a few quarterly initiatives (migrate the databases, build the app, hire four people), posted so everyone could see what the company was doing.
Still in a one-bedroom loft, short of engineers, with a barely working mobile app, Remind was clearly teacher-obsessed and quantified. In February 2014, just before the Series B closed, Doerr pitched OKRs with examples from Intel, Google, LinkedIn and Twitter, a method to keep them focused and to guide, track and support each step. Brett's reaction: why not try it?
Goals for growth
The August 2014 back-to-school explosion forced every department to ramp up at once; none of the goal setting was glamorous, Kopf says, but all of it was necessary. Remind started OKRs at fourteen people and reached sixty within two years, too many to settle priorities around one table. OKRs focused people on what would move the company up a level; to hit the teacher-engagement objective and its time-bound key result, much else had to wait.
Objective: support company hiring.
- Finance. Hire 1 director of finance and operations (talk to at least 3 candidates).
- Marketing. Source 1 product marketing manager (meet 5 candidates this quarter).
- Product. Source 1 product manager (meet 5 candidates this quarter).
- Saying no to delight. A repeated message remains one of the most requested features: a teacher could remind a fifth-grade class every Monday to bring their novel without resending. It was a classic "delight" feature, but would it move engagement? The answer was no, so it was shelved, a hard call for a teacher-centric firm that Kopf doubts they would have held without OKR discipline.
- Not all top-down. Leaders voted on the quarter's top objectives and told contributors what mattered and why; contributors worked out how to get there. Everything was written down, so everyone knew what others were doing: no confusion, no Monday-morning quarterbacking, no politics.
- The CEO's three buckets. Kopf limited himself to three or four personal objectives, printed and kept on his notepad and by his computer. Each morning he asked what he was doing today to move them, a good question for any leader.
- Open about failure. He told staff the three things he was working on and which one he was failing at miserably. People need to see the CEO's priorities to align with them, and to see that it is fine to err, correct and move on; fear of screwing up kills innovation.
- Firing yourself. Leaders of fast-growing start-ups keep firing themselves from their early jobs. Kopf had been doing accounting and payroll; an early OKR offloaded finance so he could focus on product and strategy, and OKRs eased his shift to working through a layer of executives, stopping him from backsliding into micromanagement.
An OKR legacy
OKRs are simple but not mastered at once. Early company-level objectives were off by miles, mostly too ambitious: seven or eight objectives when there was capacity for two at best. Kopf, new to strategic planning, thinks they should have eased in rather than installing the whole system at once, but would do it again: OKRs made Remind a company that executes, and three quarters after implementation it raised the $40 million Series C.
Doerr closes on Remind's unchanged core constituency, teachers, and the founders' vision to give every student an opportunity to succeed. Kopf's contrast: you can summon a cab in five minutes, yet when a child falls behind it can take weeks or months for a parent to find out.
Remind's errors are Kopf's own account: too many objectives and too fast an install. The $40 million round followed OKRs, but Remind's explosive organic growth (6 million users, 80,000 sign-ups a day) began long before them, driven by the user research of 2011. OKRs disciplined a product that already worked.
Commit: The Nuna Story
Jini Kim's health-data company Nuna shows the second half of superpower one: commitment, both to the highest-impact goals and to the OKR process itself. Its first OKR rollout failed because executives never bought in; the second worked because the CEO chased, modeled and published her own graded goals. OKRs only take hold when leaders commit publicly and model the behavior; if the officers abandon ship mid-voyage, the sailors will not bring it into port.
Doerr's frame
Nuna is the story of a founder driven by family tragedy to improve care for huge numbers of Americans, who bootstrapped through years of rejection and then recruited engineers and data scientists to a wildly audacious goal: building a new Medicaid data platform from scratch. Nuna took off in 2014; four years and one enormous Medicaid contract later it was using data to make the system work for those who need it most, and applying the lessons to help large companies improve the efficiency and quality of care in their private plans. Kim first met OKRs as a Google product manager. Doerr flags two sides of commitment here: OKRs locked in commitment to the top goals, and leaders and contributors learned to commit to the process itself.
A personal origin
Kim's brother Kimong was diagnosed with severe autism at two; a few years later he had his first grand mal seizure at Disneyland. Her parents, Korean immigrants with limited means and little English, felt helpless, and without a safety net the family would have gone bankrupt. At age nine, Jini was the one who signed the family up for Medicaid.
- Google, 2004. Her first job out of college; she had never heard of OKRs, but they became her indispensable compass for getting the most important work done.
- Google Health. One of her first products taught her how much data matters for better care, and how hard it is to get health data, even your own. That led her to found Nuna in 2010.
- No money, no customers. No OKRs at first. Kim worked full time, five others part time (including grad-student cofounder David Chen), nobody paid. They built a prototype and pitched large self-insured employers: zero orders in year one, rightly, because they did not yet understand customers well enough.
- Crashing conferences. Still no orders in year two, Kim put on a suit and crashed HR conferences to learn what benefits directors cared about and what real innovation in health care looked like.
- Product-market fit, 2012. That knowledge signed several Fortune 500 clients, after more than two years of rejection and ramen. Back from a six-month stint with Healthcare.gov, she closed $30 million, enough to pay the team for years.
The Medicaid contract
A government request for proposals sought the first-ever database of all Medicaid members: 74.5 million lives in 50 states, five territories and D.C. Previous attempts had failed multiple times. After seventy-two hours on adrenaline and Red Bull, Nuna submitted its bid to the Centers for Medicare & Medicaid Services (CMS) just in time, and two months later learned it had won.
Scaling was a three-dimensional job: upgrading the business (compliance, security, privacy), the data platform infrastructure, and the workforce, from 15 to 75 people, while still running the employer business and finishing within a year. It demanded more focus and commitment than ever.
The false start, 2015
As an ex-Googler, Kim was sold on OKRs, but she underestimated what introducing and executing them took. Goal muscle has to be built gradually; as her own private wellness OKR to run a marathon taught her, too much too soon ends in pain.
- What they did. Created quarterly and annual OKRs and rolled them out to everyone at once, about twenty people.
- What happened. It did not take. Some never set individual OKRs; others set them and stuck them in a drawer.
- What she would do now. Start with the five-person leadership team. Executives must commit first, and it may take a quarter or two to overcome managers' resistance, until they see OKRs as a practical tool for top priorities, not a necessary evil or perfunctory exercise.
Until executives are fully on board, contributors will not follow, especially when OKRs are aspirational: the harder an objective, the greater the temptation to abandon it. People look to their bosses when setting and pursuing goals; officers who jump ship in a storm cannot expect sailors to bring it into port.
The second attempt, mid-2016
Kim relaunched with renewed commitment and saw the executives buy in, but did not relax: chasing was her job. She emailed contributors to write individual OKRs; non-responders got a Slack message, then a text, then she would grab them in person and plead with them to do their OKRs.
Leaders must practice what they teach. Kim shared her own OKRs at an all-hands and was surprised how much it rallied the company: it showed she was accountable too. Contributors now feel free to critique her OKRs and suggest improvements, which she says made all the difference. Her hiring OKR, deceptively simple and make-or-break, drew much constructive input. Grades are on Google's 0.0 to 1.0 scale.
- Engineers. Recruit 10 engineers [0.8].
- Sales. Hire a commercial sales leader [1.0].
- Candidate experience. 100 percent of candidates feel they had a well-organized, professional experience even if Nuna makes no offer [0.5].
Two key results measuring commitment to professional development:
- Reviews. 100 percent of Nunas have gone through a performance review and feedback cycle [1.0].
- Scoring. 100 percent of Nunas score their individual Q3 OKRs within the first week of Q4 [0.4].
The low 0.4 is itself telling: even after the relaunch, getting everyone to grade on time remained the hardest part of the process.
Private OKRs
Nuna's commitment to OKRs is public and visible, but sometimes a private mode works better. In Q4 2016 Kim wanted a VP for the employer business to speed that unit's growth. It was a new position and she was unsure how staff would take it, so she set a private OKR that only she and Chen were accountable for. It deepened her commitment: she spoke one-on-one with key stakeholders, found potential recruits, and launched a formal recruiting process.
What OKRs changed
Start-ups live with ambiguity, and as Nuna expanded from self-insured employers to the Medicaid database to a suite of health plan products, it relied on OKRs more than ever. Sharper focus and clearer priorities are the prerequisites for deeper commitment.
- Conversations. OKRs forced many conversations that would not otherwise have happened.
- Alignment. The company is getting more aligned.
- Acting, not reacting. Instead of reacting to external events on the fly, teams act purposefully on quarterly plans.
- Deadlines. Stricter, yet they feel more attainable; Nuna does what it says it will.
You're not going to get the system just right the first time around.
Chen's full moral: it will not be perfect the second or third time either, so don't be discouraged, persevere, and adapt it to make it your own. Kim adds that commitment feeds on itself.
The mission ahead
With CMS's support Nuna has built a secure, flexible platform holding private health data for more than 74 million Americans. It wants more: to inform policy makers governing a costly, complex system, to drive analytics that predict and prevent illness, and to help improve the nation's care. Her Google lesson: the hairier the mission, the more important your OKRs.
Kimong still speaks only three words: uhma, appa and nuna, Korean for mom, dad and big sister. He gave the company its name and its mission.
In January 2017 Nuna revealed its Medicaid work. Andrew M. Slavitt, acting director of CMS, told The New York Times the cloud database was "near historic", a leap from state-level computing silos to the first systemwide view across the program. Doerr adds that Kim and Chen are just getting started.
Superpower #2: Align and Connect for Teamwork
The second superpower is transparency: everyone, from the most junior hire up to the CEO, can see everyone's OKRs. Public goals align individual work with the mission, expose duplication, and surface the best ideas wherever they start. Healthy OKR systems balance top-down alignment with bottom-up autonomy, settling at roughly half and half, and connect teams sideways as well as vertically. Doerr's fantasy football team shows both what cascading achieves and why cascading everything fails.
Why make goals public
Social media made transparency the default in daily life, and Doerr calls it the express lane to operating excellence, yet at most companies goals remain secret. Aaron Levie, founder and CEO of the cloud firm Box, voices a common frustration: at any moment a significant share of people are working on the wrong things, and the hard part is knowing which.
- Public goals get met. Research finds public goals are more likely to be attained than private ones; switching to "open" lifts achievement across the board. In a survey of 1,000 working U.S. adults, 92 percent said they would be more motivated if colleagues could see their progress.
- Open to all levels. The most junior person can read everyone's goals up to the CEO; critiques and corrections happen in public, and anyone may comment, even on flaws in the process.
- Meritocracy in sunlight. When people write down what they are working on, it is easier to see where the best ideas come from, and clear that those who rise are doing what the company most values.
- Poisons lose power. Suspicion, sandbagging and politicking fade. If sales hates marketing's plan, the dispute is aired openly instead of simmering in a silo. OKRs make objectives objective.
- Help arrives. If Employee A is visibly struggling with a quarterly objective, colleagues can see it, comment and pitch in; the work improves and working relationships deepen.
- Duplication exposed. In big organizations several people often unknowingly work on the same thing; a line of sight to every objective saves time and money.
The same page: alignment
After top-line objectives are set, managers and contributors tie daily work to the organization's vision. This linkage is alignment, and its value is hard to overstate.
- 2× more likely: per Harvard Business Review, companies with highly aligned employees are more than twice as likely to be top performers.
- 7 percent: the share of employees who fully understand their company's strategy and what is expected of them.
- No. 1: misalignment is the top obstacle between strategy and execution, according to a poll of global CEOs.
Amelia Merrill, an HR leader at RMS, a California risk-modeling firm, describes the problem: many offices and time zones, some parallel and some joint work, and everything feeling important and urgent, so employees cannot see what to do first. Doerr's answer is focused, transparent OKRs that knit each person's work to team, department and mission. People crave connection and want to know what their leaders are doing and how their own work fits; OKRs are the vehicle for vertical alignment.
The grand cascade
In the old business world goals came down the org chart like tablets from Mount Sinai: executives set objectives for department heads, who passed them to the next tier, and so on. The method still prevails in big organizations because it corrals lower-level staff onto the company's main concerns and, at best, forges unity.
The Sand Hill Unicorns
Doerr's fantasy football team, used in his Google pitch, shows cascading: each manager's key results become the next level's objectives.
- General manager. Objective: make money for the owner. KRs: win the Super Bowl; fill home stands to 90 percent+. Achieve both and profit is certain, so the OKR is well built.
- Head coach. Objective: win the Super Bowl. KRs: a passing attack of 300+ yards a game; a defense allowing fewer than 17 points a game; special teams in the top 3 for punt return coverage.
- Offensive coordinator. Objective: a 300-yard passing attack. KRs: 65 percent completion rate; fewer than 1 interception a game; hire a new quarterbacks coach.
- Defensive coordinator. Objective: give up fewer than 17 points. KRs: under 100 rushing yards allowed a game; 3+ sacks a game; develop a Pro Bowl cornerback.
- Special teams coach. Objective: top-3 punt coverage. KRs: under 10 yards per punt return; block 4+ punts in the season.
- SVP of marketing. Objective: fill stands to 90 percent. KRs: upgrade team branding; improve media coverage; revitalize in-stadium promotion.
- Marketing director. Objective: upgrade branding. KRs: target two colorful players for a campaign; create a more compelling slogan.
- Publicist. Objective: improve media coverage. KRs: players attend two charity events a season; invite 20 sports reporters to a meet-and-greet; share event photos on social media.
- Merchandise manager. Objective: revitalize in-stadium promotion. KRs: contact 10 souvenir companies; price out 5 options; present 3 giveaway ideas by August 1.
The SVP's key results are a mess: not measurable, specific or time-bound. Does better coverage mean five ESPN features, one Sports Illustrated cover, or 50 percent more social followers? And even with stronger KRs the approach stays flawed: the top objective, making a wealthy person wealthier, offers no intrinsic motivation to the general manager, let alone the East Coast scout or the PR intern at the copier.
In moderation, cascading makes an operation coherent. When every objective cascades, it becomes a mechanical, color-by-numbers exercise.
- Loss of agility. Even medium-size firms have six or seven reporting levels. Waiting for the waterfall, with meetings and reviews sprouting like weeds, each cycle can take weeks or months; tightly cascading firms resist fast goal setting, and quarterly OKRs may become impractical.
- Lack of flexibility. Cascaded goals take so much effort that people won't revise them mid-cycle; even small updates burden everyone downstream, and the system grows onerous.
- Marginalized contributors. Rigid cascades shut out the front line, who hesitate to raise concerns or promising ideas.
- One-dimensional linkages. Cascades lock in vertical alignment but barely connect peers across departments.
Bottoms up
Because OKRs are transparent, they can be shared without lockstep cascading. Levels can be skipped when that serves the purpose: an objective might jump from the CEO straight to a manager, or from a director to an individual contributor; or leaders can publish their OKRs to everyone and trust people to adapt their own goals.
With tens of thousands of staff, Google's culture would be hamstrung by rote cascading. Laszlo Bock, former head of People Operations, writes in Work Rules! that goals improve performance but hours spent cascading them do not. Because top OKRs are known and everyone's are visible, goals converge over time; grossly misaligned teams stand out, and the few initiatives touching everyone are managed directly.
- 20 percent time. The antithesis of cascading: engineers spend the equivalent of one day a week on side projects. In 2001 Paul Buchheit started one code-named Caribou; it became Gmail, the leading web email service.
- The physical therapist. To avoid soul-killing overalignment, some goals should rise from the bottom. A Unicorns therapist learns a new injury-prevention regimen at a conference and, on her own, sets an off-season OKR to adopt it. It does not match her manager's OKRs but serves the GM's top objective: healthy stars make a Super Bowl win far likelier.
- Innovation lives at the edges. The most powerful OKRs usually come from outside the C-suite. Grove: people in the trenches sense change first; salespeople read shifting demand before management, financial analysts spot changing fundamentals first.
- Micromanagement is mismanagement. Drucker: the professional employee needs rigorous standards and high goals, but how the work is done should be his own responsibility and decision. Grove warned that managerial meddling makes subordinates narrow their view, show less initiative and push problems upward, so output falls.
Who owns the how
An optimal system lets contributors set at least some of their objectives and most or all of their key results. People who choose their destination stretch further, set more ambitious targets and hit more of them; the research rule is the higher the goals, the higher the performance, and they understand better what the trip will take.
If Doerr's doctor orders him to lower his blood pressure by training for the San Francisco Marathon, he might grudgingly consider it. If he decides of his own free will to run it, he is far likelier to finish, especially running with friends. A goal whose how is defined by others does not engage people to the same degree.
There is rarely a single right answer in business; loosening the reins and backing people to find theirs helps everyone win. High-functioning teams live with creative tension between top-down and bottom-up goals, a mix of aligned and unaligned OKRs.
- Be more directive in times of operational urgency, when simply doing takes precedence.
- Use a lighter touch when the numbers are strong but the company has grown too cautious and buttoned-up.
- The usual balance. Leaders attuned to the business and their people tend to land near half top-down, half bottom-up, which Doerr thinks about right.
Cross-functional coordination
Even when goal setting transcends the org chart, unacknowledged dependencies remain the number-one cause of project slippage. The cure is lateral connection, peer to peer and team to team. Isolated individuals cannot match a connected group at innovation or hard problems: product depends on engineering, marketing on sales, and as initiatives grow more complex, interdependent divisions need a tool to reach the finish together. Connected companies are quicker companies.
Bock describes the effect of transparency: people designing a handset reach out to a software team because they spotted an interesting user-interface idea. With goals public, a team of teams can attack problems wherever they appear. Anyone hitting it out of the park gets looked at; anyone always missing gets looked at too. Clear signals start virtuous cycles, and, Bock marvels, the management tax is zero.
Align: The MyFitnessPal Story
Brothers Mike and Albert Lee built MyFitnessPal into the leading health and fitness app and then folded it into Under Armour. Both times their goals were fine but their alignment failed: teams depended on each other without saying so. Unacknowledged dependencies, not bad objectives, wrecked their execution; OKRs fixed it by making every dependency explicit and giving every OKR a single owner. Doerr's framing: OKRs are not islands but networks, vertical, horizontal and diagonal.
The company
It started with a beach wedding. Before marrying, Mike and Amy Lee wanted to lose weight; a trainer gave them a list of nutritional values for 3,000 foods and a paper pad to track calories. Mike, a programmer since age ten, knew there was a better way and conceived MyFitnessPal. For eight years he and Albert self-funded it from savings and credit cards.
- 45 million registered users when Kleiner Perkins invested in 2013; 120 million+ at the time of writing.
- 300 million pounds lost by users collectively.
- 14 million foods in the database, plus real-time links to Fitbit and dozens of other apps.
- $475 million: Under Armour's purchase price, February 2015.
The brothers' mission is a healthy planet, at the centre of a movement for quantified, digital self-care. The app reveals what used to be hidden (the calories burned on a morning run), helps users set and reach ambitious personal goals, and adds a network of friends who cheer them on. The merger married MyFitnessPal's technology to a great brand and opened access to professional athletes, the next frontier; Lee wanted to skate to where the puck is going. It also brought a maze of internal relationships, which OKRs would help them work through.
Doerr's point about alignment: when employees align with top-line goals their impact is amplified, and they stop duplicating effort or working against the grain. Goal setting came naturally to the Lees, though not easily.
Moments of clarity
Lee argues the smartphone puts a coach, nutritionist or medical adviser in your pocket at little or no cost, so people can make healthier decisions. The app gives "moments of clarity" that last a lifetime. His own: tracking his food, he learned mayonnaise has ninety calories a tablespoon and mustard five, and has not touched mayo since. Enough small changes add up.
Early focus
None of Lee's previous employers used formal goal systems. They had annual financial plans, revenue targets and broad strategies, nothing structured or continuous, and, not coincidentally, a glaring lack of alignment: he had no idea what other teams were doing, and extra meetings only wasted time.
Put two people in a boat, one rowing east and one west, and they burn a lot of energy going nowhere.
- The thousand-item list. Early on they joked about crossing off the top three items of a thousand-item to-do list and calling it a good year, leaving much below the line on purpose.
- One goal at a time. Launch the Android app, or BlackBerry, or iPhone or iPad: one at a time until done, rarely overlapping. Focused and highly measurable, if unsophisticated.
- Alignment is easy at two. With one other person on product, the brothers declared a keystone goal (launch on iPad by a date) and talked daily. Lee now wishes they had used OKRs even before funding, to be readier when opportunity came.
Entropy
Once the app took off on iPhone and Android, growth went off the charts; one day they had 35 million registered users and could no longer do one thing at a time. Lee's insight: entropy begins when you have two great people directly under you. You give each something big; each naturally pushes their own piece; soon they pull out of alignment and work on two different things. Pushing harder does not help: a good hammer splays two slightly misaligned nails sideways.
They knew they needed more structure but not how. In 2013, soon after Kleiner invested, Doerr pitched OKRs. The football analogy clicked for Lee: a simple main objective, distilled, stretched and cascaded through the organization, was how they would align the company.
Cross-team integration
Implementation was harder than expected. Choosing the right company objectives and cascading them to drive behavior took much thought, and balancing strategic with granular, directive communication was difficult.
- OKRs for people, not people for OKRs. After the Series A and a bigger leadership team, they gave each leader one big dedicated goal for accountability. Some were too narrow, others too nebulous. If an HR manager could not connect to product or revenue goals, they added a top-line objective just for him. Soon there was a cornucopia of company OKRs and nobody knew what really mattered: they had lost the forest for the trees.
- The scaling slowdown. Growing from 10 to 30 people in 2013, Lee assumed they would be 200 percent more productive. New engineers needed long training, and several engineers on one project needed new processes so they did not override each other. Productivity fell during the transition.
Alignment, Lee says, is helping people understand what you want them to do. Most contributors will happily ladder up to top-line OKRs, if they know where to set the ladder.
The shared engineering problem
As the team grew more layered, three product managers each had sound individual OKRs: one for Premium, the paid subscription tier; one for the API platform that let partners like Fitbit connect, write data or build on MyFitnessPal; and one for the core login experience.
- Caught in the middle. The shared engineering team was not aligned with these objectives; it had its own infrastructure OKRs to keep the plumbing working and lights on. Management assumed it could do everything, a big mistake.
- Shifting priorities. Engineers did not know what to work on, and it changed without notice, sometimes to whichever product manager yelled loudest. Switching projects weekly killed efficiency; returning to a product meant relearning how it worked. Revenue-critical Premium went in fits and starts.
- Lee's frustration. Talented hires and heavy spending, yet no faster.
A top-priority marketing OKR for personalized, targeted emails was well built: drive a minimum number of monthly active users to the blog, with a key result to raise email click-through. But nobody in marketing told engineering, which had already set its quarter's priorities. Without engineers' buy-in the OKR was doomed before it started, and the founders did not learn this until the quarterly postmortem. It was done a quarter late.
That was the wake-up call. The OKRs were well crafted but implementation fell short: when departments relied on each other, the dependency was never made explicit, coordination was hit-and-miss and deadlines were routinely blown.
- Quarterly integration meetings. Each quarter department heads presented goals and identified dependencies. Nobody left until three questions were answered: is everyone's buy-in need met? Is any team overstretched? If so, how can its objectives become realistic?
- One owner per OKR. Alignment is not redundancy: every OKR has a single owner, with other teams linking in. Co-ownership weakens accountability; Lee does not want two people blaming each other. Parallel objectives must have distinct key results.
- Deadline KRs for the new. Market response to something conceptually new is unpredictable; they would wildly beat or miss a metric. So they keyed KRs to dates ("Launch MFP Premium by 5/1/15"), then, once real data arrived, set the next round against projected outputs, realistic or stretched.
- Watch for timid KRs. The more ambitious the objective, the more conservative the KRs people chose (an email here, a push notification there), a classic unintended consequence. Goals were fitted to context: incremental where right, and sometimes the instruction to ignore monthly active users and swing for the fences.
Each cycle got better: objectives sharper, KRs more measurable, achievement rates higher. It took two or three quarters to get the hang of it, especially for product features tied to broad objectives.
Under Armour: unacknowledged dependencies, writ larger
Under Armour had a different goal-setting mode. Lee now had a boss and a new division, UA Connected Fitness-North America, charged with using digital technology to improve fitness and performance, with three more apps, each with its own culture. At scale, alignment grows exponentially more complex: how to show 400 people what the group was trying to achieve so they could align with it and each other? Lee wonders how Amazon or Apple manage; introducing OKRs across the division made a big difference.
A few weeks after the acquisition, Lee's boss called a twenty-person leadership off-site of Connected Fitness stakeholders. Under Armour ran an annual cadence, so each department head presented the year's aims. The Lees were prepared, and surprised:
- E-commerce counted on the apps to drive significant traffic.
- The data team assumed a mass of data.
- Media sales had booked the group for a set dollar amount of new ad revenue.
Each had fixed expectations, with no visibility into the others' asks and no view of how they fit Connected Fitness's growth goals or the company picture. It was the old MyFitnessPal problem on steroids; there was no way to do it all.
Straightening out alignment took eighteen months, impossible without OKRs.
- Define capacity. First, the group's capacity constraints for new software.
- Clarify core priorities. Then what mattered most.
- One-way transparency. Lee shared Connected Fitness's high-level OKRs, explained why projects needed the time allotted and where to double down on company goals, and invited others to say if anything was missing or wrong. Nervous at first, he found it worked: others recognized the limits and adjusted, while his teams sought projects that met cross-departmental objectives.
- Cut the road map in half. When Albert took over the MapMyFitness team, he said half the road map had to go. Features are now judged the MyFitnessPal way: if this comes off the road map this quarter, what happens to the user experience? Usually little. The calls rest on metrics, not opinion.
- Choose. Focus and alignment are binary stars. Premium finally launched in May 2015, three months after the deal, only once the group openly admitted it could not do everything and declared Premium its number-one objective above all else.
- Still a work in progress. Soon after the merger, two of the four apps built maps into run tracking at the same time, in different ways with different providers: inefficient and inconsistent for customers. The teams set up a monthly check-in, and OKRs then went division-wide. Knowing the top priorities gives everyone freedom to say no.
North Star alignment
The group still prizes transparency and accountability: objectives sit on a wiki anyone can read and are discussed at weekly all-hands. When Lee demonstrated the OKR process at a leadership retreat, one executive called it the best off-site ever. With OKRs the operating foundation of Connected Fitness, he hopes to spread them across Under Armour by example: the larger the organization, the more value they offer.
Alignment's deeper meaning is keeping goals true to North Star values: Under Armour's mission "to make all athletes better", and the MyFitnessPal mantra that when customers reach their health goals, the company succeeds. The founding question still applies: will this feature or partnership help customers succeed? Users do the hard work, like the woman who stood up from her chair without using her hands for the first time in twenty years; the company's success lies in providing such moments, and it writes that into high-level objectives.
Objective: help more people around the world.
- Growth. Add 27M new users in 2014.
- Scale. Reach 80M total registered users.
Every decision must square with the vision. In a trade-off between customers and a business goal, they align with the customer; an objective out of line with the mantra gets extra scrutiny before it goes ahead.
The North Star story reads differently in hindsight. In 2018 MyFitnessPal disclosed a breach affecting about 150 million user accounts, and in 2020 Under Armour sold the app to Francisco Partners for about $345 million, well below the $475 million it paid, abandoning the Connected Fitness strategy. Alignment inside the division did not make the acquisition pay off for the parent.
Connect: The Intuit Story
Atticus Tysen, CIO of Intuit, piloted OKRs in a 600-person IT department during the company's shift to the cloud, to open platforms and to global operations. Public goals connected remote teams to headquarters and, more surprisingly, to each other. The payoff of transparent OKRs is horizontal: people stop linking every goal to the boss and start linking to the peers they actually depend on. The story also shows the value of an OKR pilot before, or even instead of, a company-wide rollout.
The company
Intuit has been on Fortune's World's Most Admired Companies list for fourteen years running. It made its name in the 1980s with Quicken, personal finance on the desktop, then added TurboTax and the accounting program QuickBooks, which later moved online. It has survived one competitive threat after another by staying a step ahead; most recently it sold Quicken and rebuilt QuickBooks Online as an open platform, and subscriptions rose 49 percent. UBS analyst Brent Thill told The New York Times that whenever Intuit takes a wrong turn it gets quickly off the gravel and back onto the blacktop.
People cannot connect with what they cannot see, and networks do not grow in silos. Because OKRs are open to every level of every department, companies that stick with them become more coherent; adaptable organizations tend to be openly connected ones.
Intuit's transparency was instilled by cofounder Scott Cook and reinforced by "Coach" Bill Campbell, its CEO and longtime chairman. Tysen remembers Campbell as one of the most open people he has met: he read people, invested in them, and you always knew what he thought and that he was in your corner. CEO Brad Smith posts his own goals in his office for anyone to see and calls connected goal setting essential to letting employees do the best work of their lives.
The pilot
To help IT adapt to the cloud, Tysen introduced OKRs in stages, refusing to force them: he wanted enthusiastic compliance, not bureaucratic compliance, and to see whether the system would succeed on its own. It did.
- Quarter 1: OKRs for Tysen's direct reports; quarter 2: down to directors; quarter 3: all 600 IT employees.
- ~2,500 active objectives tackled each quarter, supported by real-time automated data and routine check-ins.
- About half of OKRs aligned to higher-ups' or the department's goals.
- 4,000+ views of managers' OKRs a quarter, about seven per employee, a strong sign of frontline engagement.
- 3 to 5 business objectives per employee per quarter, plus 1 or 2 personal ones.
Doerr calls it the Lasik surgery of OKRs: contributors now see clear links between their daily work, colleagues' priorities, the team's quarterly objectives and the company's "True North" mission. A few hundred users can serve as an OKR laboratory to iron out kinks before deployment at scale.
Tysen's challenge: IT in a disruption
Tysen spent eleven years on the product side before becoming CIO in 2013, moving because he loved the company and saw IT had to evolve. Intuit was pivoting several ways at once:
- Desktop to cloud software.
- Closed to open platform, welcoming thousands of third-party apps.
- North American to global.
- House of brands to branded house: TurboTax, Quicken and QuickBooks becoming one integrated Intuit ecosystem.
In any disruption IT takes the brunt of internal frustration, partly because it is opaque. A company over thirty years old, especially a tech company, piles up layers of complex technology. IT juggles internal partners and end users, bridges technology and business outcomes, and, hardest of all, must keep today's systems working perfectly while investing in the future. Intuit once ran nine different billing systems, each with its own problems; when you fight fires daily, building next-generation billing is hard. In a siloed organization, people who try to see what is happening elsewhere do not know where to start or lack time to follow up.
Change from the top, experiment from below
Change began with Brad Smith, chairman and CEO, who installed a company-wide goal-setting system: managers meet reports monthly to discuss individual goals, with built-in 360-degree feedback as both sides compare notes. Intuit has a culture of learning and experimentation: try many things, keep what works, adapt it. Tysen partnered with HR to pilot OKRs in Enterprise Business Solutions (EBS), Intuit's name for IT. He had found OKRs in 2014 by Googling "goal setting", and his research suggested they could change how EBS operated and how it saw itself.
- Beyond tickets. Modern IT is not ticking off help tickets and change requests; it is adding business value: shedding redundant clone systems, creating functionality, finding future-oriented solutions. EBS needed to change root and branch.
- Air cover. Leaders had to give people permission to back-burner routine tasks for longer-term, higher-value work.
- Visible beyond IT. Though nobody outside EBS used them, Tysen made EBS OKRs visible across Intuit, so the company knew what IT was doing, how and why. People who understand your priorities and constraints are likelier to trust you when something goes sideways.
Tysen at first assumed his personal goals should equal the department's top OKRs. Bad optics: top objectives typically lasted about eighteen months, while teams and individuals updated theirs as conditions changed. Staff reasonably asked what the CIO was doing if his goals never changed. He took the point; he now has his own objectives and ladders up to the top-level OKRs like everyone else.
Bringing the edges in
EBS has formal teams in four U.S. regions and Bangalore, southern India's tech hub, plus support staff in every Intuit location. People far from the centre wonder what headquarters does, and headquarters wonders about them. OKRs ended the mystery and made the department cohesive.
One top-level objective is to rationalize, modernize and secure all technology used to run Intuit. Visiting teams in Texas or Arizona, Tysen now hears staff describe projects as rationalizing the portfolio or modernizing a system: the same three verbs everywhere. When a new project is proposed, they ask how it fits the OKR template, and if it does not, they rightly ask why are we doing this?
- ERP. Complete migration of Oracle eBusiness Suite to R12 and retire 11.5.9 this quarter.
- Billing. Deliver wholesale billing as a platform capability by end of FY16.
- CRM. Complete onboarding of the small-business unit's agents to Salesforce.
- Legacy. Create a retirement plan for all legacy technology.
- Strategy. Draft and align new Workforce Technology strategies, road maps and principles.
Live data from the cloud
Intuit calls itself a thirty-four-year-old start-up; its history is a run of platform disruptions, each upending the last: DOS, then Windows and Macintosh, then mobile, now the cloud. OKRs work even better in the cloud era because horizontal alignment comes naturally: with open goals, the data and analytics team saw from the start what the financial systems team planned, and the two linked objectives in real time instead of after the fact, a sea change.
Desktop software company. Leaders read operations through a twentieth-century retail lens: postmortems of sales reports and channel flow. They try to predict, but mostly look in the rearview mirror.
Cloud business. Leaders want to know what is happening now: subscriptions this week, trials running, conversion rate. A customer can find a product, skim the page, try it and buy in ten minutes or less, so leaders should check the funnel daily.
EBS therefore needs real-time reporting, data and analytics even while it builds features like wholesale billing, which it wrote into a top-level objective.
- Data marts. Deliver functional data marts for HR and Sales.
- Warehouse. Complete migration to a new Enterprise Data Warehouse built for real-time access.
- One team. Create a single team operating all data visualization tools across Intuit, for a unified strategy.
- Teaching. Create a teaching module to help other teams use visualization tools.
A tool for global collaboration
As Intuit globalizes, asynchronous collaboration is a way of life. With a thirteen-hour gap to Bangalore, live video has limited use: one side sleeps while the other works. Three years earlier there were few practical options; Intuit had invested in workplace tools but lacked persistent chat, collaborative authoring and videoconferencing. People improvised unevenly and productivity slipped.
EBS promoted a workforce-technology key result to its own top-level OKR. Within six months the new emphasis added several tools on a single sign-on: Slack for persistent chat, Google Docs for collaborative editing, Box for content management, BlueJeans for video. The open OKR platform helped teams make the switch, and people now work instead of puzzling over which tool to use.
- Box. Pilot for the first 100 users by mid-quarter.
- BlueJeans. Complete rollout to final users by quarter end.
- Google. Move the first 50 individual-account users to the enterprise account by quarter end.
- Slack. Finalize the contract by end of month 1; complete the rollout plan by quarter end.
Goal setting involves judgment calls. If you temporarily elevate a key result to an objective, say so openly: it is top-level now, and when it no longer needs extra attention it will drift back down into a KR. The system is dynamic; you are always adjusting the altitude.
Studies have long shown frontline staff thrive when they see how their work aligns with company goals, and Tysen finds this especially true at remote sites. A Bangalore employee put it this way: my objective is a key result of my manager's OKR, which ties to the top EBS objective, which ties to Intuit's shift to the cloud, so now I see how my work in India connects to the mission.
Horizontal connections
Intuit was flat from the start, with few tiers between CEO and front line. Cook's rule, the best idea wins, not the biggest title, still holds. Even in its siloed days Intuit was vertically open: you could speak freely to your manager or your manager's manager and get a respectful hearing. OKRs opened EBS horizontally, across teams.
- Too much upward linking. At first everyone instinctively aligned to their managers' goals or Tysen's. He found literally hundreds of key results linked to one of his top-level objectives, and told people their manager was still their manager, but they should disconnect from us and connect to each other.
- E-commerce and billing. The two teams sit under different vice presidents who report to Tysen. If e-commerce builds a shopping cart, billing must ship matching features. Previously the engineering teams ran separately, and their program managers tried, with mixed success, to connect from above; the people doing the work never talked.
- Now. With horizontally transparent OKRs, engineers link to each other's objectives deliberately, iterating each quarter against department goals while working out how to coordinate with peers. EBS is moving away from senior committee mandates toward real autonomy: leaders set context, ask the big questions and supply data, while interconnected groups drive progress.
Intuit's evidence is usage (2,500 objectives, 4,000 views a quarter), not outcomes; Tysen reports no before-and-after numbers on delivery speed or cost. The story also comes from an IT department, where much work is internal service with clear deliverables, so it shows OKRs connecting teams more than it shows them driving market results.
Superpower #3: Track for Accountability
Unlike frozen, set-and-forget business goals, OKRs are living things: tracked, revised and graded through a three-phase life cycle of setup, midlife tracking and wrap-up. Tracking turns goals into a feedback loop that catches drift early and stops effort in the wrong direction. Scores pinpoint what happened, honest self-assessment explains it, and reflection improves the next cycle: no judgments, only learnings.
The setup: give goals a home
General-purpose software can launch an OKR process, but it does not scale.
A Fortune 500 company trying to speed up its goal cadence hit a wall: all 82,000 contributors had recorded their annual objectives in Microsoft Word files. Quarterly OKRs would have meant 328,000 files a year, public in theory but impossible to search for connections or alignment. If you share a goal nobody sees, Doerr asks, is the system really transparent?
When Bill Pence arrived at AOL as global CTO in 2014, company and division goals sat on a spreadsheet and rolled down from there, but never had a home where they connected daily with people. Without frequent updates, goals drift into irrelevance and the gap between plan and reality widens by the day; at quarter's or year's end you are left with zombie OKRs, whats and hows on paper with no life. CEO Tim Armstrong found the goals too disconnected: unlinked, not cascading, not tethered to people's work through the year. In 2016 he brought in a dedicated platform and rolled out OKRs, producing, Pence says, radical transparency, real-time connection and routine coordination.
Contributors are most engaged when they can see how their work contributes. They look for tangible measures of achievement day to day; the year-end bonus only validates what they already know. OKRs tap the more powerful intrinsic value of the work itself.
Best-in-class cloud platforms offer mobile apps, automatic updating, analytics, real-time alerts and integration with Salesforce, JIRA and Zendesk; in three or four clicks users create, track, edit and score OKRs on a dashboard.
- Visibility. Seamless access to the OKRs of your boss, your reports and the whole organization.
- Engagement. Knowing you are working on the right things makes it easier to stay motivated.
- Internal networking. Transparency steers people to colleagues with shared professional interests.
- Savings. No more hours lost digging through meeting notes, emails, Word files and slide decks; the information is ready when you are.
The OKR shepherd
For the system to work, the group using it, executives or a whole organization, must adopt it universally: no exceptions, no opt-outs. There will be late adopters, resisters and procrastinators, so a best practice is to name one or more OKR shepherds. In Google's products department that was SVP Jonathan Rosenberg.
On August 5, 2010 Rosenberg emailed the "Product Gang" under the subject line "Amidst boundless opportunities, 13 PMs whiff on OKRs (names included)." The gist:
- Why it matters. He believes good quarterly OKRs are part of succeeding at Google, so he sends reminders and has managers review them for quality.
- Carrot and stick. He had tried nice notes and mean ones, from threatening his "Pit of Despair" in October 2007 to celebrating near-perfection in July 2008, until compliance reached near 100 percent.
- The relapse. Then he stopped sending notes, and that quarter several PMs missed the deadline and others did not grade their Q2 OKRs. Lesson: what matters is not the type of note but that he sends one at all. (New AdMob staff and those who finished late in July got a pass.)
- The challenge. With so many opportunities (search, ads, display, YouTube, Android, enterprise, local, commerce, Chrome, TV, mobile, social), anyone who cannot write OKRs that excite them should come see him.
- The ask. Write OKRs on time, grade last quarter's, do it well, and post them so the link on the intranet page works. This is not administrative busywork but how priorities are set and teams kept together.
Midlife tracking
The Fitbit craze shows people want to see their progress, visually and to the percentage point. Research suggests measured headway can motivate more than public recognition, money or even reaching the goal. Daniel Pink, author of Drive, says the single greatest motivator is making progress in one's work; the days people progress are the days they feel most engaged.
- Weekly, not daily. Most platforms show progress visually. Unlike Fitbit steps, OKRs need no daily tracking, but regular, preferably weekly, check-ins prevent slippage.
- Drucker on check-ins. Without an action plan the executive is a prisoner of events; without check-ins to re-examine the plan, there is no way to tell events that matter from noise.
- Write, monitor, share. Writing a goal down raises the odds of reaching it; monitoring progress and sharing with colleagues, both built into OKRs, raises them more. In a California study, people who wrote their goals and sent weekly progress reports to a friend achieved 43 percent more of them than those who only thought about their goals.
OKRs are guardrails, not chains or blinders.
- Continue. A green ("on track") goal: if it isn't broken, don't fix it.
- Update. A yellow ("needs attention") KR or objective: adjust to changes in the work or environment. What could be done differently? Does it need a revised timeline? Should other initiatives be back-burnered to free resources?
- Start. Launch a new OKR mid-cycle whenever the need arises.
- Stop. When a red ("at risk") goal has outlived its usefulness, the best answer may be to drop it.
A real-time dashboard quantifies progress against target and flags what needs attention. OKRs mostly push for more, but they also stop people persisting in the wrong direction. Stephen Covey: if the ladder leans on the wrong wall, each step takes you to the wrong place faster. Continuous tracking makes late surprises less likely, and people can learn from failure and turn a setback into the seed of a new success.
Remind's first revenue service, a peer-to-peer payment system, was a total flop: nobody used it, because it solved no clear problem. The goal was changed at once to an event-driven system: a teacher announces a field trip next week, asks who is coming and whether they want to pay. That changed everything, and usage grew fast.
- End obsolete goals midstream. No need to cling to an outdated projection: strike it and move on. Goals serve purpose, not the reverse.
- The proviso. When an objective is dropped early, notify everyone who depends on it, then reflect: what did I learn that I did not foresee at the start of the quarter, and how will I apply it?
- Check-in rhythm. Contributors and managers review OKRs several times a quarter, reporting progress, naming obstacles and refining KRs; teams and departments meet regularly on shared objectives, and a failing committed OKR gets a rescue plan.
- Google's cadence. Team check-in frequency depends on business needs, the gap between prediction and execution, communication quality, and team size and location: the more dispersed, the more frequent. The benchmark is monthly at minimum, though goal talk is so pervasive that formal meetings sometimes lapse.
Wrap-up: rinse and repeat
OKRs do not expire when the work is done; like any data-driven system, they yield value in post hoc analysis. Wrap-ups, in one-on-ones and team meetings, have three parts: objective scoring, subjective self-assessment and reflection.
Scoring
Scoring records what was achieved and how it might be done differently. A low score forces reassessment: is the objective still worth pursuing, and if so, what must change? On state-of-the-art platforms scores are system-generated, untouched by human hands; homegrown tools may need manual sums.
Objective score = average of its key results' percentage completion, on a scale of 0 to 1.0:
- 0.7 to 1.0 = green. We delivered.
- 0.4 to 0.6 = yellow. We made progress but fell short of completion.
- 0.0 to 0.3 = red. We failed to make real progress.
Intel used a similar formula. Grove's actual Q2 1980 Crush OKR, endorsed by his executives, was graded:
- Five benchmarks published: three done, 0.6, borderline green.
- Repackage the 8086 family: done, as a new product line called iAPX, 1.0.
- 8MHz part into production: set for early May, a fiasco; polysilicon problems pushed it to October, 0.
- Coprocessor sampled by June 15: the goal was 500 parts, Intel shipped 470, 0.9, green.
The average, 62.5 percent (0.625), was respectable. The board judged it somewhat below expectations, but not far, knowing how aggressively goals were set. Intel entered each quarter expecting not to hit everything; a department approaching 100 percent was presumed to be aiming too low, and there would be hell to pay.
Self-assessment
Objective data improves with the goal setter's thoughtful subjective judgment. There may be extenuating circumstances: weak numbers can hide strong effort, and strong numbers can be artificially inflated.
- Fifty calls. The team objective is new customers; your KR is fifty calls. You make 35, a raw 70 percent. If a dozen calls ran for hours and signed eight customers, you might give yourself 1.0. If you procrastinated, rushed all fifty and signed one, perhaps 0.25, since you could have pushed harder. And on reflection, should the KR have counted customers rather than calls?
- Three articles. A PR manager's team KR is three national articles; it places two, but one is a Wall Street Journal cover story. The raw score is 67 percent, but the manager rates it 9 out of 10.
- Guides, not grades. Googlers use OKRs in self-assessments as guides, not grades. Shona Brown, former SVP of business operations, told Doerr the point was not red, yellow or green but a list of what people delivered above business as usual and connected to company goals; the aim is everyone working on the right things.
| KR: bring in ten new customers | Progress | Score | Self-assessment |
|---|---|---|---|
| Market slump | 70% | 0.9 | A slump made it far tougher than expected; seven customers was an exceptionally good effort and outcome |
| Done in eight weeks | 100% | 0.7 | Reaching it eight weeks into the quarter showed the OKR was set too low |
| Lucky referral | 80% | 0.6 | Eight customers came more from luck than work: one brought in five others |
| Low-value wins | 90% | 0.5 | Nine customers landed, but seven bring in little revenue |
Some people grade themselves too harshly; others need to be challenged. An alert facilitator or team leader helps recalibrate. The numbers matter less than contextual feedback and broad team discussion. Scores pinpoint what went right or wrong; self-assessments drive a better goal-setting process next quarter.
Reflection
OKRs are action-oriented, but relentless action becomes a hamster wheel of grim striving. Doerr's formula for satisfaction: set aggressive goals, achieve most of them, pause to reflect, repeat. A Harvard Business School study found learning from direct experience works better when coupled with reflection, the deliberate effort to synthesize and articulate its lessons. John Dewey went further: we learn not from experience but from reflecting on experience.
- Did I accomplish all my objectives? If so, what contributed to my success?
- If not, what obstacles did I encounter?
- If I rewrote a goal achieved in full, what would I change?
- What have I learned that might alter my approach to the next cycle's OKRs?
Wrap-ups look back and forward at once. An unfinished objective may roll over with fresh key results, or its moment may have passed and it is dropped; either way, management judgment comes first. Finally, after appraising the work and owning the shortfalls, savor the progress: throw the team a party.
Doerr's pitch for dedicated cloud OKR software comes from an interested party: Kleiner Perkins backed the OKR platform BetterWorks in 2014, with Doerr on its board, which the chapter does not mention. The 43 percent "California study" (Gail Matthews, Dominican University) was a small, unreviewed conference study. And the advice to adjust raw scores by self-assessment helps learning but also opens room for grade inflation unless managers recalibrate.
Track: The Gates Foundation Story
In 2000 the Bill & Melinda Gates Foundation became a $20 billion start-up obliged to spend at least $1 billion a year toward the boldest of missions: everyone deserves a healthy and productive life. CEO Patty Stonesifer brought in OKRs in 2002 after hearing Doerr's pitch at an Amazon board meeting. Tracking let the foundation be ambitious and disciplined at once: when key results showed no progress, the money moved. Bill Gates adds the chapter's sharpest rule: a mission is not an objective.
The $20 billion start-up
Gates had just stepped down as Microsoft CEO but remained chairman and chief product strategist. He needed a way to channel the foundation's ambition, adapt to shifting conditions in the field, and let himself, extremely busy and fast-moving, make the best choices. Doerr's lesson: the higher the stakes, the more important tracking becomes, to flag looming problems, back out of dead ends and change goals on the run. The leaders recruited scores of global-health experts and told them to stop thinking incrementally and ask what they would do with unlimited resources. By 2002 the foundation had grown enough to need structured goal setting.
Stonesifer: a blank sheet needs a yardstick
The foundation had the gift of a blank sheet of paper, how do you want to change the world?, but the gift was heavy: with so big a goal, how do you know you are making progress? It felt bound to be responsible with the capital, and Bill and Melinda wanted a disciplined system behind hard choices.
- Best in the world. It borrowed Jim Collins's question, what can you be best in the world at?, then laid OKRs on top. Its DNA was the belief in a healthy, productive life for all and Bill and Melinda's faith in technology as a driver of change.
- DALYs. For a time it used Disability-Adjusted Life Years (DALYs), a global-health metric giving a data-driven basis for key results, for example to weigh an investment in micronutrients against one to fight river blindness. DALYs pointed it to vaccines, which add enormous numbers of productive life years. A credible metric plus key results made things clear.
Gates: aim higher
For Gates, ambitious directional goals were always central at Microsoft. From youth he thought software was magic. In the early days the exponential growth in transistors mapped onto device performance; chips, storage and communications were all improving exponentially, displays less so, but the graphical interface would be fast enough. The missing element was the magic software. He gave up surer paths in law or science because the prospect he called "information at your fingertips" was so fascinating.
Even before forming their partnership, Gates and Paul Allen talked of a computer on every desk and in every home. IBM and others, with resources and skills far beyond theirs, did not aim for it because they did not see it as possible, so they pushed less hard. Gates and Allen saw that Moore's law would make machines cheap and bring software to critical mass. Their biggest advantage: they aimed higher.
Making goals concrete
When Bill and Melinda put $20 billion into the foundation in 2000, it was at once a start-up and the world's biggest foundation, and payout rules required it to spend a minimum of $1 billion a year.
- Learned, not invented. Gates watched Andy Grove manage people on subgoals (key results) and studied Japanese companies, learning how to handle people falling short.
- Green-yellow-red. Stonesifer brought in OKRs and their color scoring, and it worked. Using OKRs in grant reviews made Gates confident about what they were pursuing. Still running Microsoft with little time, he relied on Stonesifer to make their agreement efficient, and the goals process was central to that.
- Two rejected grants. Twice he turned a grant down because its goals were not clear enough; OKRs gave him confidence it was the right call.
- Too aspirational. He loves goals but says they must be handled well. The malaria team once believed it would eradicate malaria by 2015, which was unrealistic; an over-aspirational goal damages credibility.
Gates sees philanthropy confuse objectives with missions all the time. A mission is directional. An objective is a set of concrete steps you are deliberately engaged in and actually trying to achieve. Ambition is fine, but you must know how to scale it and how to measure it. Philanthropy is improving as people from high-performance businesses tilt its culture: a good mission is not enough without a concrete objective and a route to it.
Ambitious and disciplined at once
Stonesifer's account of how tracking worked in practice:
- Reallocate. When measurable key results showed no progress, or that an objective was unachievable, the foundation moved the capital.
- Beats against huge goals. For a very ambitious objective like eliminating Guinea worm disease, it set quarterly and annual beats for substantial key results, to know whether dollars were making headway.
- Big goals reveal levers. Until you set a goal as big as vaccinating every child everywhere, you cannot find which lever or mix of levers matters most. Annual strategy reviews began by asking whether the objective was eradication or expanding vaccine reach.
- The 80/90 rule. Then key results got practical, such as the rule at the Global Alliance for Vaccines and Immunization (GAVI): 80 percent of districts reach 90 percent or more coverage. KRs align daily work, and over time they are ratcheted up toward the big goal.
- Breaking it down is the hard part. Setting big goals was easier than working out which rocks must be moved. Bill and Melinda want progress, but bold goals do not faze them.
Stonesifer admits they were sometimes measuring the wrong thing, though always trying to hold themselves accountable. Private foundations lack a market signal of impact, so they must watch whether their data actually leads to the ultimate goal, and they learned so fast they sometimes changed data sets midstream. Her example: a seed that doubles yam production looks like success on that number, until it turns out nobody will plant it because the yams take four times longer to cook at night.
Malaria: tracking what matters
Doerr's case in point is the fight against the mosquito, the deadliest animal on the planet. In 2016 the foundation and the British government launched a five-year, $4.3 billion campaign against malaria, the deadliest tropical disease. Driven by data, they widened their focus from a transmission-blocking vaccine to a comprehensive eradication strategy.
Objective: global eradication of malaria by 2040.
- Proof. Prove to the world that a radical cure-based approach can achieve regional elimination.
- Tools. Prepare for scale-up by creating the necessary tools, notably the SERCAP (Single Exposure Radical Cure and Prophylaxis) diagnostic.
- Momentum. Sustain current global progress so the environment favors an eradication push.
The top-line aim is to remove the Plasmodium parasite from the human population, with special emphasis on drug-resistant strains. Gates concedes it will not be easy; Doerr thinks it has a real chance because the team is tracking what matters.
Stonesifer's own admission that they were sometimes measuring the wrong thing is the chapter's most useful caveat. WHO data show global malaria progress stalled after about 2015, so the 2040 objective illustrates Gates's own warning about goals that outrun the evidence; tracking is what exposes that gap. The institution has also changed: after Melinda French Gates left in 2024 it became simply the Gates Foundation, and in 2025 Gates announced it would spend down its endowment and close by 2045.
Superpower #4: Stretch for Amazing
The fourth superpower is stretch: goals set beyond apparent reach unlock capacity, creativity and new business models that safe goals never will. Bill Campbell: companies that don't innovate (not iterate) die, and conservative goal setting stifles innovation. Hard goals reached only 60 to 70 percent of the time still beat easy goals reached in full, so Google and Intel deliberately set aspirational OKRs they expect to miss.
Why stretch
OKRs push people past their comfort zones, to the border between abilities and dreams. For companies that want to live long and prosper, stretching is compulsory; Doerr likens innovation to oxygen, without which nobody wins.
- Big Hairy Audacious Goals. In Good to Great Jim Collins describes the BHAG: a huge, daunting goal, like a mountain to climb, that is clear and compelling, understood at once, a unifying focal point that builds team spirit toward a finish line. Like the 1960s NASA moon mission, it grabs people in the gut.
- Locke's evidence. Edwin Locke, patriarch of goal-setting research, mined a dozen studies across many arenas for the link between goal difficulty and achievement. The result was unequivocal: the harder the goal, the higher the performance. People with very hard goals reached them far less often than those with very easy ones, yet consistently performed better. Stretched workers were also more motivated and engaged: challenging goals raise task interest and help people find the pleasure in an activity.
- Grand challenges. In 2007 the National Academy of Engineering asked a panel including Larry Page, futurist Ray Kurzweil and geneticist J. Craig Venter to pick fourteen Grand Engineering Challenges for the century. After a year's debate they chose quintessential stretch goals such as fusion energy, reverse-engineering the brain, preventing nuclear terror and securing cyberspace.
- Ordinary work, extraordinary level. Stretch goals need not be rarefied. They fit Doerr's favorite definition of entrepreneurs: people who do more than anyone thinks possible with less than anyone thinks possible. In start-ups and market leaders alike they sharpen an entrepreneurial culture and drive operating excellence.
- Edmunds.com. Chief digital officer Philip Potloff says Edmunds is trying to change how car retailing works, and the only way to boil down its crazy, industry-changing goals is through OKRs.
Aspirational goals draw on all the others: focus and commitment to target goals that matter; a transparent, aligned and connected organization to achieve far beyond the norm; and quantifiable tracking, without which you cannot know when you have reached the amazing objective.
Two OKR baskets
Google splits its OKRs into two categories, a distinction with real consequences.
Committed OKRs. Tied to Google's metrics: product releases, bookings, hiring, customers. Management sets them at company level, employees at department level. Generally, such as sales and revenue goals, they must be met in full (100 percent) within the set time.
Aspirational (stretch) OKRs. Bigger-picture, higher-risk, future-tilting ideas that can come from any tier and aim to mobilize the whole organization. Hard by definition; failure, at an average rate of 40 percent, is part of the territory.
The weighting between the baskets is a cultural question that varies by organization and by quarter. Leaders should ask what kind of company they need to be in the coming year: agile and daring to crack a new market, or conservative and operational to firm up an existing position? Are they in survival mode, or is there cash to bet big for a big reward? What does the business need right now?
Our need to stretch
Andy Grove admired psychologist Abraham Maslow, whose hierarchy of needs holds that only after basic concerns are met (food and shelter, then safety, then love and belonging) do people move to higher motivations, topped by self-actualization.
Grove was fascinated that some people, unprompted, kept testing the outer limits of their abilities to reach a personal best; never self-satisfied, they were a manager's dream. But not everyone is a natural achiever, and for the rest stretched goals elicit maximum output; such goal setting, he wrote, is essential for peak performance from yourself and your subordinates. Intel prized calculated risk takers; it was where Doerr learned to stretch and to dare to fail.
In Operation Crush, salespeople were measured by design wins, products designed around the 8086. The task force under Bill Davidow first set 1,000 design wins in a calendar year, already 50 percent above the previous year. As Dave House, general manager for microprocessors, recalls, it was Jim Lally or Davidow who named 1,000, an enormous-seeming number, and during planning it somehow became 2,000, the figure taken to field sales.
- The ask. 2,000 meant one win per salesperson per month, roughly tripling output for a chip so unpopular that longtime customers were hanging up on reps. A beaten-down sales force now faced Mount Everest. Davidow says he chose 2,000 because Intel needed a rallying point.
- The carrot. Everyone who hit the mark won a trip for two to Tahiti.
- The twist. Lally added that if a single rep missed quota, that rep's entire district office lost the trip.
- The nudge. Early numbers trailed so badly that the task force considered relaxing the design-win criterion. Then, that summer, full-color Tahiti brochures mysteriously arrived in every salesperson's home mailbox. By Q3, peer pressure on laggards was enormous.
- The outcome. The year's tally passed 2,300, the 8086 dominated, Intel's future was secured, and virtually the whole sales force went to Tahiti.
The gospel of 10x
If Grove is the patron saint of aspirational OKRs, Larry Page is their high priest. Google is synonymous with exponentially aggressive goals, what Steven Levy calls "the gospel of 10x."
Earlier webmail offered meager storage, typically 2 to 4 megabytes; users deleted old mail to make room, and archives were a pipe dream. During development Google considered 100 MB, already a huge upgrade. By the public release in 2004, that goal was forgotten: Gmail offered a full gigabyte, up to 500 times the competition, so users could keep mail forever. It did not improve the category; it reinvented it and forced rivals to raise their game by orders of magnitude.
Page observes that most people assume things are impossible instead of starting from real-world physics and working out what is actually possible. Levy's gloss in Wired: to Page, a 10 percent improvement means doing what everyone else does; you probably won't fail spectacularly, but you are guaranteed not to succeed wildly. So Page expects products ten times better than rivals, which rules out hidden efficiencies and code tweaks and requires rethinking problems, exploring what is technically possible, and having fun doing it.
- The 70 percent zone. In line with Grove's standard, Google sets aspirational OKRs at 60 to 70 percent attainment: performance is expected to fall short at least 30 percent of the time, and that counts as success.
- Misfires come with it. Google has had colossal failures, from Helpouts to Google Answers; living in the 70 percent zone means many moonshots and a willingness to court failure.
- Impossible goals force better questions. At the start of a period, not one goal may look possible, so Googlers must ask what radical, high-risk action to consider, what to stop doing, and where to move resources or find partners. By the deadline, a healthy fraction of the impossible goals are somehow met in full.
Stretch variables
A stretch goal must not feel like a long march to nowhere, nor be imposed from above regardless of realities on the ground; stretch a team too fast and too far and it may snap. High-effort, high-risk goals need employee commitment, so leaders must convey two things: the importance of the outcome and the belief that it is attainable.
- Margin for error. Few organizations have Google's resources to absorb a crashed moonshot. Risk tolerance varies and changes over time; the greater the margin for error, the further a company can extend itself.
- Morale. A 40 percent OKR failure rate may seem too risky and too discouraging whatever leaders say. For high achievers, anything short of perfection saps morale. At Risk Management Solutions in California, former HR leader Amelia Merrill says there are "more degrees than employees": people used to A's, not B's, find less than 100 percent culturally very hard.
- The all-committed approach. At MyFitnessPal, Mike Lee treats every OKR as committed: hard but fully attainable, the bar set where he thinks it belongs, so finishing them all means real progress. Reasonable, Doerr says, but risky: will people shy away from objectives that might top out at 90 percent? Better to set at least a modest stretch; with experience, key results become more precise and more aggressive.
- No magic number. Ask how your team can create maximum value and what amazing would look like.
The reward of having met one of these challenging goals is that you get to play again.
Grove's full coda: set uncomfortably tough objectives and meet them, then after ten milliseconds of celebration set another highly difficult set and meet those too. Stretching for amazing is the place to start greatness, and by no means the place to stop.
The Tahiti story ties a stretch goal directly to a reward plus collective punishment, the opposite of Grove's rule to keep OKRs separate from compensation. It worked once, under existential threat; as a routine design it invites the sandbagging and gaming the book warns against. Locke's "harder is better" finding also assumes commitment and ability: his own later work with Latham notes that goals beyond a person's capability depress performance, which is why the "belief it is attainable" condition matters.
Stretch: The Google Chrome Story
Sundar Pichai tells how Google built Chrome from nothing on a stretch OKR it missed twice (2008 and 2009) before hitting it in 2010. Astro Teller of Google X supplies the logic: asking for 50 miles per gallon means retooling a car, asking for 500 means starting over. A stretch target matters less for the date it is reached than for the direction, urgency and rethinking it forces, and a miss handled well becomes the next plan.
The stretch idea and the man
Doerr opens with Teller, head of the X lab behind Project Loon and the self-driving car: a modest gain lets you tinker with the existing design, a tenfold gain forces a clean sheet. In 2008 Pichai, then vice president of product development, was in exactly that clean-sheet position with Chrome. Doerr notes that Chrome is now the most popular browser on both desktop and mobile, and quotes Larry Page: set a crazy goal, miss it, and you will still achieve something remarkable. Pichai's own career is cast as a stretch goal in person: in October 2015, at 43, he became Google's third CEO, over 60,000+ employees and $80 billion in revenue.
Pre-phone and post-phone
Pichai grew up in Chennai in the 1980s, son of an electrical engineer, in modest circumstances. The waiting list for a rotary telephone was three to four years; the family got one when he was twelve, and neighbours came in to use it. Before the phone, checking whether his mother's blood test was ready meant a bus ride, a queue and often a "come back tomorrow", about three hours in all; after it, a call. He remembers technology as discrete before-and-after moments. He read every book on computers and semiconductors he could find and set himself one goal, Stanford, as the route into Silicon Valley; scarcity, he thinks, made him dream harder.
The new applications platform
After Stanford he spent five years at Applied Materials in Santa Clara in process engineering and R&D. Visits to Intel let him feel Andy Grove's discipline at the door (he half remembers paying for each cup of coffee). Semiconductor process work demands methodical, sequenced goals, and it taught him to think about goals precisely.
He followed Google closely and was taken with Deskbar, which let Windows users search from a small taskbar window without opening a browser: present when needed and only then, an early growth tool. He joined Google as a product manager in 2004, when the company still revolved around search but Web 2.0, user-generated content and AJAX were turning the web from a content platform into an applications platform. His first job was expanding the distribution of Google Toolbar, which added Google Search to any browser. Within a few years Toolbar users grew more than 10x, his first look at what an ambitious stretch OKR could do.
Rethinking the browser
Google had formed a client-software team, some of whom worked on Firefox to help Mozilla. By 2006 the team was rethinking the browser as a computing platform, almost an operating system, on which people could write applications. That insight produced Chrome, and it set two technical requirements:
- Multiprocess architecture. Each tab runs as its own process, so a crash in one application cannot take down a user's Gmail.
- Much faster JavaScript. Web applications would only feel like desktop software if the scripting engine ran far faster.
CEO Eric Schmidt knew how hard a browser built from scratch would be and told the team that if they did it they had better be serious: unless Chrome was dramatically different, better and faster than existing browsers, there was no point.
Objective: develop the next-generation client platform for web applications, an objective that would shape Google's future. Main key result: Chrome reaches 20 million seven-day active users by year end, starting from zero.
Upping the goal
In Google's OKR culture, 70 percent average attainment counted as success; all-green scorecards meant the team was not stretching. The tension is that Google hires people driven to succeed, and no leader wanted to stand before the company at quarter end with a big red on the screen explaining the failure. That discomfort drove a lot of heroic effort, but a correctly set objective sometimes made a red unavoidable.
Page was good at raising the company OKRs. He wanted Googlers "uncomfortably excited" and to show "a healthy disregard for the impossible," and Pichai tried to do the same for the products team. Writing an OKR that might well fail took courage, but there was no other way to be great.
Pichai's leadership rule: challenge the team without making the goal feel unachievable. Privately he thought there was no way to hit 20 million. He set it anyway, because the number did four things that mattered more than a somewhat arbitrary target on a fixed day:
- Direction. It told the team where to go.
- A barometer. It gave a measure of progress.
- No complacency. It made coasting impossible.
- Daily rethinking. It forced the team to rethink its framework every day.
Then came setbacks. Chrome struggled to reach 3 percent market share, and the Mac version slipped badly, so only Windows users could count toward the target. The good news was that people who tried Chrome loved it, which began to compound growth. The task was to find more users, sooner.
Digging deeper: V8
Google treats speed as identity and wages constant war on latency. In 2008 Page and Sergey Brin wrote a company OKR that caught everyone's attention: make the web as fast as flipping through a magazine. Under it, Chrome set a sub-OKR to make web applications run as smoothly as desktop downloads, with a 10x JavaScript moonshot named V8 after the car engine.
Google found Lars Bak, a Danish programmer who had built virtual machines for Sun Microsystems and held more than a dozen patents. He told them calmly he could make something much faster. Within four months JavaScript ran 10x as fast as on Firefox; within two years, more than 20x. Bak later told Steven Levy (In the Plex) that the team had underestimated what it could do: some stretch goals are less wild than they look.
Pichai calls stretch OKRs an intense exercise in problem solving. Toolbar had taught him to work through troughs: sleepless nights, but cautious optimism in front of the team. Lost users meant an experiment to find out why; compatibility problems meant a group assigned to fix them. He tried to be systematic, not emotional.
Google's moonshot culture asks of every invention how it could scale to a billion, a number that feels abstract early on. Setting a measurable annual objective and chunking it quarter by quarter makes moonshots doable; OKRs supply clear, quantitative targets on the road to qualitative leaps.
Missing, then reframing
Missing 20 million in 2008 made the team dig deeper. They never dropped the objective but reframed it: we did not reach the goal, but we are laying the foundation to break through, so what will we do differently? Among smart people, Pichai says, you need real answers to that; you cannot tap-dance. The basic question was why people would not try a new browser. The answers:
- Distribution deals to put Chrome in front of users.
- Television advertising, once it emerged that many people did not know what a browser did. It became the largest offline campaign in Google's history, including "Dear Sophie," a father's digital scrapbook of his growing daughter that showed how the browser led to Gmail, YouTube and Maps, the web as an applications platform.
Try-fail, try-succeed
| Year | Target (7-day actives) | Outcome | What changed next |
|---|---|---|---|
| 2008 | 20M | Missed | Distribution deals; TV ads ("Dear Sophie") |
| 2009 | 50M | Missed, 38M | Bigger goal, growth reinvented |
| 2010 | 111M | Hit | OEM deals, "Chrome Fast", Mac and Linux, dormant-user alert |
For 2010 Pichai proposed 100 million. Page pushed harder: that was only 10 percent of the world's roughly one billion internet users. Pichai argued 100 million was already very aggressive; they settled on 111 million, a classic stretch that meant reinventing Chrome's business. February: broader distribution deals with OEMs (original equipment manufacturers). March: a US "Chrome Fast" awareness campaign. May: launch on OS X and Linux, ending Windows-only status. Well into Q3 the result was in doubt, until a small change, a passive alert to dormant former users, lifted the total from 87 million to 107 million by the end of Q3, and to 111 million shortly after.
Today Chrome has over a billion active users on mobile alone. Pichai credits OKRs: they are how Google thinks about everything, and always has.
The two misses were not failures of the method. Each forced a sharper question (why won't people switch?) and a new lever (distribution, advertising, new platforms, re-engagement). The stretch number kept the team reinventing the business rather than defending a plan.
The next frontier
Pichai's father came of age when computing meant mainframes, big teams and administrators; all he wanted was an easy way onto the web. Pichai has always prized simplicity: Google Search did complex things behind a plain interface, and he wanted Chrome equally simple whether the user was a child in India or a Stanford professor. In 2008 he gave his retired father a netbook with Chrome. The technology faded away: his father never opened another app or downloaded software, he lived in Chrome. Pichai's lesson is to keep imagining the next frontier (Toolbar to Chrome, and on): his father's experience prompted the idea of an operating system with the browser as its interface, and a laptop built around it, the Chromebook, tapping applications in the cloud. Those, he says, were stretch goals for another day.
The story is told by a winner: Chrome's later dominance makes the misses look like stepping stones. Doerr offers no Google stretch goal that missed and stayed missed, though the previous chapter concedes products like Helpouts died. The decisive 2010 jump also owed much to paid distribution and bundling deals, a lever few companies can buy, and the 87 to 107 million surge from a re-engagement alert suggests the metric (seven-day actives) could be moved by tactics as well as product quality.
Stretch: The YouTube Story
YouTube set a four-year OKR to grow daily watch time tenfold, to one billion hours a day by the end of 2016, and reached it ahead of schedule. CEO Susan Wojcicki and engineering VP Cristos Goodrow show what made it work: first choosing the right metric, then guarding a single goal through a change of leadership. Before YouTube could chase its audacious objective, it had to figure out how to measure what mattered: watch time, not views.
A second Google stretch
Doerr adds a second Google stretch story because the company is full of them. Wojcicki, whom Time called the most powerful woman on the internet, rented her Menlo Park garage to Google in September 1998, days after it incorporated; she became employee No. 16 and its first marketing manager. Eight years later, when analysts doubted YouTube would survive, she was a leading voice persuading Google's board to buy it, convinced online video would disrupt network television. As senior VP of advertising and commerce she reworked AdWords and conceived AdSense, in effect driving Google's two main revenue streams.
By 2012 YouTube led the market but its pace of innovation had slowed, and an organization that brakes finds it hard to reaccelerate. When Wojcicki became CEO in 2014 she inherited a goal of growing 10x in four years to a billion daily hours, and she wanted to grow responsibly, not at any cost.
- More than 1 billion users, nearly one third of everyone on the internet.
- Navigable in 70+ languages in 80+ countries.
- Its mobile platform alone reaches more 18-to-49-year-olds than any cable or broadcast network.
Doerr's point: none of this came from a single insight. It took years of rigorous execution, close attention to detail and the structure of OKRs, plus the prior step of finding the right measure.
Wojcicki on OKRs at Google
She leased the garage wanting only rent, then came to know Page and Brin and saw they were better placed than she was to execute ideas she had for her own company. The clincher was a day Google Search went down and she could not work: it had become indispensable, and she reasoned it would be for everyone.
She was present in fall 1999 when Doerr presented OKRs at 2400 Bayshore, Mountain View, an old Sun Microsystems plant of perhaps 42,000 square feet, less than half of it occupied; the meeting was held in the all-hands half. She remembers his "this is an objective, this is a key result" explanation and the football analogy, and recently found his slides, plastic laminates for an overhead projector. Page and Brin had never run or worked in a company; they argued but listened, found measurables intuitive, and were impressed that Intel, a great company, used OKRs.
- Start young. OKRs suit young companies building culture, where clarity matters more with fewer resources; like raising children, rules from the start beat rules imposed on teenagers. Yet mature companies do turnarounds, so no company is too young or too late.
- Leader plus lieutenant. The process needs a leader who embraces it and a lieutenant to ride herd on scoring and reviews; Wojcicki ran OKRs for Page.
- Top-down, with give-and-take. In four-hour meetings Page debated every company objective and people had to defend theirs; experts negotiated the key results: here is the direction, now tell us how you will get there.
- A pulpit for priorities. The meetings let Page stress what he cared about, especially product speed: tell me your speed now; why can't you cut that in half?
- Keep them public. A quarterly all-hands videocast still covers top-level OKRs (Salar Kamangar once ran through the entire roster memorably), detailed discussion happens in teams, and OKRs sit on company and team intranet pages, updated in real time, open to anyone.
If you can't beat 'em
Google Videos launched in 2005, a month before YouTube. The first clip Wojcicki's team put up was a purple Muppet singing a nonsense song from an Italian film; she and Brin were puzzled until her children demanded it again. That revealed a next-generation opportunity: anyone creating video for global distribution. The first real hit was two students singing Backstreet Boys in a dorm room with a roommate studying behind them; user-generated content beat the professional videos.
The fatal flaw was an upload delay, a breach of Google's product rule make it fast. YouTube's uploads were watchable at once. By the time Google fixed it, YouTube was out-streaming Google Videos three to one, but it was swamped by demand and needed capital for infrastructure, so it had to sell. Wojcicki built the spreadsheets justifying the $1.65 billion price and persuaded the founders, who at the last minute asked her to present them to the board. The board approved while doubting her year-on-year user-growth assumption, which is the one thing YouTube has delivered ever since.
Big rocks (Goodrow)
Goodrow arrived from Google Product Search in February 2011, three years before Wojcicki. YouTube's roughly 800 people produced hundreds of OKRs per quarter: a ten-person team would open a Google doc, type 30 or 40 objectives, and finish less than half.
He names two habits engineers bring to goal setting: they hate crossing off any good idea, and they chronically underestimate how long things take (smart people insisting they can surely do more). Forcing teams down to three or four objectives took discipline but made OKRs rigorous and made clear what counted most; he applied the same rule after taking over search and discovery.
When Kamangar handed day-to-day tech leadership to Shishir Mehrotra, Mehrotra brought focus to the whole company with Stephen Covey's metaphor.
You must fit rocks, pebbles and sand into a wide-mouthed one-gallon jar. Pour in sand first, then pebbles, and the rocks will not fit. Put the big rocks in first, then pebbles, then sand, and the sand fills the gaps: everything fits. The most important things must be done first or they will not get done at all.
Until then people let a thousand flowers bloom and nobody could name the top-level OKRs. Leadership now said that all the ideas were good, but asked teams to pick a few as big rocks for the quarter and the year. After that everyone at YouTube knew the top priorities, and Goodrow calls it a giant step toward the goal that consumed his next four years.
A better metric
YouTube knew how to make money but not how to grow viewership. In Google Research, Jim McFadden of the Sibyl team was building a system for "watch next" recommendations (related videos), with large potential to boost views. The question was whether views were the right thing to boost.
Goodrow borrows Satya Nadella's line that when computing is nearly limitless, the scarce commodity is human attention. If people spend more of their time watching, they must be happier with what they watch, which starts a virtuous circle. YouTube's true currency was watch time, not views or clicks, and it needed a new core metric.
Watch time, and only watch time
In September 2011 Goodrow emailed his boss and YouTube's leadership with the subject line "Watch time, and only watch time": other things equal, the goal was to raise watch time. To many at Google it was heresy, for three reasons:
- Search philosophy. Google Search is a switchboard designed to send you off the site to your best destination as fast as possible; maximizing time on site contradicts it.
- Views. Watch time would hurt views, the key metric for users and creators.
- Money. Ads ran only before videos, so fewer video starts meant fewer ads and less revenue, at least initially.
Google Search logic. A user asks "how do I tie a bow tie?" and there are two videos. Rank the precise one-minute tutorial first: people want to tie the tie and get to their event.
Goodrow's logic. Rank the funny ten-minute video first, even if you may not learn the knot. YouTube's job is to keep people engaged; viewers are happier watching seven, or even two, minutes of a ten-minute video than all of a one-minute one. Those who want only the knot will choose the short one.
He won after six months. On the Ides of March (March 15), 2012, YouTube launched a watch-time-optimized recommendation algorithm aimed at engagement and satisfaction, which especially improved music, how-to, entertainment and late-night comedy.
A big round number
At the November 2012 YouTube Leadership Summit in Los Angeles, Mehrotra told a few colleagues he was about to announce a stretch goal for the coming year: one billion hours of daily watch time. Round numbers are simple, and simplicity has power. It meant 10x, so years rather than months. 2015 was too soon and 2017 sounded odd (primes generally do), so minutes before he went on stage they settled on end of 2016: a four-year OKR with rolling annual objectives and quarterly incremental key results.
Objective: reach 1 billion hours of watch time per day by 2016, with growth driven by:
- Search team and Main App (+XX%) and Living Room (+XX%).
- Kids' engagement and gaming watch time (X watch hours per day).
- Launch YouTube VR and grow the VR catalog from X to Y videos.
Principled stretching
Stretch goals crush people who do not believe in them, so framing matters. Mehrotra cut the BHAG down to size: a billion daily hours was under 20 percent of the world's daily television watching. YouTube was not chasing size for its own sake; it was scaling toward something already much bigger.
The team were not 10x absolutists. They accepted some watch-time-negative decisions for users' sake, such as no longer recommending tabloid clickbait like "World's Worst Parents" (a thumbnail of a baby in a pot on a stove). After three weeks watch time in that group was down 0.5 percent; they held firm, and within three months it had recovered and risen, as viewers found more satisfying content. Once the goal was set, though, every change was measured for watch-time impact; anything that might slow progress was estimated precisely and needed internal consensus first.
Getting up to speed (Wojcicki)
Kamangar enjoys early-stage companies and moved on in 2012. YouTube had split into business and technology factions and needed unifying; after a decade on AdWords' complex ecosystem, Wojcicki wanted the job. Most of YouTube thought the billion-hour goal impossible ("it would break the internet"), but she believed a clear, measurable objective would energize people.
When she arrived in February 2014, YouTube was nearly a third of the way through the four years and off pace: watch-time growth had fallen well below what the deadline required. Google may aim for 0.7 on stretch goals in aggregate, but no team sets out to settle for 70 percent, especially once a goal is within reach; no one would have accepted 700 million hours.
She was not sure they would make it, and thought a small miss would be acceptable if the team stayed united and aligned. Google had missed before and simply rolled objectives forward: in her telling, AdSense was a quarterly OKR, and despite huge effort it launched two days late, harmlessly.
For her, OKRs are most useful when you are behind. When she ran Google's mid-quarter OKR updates, the point was to fix things: gather the leadership team, ask each member to name five projects that would close the gap, then extend the OKR and reward positive behaviour. So she was not overly worried about the deadline.
The guardian
Goodrow, the OKR's guardian, saw it as his white whale. At Wojcicki's "up-to-speed" meeting he brought a 46-slide deck and made his point by slide five: we need to catch up. From 2013 through 2016 the billion-hour OKR headlined every annual presentation, with clear interim milestones. He thanked her for keeping the 10x goal, then said they were way behind, that he was freaked out and hoped she was at least a little, and asked her to weigh every priority decision against it.
Bandwidth and the gray
- Protecting egress bandwidth. Delivering video takes far more bytes than email or social media (egress bandwidth), so YouTube reserves server capacity in advance. After the goal was announced, leadership had lobbied to reserve bandwidth through 2016. When Wojcicki took over, Google's server group wanted to renegotiate the apparently exorbitant spend. New and behind plan, she knew lost machines would be hard to recover, so she kicked the can: stay with the plan and meet again in three months. By then more data and more growth made her case easier.
- Keeping an eye on the gray. The goal had become a religion at YouTube, and she respected it, but a black-and-white target can do harm if unmanaged. Daily watch time = daily active viewers (DAVs) × average time per viewer. YouTube was good at the second, the lower-hanging fruit (deepening a relationship is easier than starting one), but research showed more potential in enlarging the user base than in getting teenage boys to watch twice as much. Advertisers wanted new users too.
Mutual support (Goodrow)
New leadership puts everything up for review. Wojcicki could have gone back to a views goal, chosen a revenue goal, or kept watch time alongside three other equal priorities. Any of these, Goodrow says, and YouTube would have been distracted and never caught up. Continuity of one stretch objective across leaders was decisive.
- Names on goals. After Wojcicki arrived, owners' names sat next to company goals with green, yellow or red progress bars; "Cristos" stood beside "one billion hours" at every weekly staff meeting, quarter after quarter. He felt personally responsible.
- Safe to fail, but. He valued Google's creed that it is fine to fail against risky goals, and good things had already come: much better search and recommendations, a higher profile for YouTube inside Google, record morale, even marketers discussing watch time with fervour.
- Total commitment. He had told Mehrotra early that he would resign from Google if they missed the deadline, and meant it; he thinks that intensity kept him going.
At New Year 2016, the start of the final lap, YouTube was on schedule but barely. The summer doldrums followed, people outdoors watching less, and as late as July growth lagged the year-end goal; Goodrow asked teams to consider reordering projects to reaccelerate. In September returning and new viewers amplified every search and recommendation improvement. Engineers hunted for changes worth as little as 0.2 percent of watch time and found about 150 in 2016, needing nearly all of them. By early October growth beat the target rate. Goodrow still checked the graph daily, on vacation and when sick, until one Monday that fall it showed the billion hours crossed over the weekend, ahead of schedule. The next day, for the first time in over three years, he did not look.
What the stretch produced
- Unplanned dividends. Daily views soared in parallel with watch time; stretch OKRs set forces in motion whose destination you cannot predict.
- Cover from the top. Page and Brin were not involved in YouTube's business and could not know the course was optimal, but liked that it was ambitious and clear. When many in Google Search openly doubted the goal, both said on stage that they supported it, giving YouTube autonomy.
Thinking bigger (Wojcicki)
The aspirational goal reset the whole organization: if YouTube was going to be that big, teams asked, perhaps it needed a new architecture and new storage. It pushed the company to prepare for the future and think bigger. She doubts the goal could have been met in four years without the OKR's process, structure and clarity; in a fast-growing firm people need a benchmark, and the trick is choosing the right one. The billion hours gave the engineers a North Star.
Metrics age, though. In 2013 watch time was the best gauge of the YouTube experience; now the company weighs viewer satisfaction and social responsibility too. Two ten-minute videos give the same watch time, but which made you happier?
2015 User satisfaction surveys plus likes and dislikes factored into recommendations, so users feel their time was well spent.
2017 A breaking-news shelf on the homepage surfaces content from authoritative news sources.
Ongoing A broader set of meaningful signals in recommendations; as YouTube's role in society changes, the hunt for the right metrics, and the right OKRs, continues.
- Engagement's side effects. Critics and former engineers later argued that optimizing recommendations for watch time pushed viewers toward sensational and extreme content; the chapter's brief nods to "responsibility" understate that debate.
- A date slip. Wojcicki places AdSense's launch in 2007; Google introduced AdSense in 2003.
- Since publication. Wojcicki stepped down as YouTube CEO in 2023 and died in 2024.
Continuous Performance Management: OKRs and CFRs
Part Two, "The New World of Work", opens with Doerr's claim that annual reviews are costly, exhausting and mostly futile, and that quarterly OKRs need a matching human system: continuous performance management, delivered through CFRs: Conversations, Feedback and Recognition. OKRs and CFRs reinforce each other, and the first step toward both is blunt: divorce compensation from OKRs.
The case against the annual review
Doerr's epigraph is Sheryl Sandberg's chain: talk changes minds, minds change behaviour, behaviour changes institutions. The annual review fails on cost, fairness and value.
- 7.5 hours of manager time per direct report, on average.
- Only 12 percent of HR leaders call the process "highly effective" at driving business value; only 6 percent think it is worth the time.
- 10 percent of the Fortune 500 had already dropped it at the time of writing, and the number was growing.
- 31 percent lower voluntary turnover at "high-recognition" companies than at poor-recognition ones.
Year-end ratings are distorted by recency bias and weighed down by stack rankings and bell curves, so they cannot be fair or well measured. The painful lesson is that people cannot be reduced to numbers. Even Peter Drucker, champion of measured goals, said a manager's first role is the personal one: relationships, mutual confidence, building a community. Doerr adds the line usually credited to Einstein: not everything that can be counted counts, and not everything that counts can be counted.
Just as quarterly OKRs made annual goals obsolete, Doerr wants an equivalent upgrade for performance management, a new HR model for a new world of work: continuous performance management, implemented with CFRs.
- Conversations. An authentic, richly textured exchange between manager and contributor, aimed at driving performance.
- Feedback. Bidirectional or networked communication among peers to evaluate progress and guide future improvement.
- Recognition. Expressions of appreciation to deserving people for contributions of all sizes.
Like OKRs, CFRs promote transparency, accountability, empowerment and teamwork at every level. Doerr's image: CFRs ignite OKRs and boost them into orbit, completing Grove's method and giving OKRs a human voice.
The marriage of the two
Doug Dennerline, CEO of BetterWorks (the pioneer in putting both tools on the cloud and smartphones, used by hundreds of organizations), calls the marriage of the two the real home run. A conversation limited to whether a goal was hit loses context; continuous management surfaces the critical questions:
- Was the goal harder than expected when it was set?
- Was it the right goal in the first place?
- Is it motivating?
- Should we double down on the two or three things that worked last quarter, or pivot?
Those insights must come from all over the organization. The reverse also holds: without goals, what are you even talking about? People feel more fulfilled with clear, aligned targets, because they can see how their work connects to the whole. In Doerr's football image, objectives are the goalposts and KRs the yard markers; CFRs are everything that binds the team between games: the Monday tape reviews, midweek squad meetings, pre-play huddles and end-zone celebrations.
Reinventing HR
The change is under way, among big firms and especially among start-ups less bound by tradition, and Doerr argues almost every HR custom needs rethinking for a mobile, agile, non-hierarchical workforce. Replacing or augmenting the review with ongoing conversation and real-time feedback lets companies improve all year; alignment and transparency become daily habits, and managers tackle struggling employees like firefighters, at once, instead of waiting for a day of reckoning. Doerr claims this lifts performance from bottom to top, helps morale and development for leaders and contributors alike, and is stronger still combined with OKRs' quarterly goals and tracking.
In the transition, many organizations add criteria such as competencies and team play. Many large firms run parallel tracks, annual reviews alongside continuous management, and some may stay there for good; others cut the cord, dropping ratings and rankings for transparent, collaboratively built, multidimensional criteria.
| Dimension | Annual performance management | Continuous performance management |
|---|---|---|
| Cadence | Annual feedback | Continuous feedback |
| Pay | Tied to compensation | Decoupled from compensation |
| Style | Directing, autocratic | Coaching, democratic |
| Focus | Outcome focused | Process focused |
| Lens | Weakness based | Strength based |
| Fairness | Prone to bias | Fact driven |
The Washington, D.C. international development nonprofit Pact adopted OKRs because its review process was already moving to a faster cadence; director Tim Staffa says it killed its annual review and replaced it with more frequent touch points, called Propel:
- Monthly one-on-ones between employee and manager on how things are going.
- A quarterly OKR review: what did you set out to do, what did and didn't you do, why or why not, and what can change?
- A semiannual professional-development conversation: where you have been, are, and want to go, and how manager and organization can help.
- Ongoing, self-driven insight: when a colleague says "nice job" after your presentation, do not just say thanks; ask what one thing they liked, to capture specific feedback in real time.
An amicable divorce
The first step to continuous management is to divorce compensation (raises and bonuses) from OKRs. They become two conversations with their own cadences and calendars: a backward-looking assessment, typically at year end, and an ongoing, forward-looking dialogue between leaders and contributors built on five questions.
- What are you working on?
- How are you doing; how are your OKRs coming along?
- Is anything impeding your work?
- What do you need from me to be (more) successful?
- How do you need to grow to reach your career goals?
Doerr does not want reviews and goals fully severed. A data-driven summary of achievements is an antidote to ratings bias, and because OKRs capture someone's most meaningful work they are reliable input for the next cycle. But when goals are used to set pay, people sandbag: they play defence, stop stretching, get bored, and the organization suffers most.
A set extreme stretch goals and hit 75 percent of them. B hit 90 percent of his key results, but his manager knows he did not push himself and that he skipped several important team meetings. Should B get the larger bonus? No, says Doerr, not if you want to keep initiative and morale.
At Google, according to people chief Laszlo Bock, OKRs make up a third or less of a performance rating. They rank behind cross-functional feedback and, above all, context: goals can be wrong, markets can go haywire, a client can leave and force a rebuild. Google keeps raw goal scores apart from pay decisions, and wipes OKR scores from the system after each cycle. No formula captures complex human behaviour; that is what judgment is for. OKRs and pay can stay friends, Doerr says, but they no longer live together.
Conversations
Drucker was among the first to stress regular one-on-ones. Grove estimated that ninety minutes of a manager's time can raise the quality of a subordinate's work for two weeks, and made one-on-ones mandatory at Intel. In his account the meeting is mutual teaching: the supervisor passes on skills and approaches through specific problems, the subordinate reports in detail what he is doing and worried about. It is the subordinate's meeting, with agenda and tone set by him; the supervisor is there to learn and coach. It is also the place for heart-to-heart issues: is he satisfied with his performance, is something frustrating him, does he doubt where he is going?
Modern tools make Grove's rules more practical. Good one-on-ones dig beneath day-to-day work and keep a set cadence, weekly to quarterly, as needed.
- Goal setting and reflection. Set the employee's OKR plan for the coming cycle and align it with organizational priorities.
- Ongoing progress updates. Brief, data-driven check-ins on real-time progress, with problem solving as needed.
- Two-way coaching. Help contributors reach their potential and managers do a better job.
- Career growth. Build skills, find growth opportunities, and widen the employee's view of a future at the company.
- Lightweight performance reviews. Gather inputs and summarize accomplishments since the last meeting against the organization's needs, held apart from the pay and bonus review.
Managers thus shift from taskmasters to teachers, coaches and mentors. Doerr's example: a head of product has waffled over a design decision and put a release date at risk. Before the next executive meeting, a CEO-as-coach asks whether he could be more decisive: lay out the two best options and state his preference. If he agrees, there is a plan. Criticism looks back; coaching aims at future improvement.
Feedback
Sandberg (Lean In) defines feedback as an opinion grounded in observation and experience that tells us the impression we make on others. Without it, nobody knows how to improve. Doerr cites research that today's workers want to be empowered and inspired rather than directed, to give feedback to managers rather than wait a year to receive it, and to discuss, share and track goals with peers. Public OKRs prompt questions from all directions: are these the right things to focus on, would completing them count as a big success, how could we stretch further?
Feedback works only if it is specific:
- Negative. You started last week's meeting late, and it came across as disorganized.
- Positive. Your opening anecdote grabbed the audience, and closing with next action steps worked well.
In developing organizations feedback is run by HR and scheduled; in mature ones it is ad hoc, real-time and multidirectional, open dialogue anywhere in the company. If we rate Uber drivers (and they rate us) and rate the raters on Yelp, workplaces can support two-way feedback: what do you need from me, and here is what I need from you. The anonymous suggestion box has given way to always-on tools: quick surveys, anonymous social networks, even apps rating meetings and their organizers.
Peer-to-peer (360-degree) feedback adds another lens and can be anonymous, public or in between, depending on purpose: if it is meant to develop careers it goes privately to the individual; if it is meant to expose organizational problems it goes to HR. It is especially valuable in cross-functional work: as horizontal communication opens up and OKRs combine with 360 feedback, Doerr predicts the silo becomes a relic.
Recognition
Doerr calls recognition the most underestimated and least understood CFR. The gold watch for long service is gone; modern recognition is performance-based and horizontal, crowdsourcing meritocracy. When JetBlue installed a values-driven peer-to-peer recognition system and leaders started noticing people who had flown under the radar, employee-satisfaction metrics nearly doubled. Saying thank you, however soft it sounds, builds engagement (the 31 percent turnover gap above).
- Peer-to-peer. When peers consistently recognize achievements, a culture of gratitude forms; Zume Pizza's Friday all-hands "roundup" ends with unsolicited, unedited shout-outs from anyone to anyone.
- Clear criteria. Recognize actions and results: special projects completed, company goals met, values shown. Replace "Employee of the Month" with "Achievement of the Month."
- Share the stories. Newsletters and company blogs give the narrative behind an accomplishment, adding meaning.
- Frequent and attainable. Hail small wins too: the extra push to a deadline, the polish on a proposal, things managers take for granted.
- Tie it to strategy. Customer service, innovation, teamwork, cost cutting: any priority can be backed by a timely shout-out.
OKR platforms are built for peer recognition: quarterly goals keep resetting where praise matters most, and transparent OKRs make it natural to celebrate big wins and small. As teams connect this way, a recognition engine revs up across the company; anyone can cheer anyone's goal regardless of title or department, and each cheer is a step toward operating excellence, the shared purpose of OKRs and CFRs.
Doerr's chief witness, BetterWorks, is a company he backed through Kleiner Perkins, as is Zume; the chapter reads partly as a sales case for OKR software. The survey figures are cited without methodology, and the evidence that CFRs raise performance (rather than satisfaction) is thinner than the confident tone suggests.
Ditching Annual Performance Reviews: The Adobe Story
In 2012 Adobe executive Donna Morris, jet-lagged in India, told a reporter that Adobe would abolish annual reviews and stack rankings, before she had told her HR team or the CEO. Her scramble produced Check-in, a lightweight system of continuous conversations decoupled from pay. Adobe swapped eighty thousand manager hours of demoralizing paperwork for feedback at least every six weeks, and voluntary attrition dropped sharply.
The problem
About six years before the book, Adobe ran the usual annual review. It cost managers eight hours per employee, demoralized everyone, and created no visible value. Each February voluntary attrition spiked as people disappointed by their reviews took their talent elsewhere. Meanwhile Adobe was racing to a cloud subscription model, a real-time business it needed to win, while HR stayed shackled to an annual rhythm.
- 8 hours of manager time per employee per review.
- 80,000 manager hours a year in total, nearly 40 full-time hires' worth.
- Voluntary attrition spiked every February, after disappointing reviews.
The slip that started it
On a 2012 business trip to India, Morris vented to a reporter about traditional performance management and, guard lowered by jet lag, said Adobe planned to abolish annual reviews and stack rankings for more frequent, forward-looking feedback. Good idea; she had not yet discussed it with her HR staff or Adobe's CEO.
She hustled to bring the company round. On Adobe's intranet she framed the task as reviewing contributions, rewarding accomplishments and exchanging feedback, asked whether these had to be fused into one cumbersome process, and invited people to say what they would want in its place: what would inspire, motivate and value contribution better? The post set off one of the most widely engaged discussions in Adobe's history.
How Check-in works
Under Check-in, managers, employees and peers hold several conversations a year, and leaders across the company own the process rather than leaning on HR.
- Three focus areas. Quarterly "goals and expectations" (Adobe's term for OKRs), regular feedback, and career development and growth.
- Light by design. Lightweight, flexible and transparent, with minimal structure and no tracking or paperwork; sessions are called by contributors.
- Pay separate. Check-ins are decoupled from compensation. Forced-distribution stack rankings gave way to an annual Rewards Check-in, where trained managers scale pay by performance, business impact, skill scarcity and market conditions, with no fixed guidelines.
Since Check-in started in fall 2012, Doerr reports, voluntary attrition has dropped sharply, and CFRs have invigorated the whole operation.
Morris: living the values
Adobe's four founding values are genuine, exceptional, innovative and involved, and the old review contradicted every one. Her pitch to staff: what if there were no ratings, no rankings and no forms, and instead everyone knew what was expected and could grow a career at a company that valued them?
- Training in waves. A series of 30-to-60-minute web trainings went to senior leaders, then managers, then employees, with 90 percent participation; each quarter addressed a new phase, from setting expectations to giving and receiving feedback.
- Employee Resource Center. Templates and videos build constructive-feedback skills, easing in Adobe's many engineers who were not used to open dialogue.
- Leaders as models. Executives showed they were open to feedback and comfortable having their vision questioned.
- Managers as business leaders. Each manager gets budgets for base pay, incentives and equity to distribute as they see fit, which makes them truly responsible for their reports; employees, through regular Check-ins on goals, action items and development, have input too. With no fixed pots, teammates stop being competitors.
Feedback in every direction
People want to drive their own success and to know how they are doing while they are doing it, not be graded at year end. Check-in guarantees specific feedback at least every six weeks, and in practice weekly, so the process leads rather than lags. Feedback flows three ways:
- Manager to employee, the default.
- Employee to manager, such as: I felt out on a limb on project X and needed more support.
- Peer to peer, essential in Adobe's matrix: Morris has communication, finance and legal partners who report elsewhere but sit on strong dotted lines to her, and they review expectations and give each other feedback.
- Executive support.
- Clarity on company objectives and how they align with individual priorities (Adobe's goals and expectations, i.e. OKRs).
- Investment in training for managers and leaders: not courses away from work but one-hour online sessions with role-played vignettes, such as the steps for giving difficult feedback.
Morris calls well-delivered corrective feedback the greatest gift you can give, because it can change mindset and behaviour for the better. The culture shift is toward "it's okay to make a mistake, because that's how I'll grow."
For Morris, HR exists for the success of the business: its job is to consult with other leaders on making everyone successful. Success is not built by forms, rankings and ratings, or by policies that bog people down, but by mechanisms that build capability.
| Dimension | Before: annual review | After: Check-in |
|---|---|---|
| Priorities | Set at the start of the year, often not revisited | Set and adjusted with manager regularly |
| Feedback | Long cycle of submitting accomplishments, soliciting feedback, writing reviews | Ongoing dialogue, no formal written review |
| Pay | Onerous rating and ranking of everyone for salary and equity | No rating or ranking; manager sets salary and equity annually on performance |
| Cadence | Sessions inconsistent and unmonitored; productivity spikes at year end | Quarterly at least, ongoing as the norm; steady productivity |
| HR role | Manage paperwork and make sure steps are done | Equip people for constructive conversations |
| Training | HR partners who could not reach everyone | Central Employee Resource Center, on demand |
Why it matters
For a service business, nothing is worth more than engaged employees who feel they make a difference and want to stay; turnover is costly. The best turnover is internal turnover, people growing careers inside the company. People are not natural nomads, Morris argues; they want a place where they can make a real impact, and Check-in is how Adobe provides it.
The chapter says attrition "dropped sharply" but gives no figure, and offers no measure of whether manager-set pay without guidelines introduced new bias. The account comes entirely from the executive who championed the change.
Baking Better Every Day: The Zume Pizza Story
Co-CEOs Julia Collins and Alex Garden used OKRs and CFRs at Zume Pizza, a robot-assisted Silicon Valley start-up, less to grow fast than to build the company's people. Doerr's framing: beyond growth, OKRs and CFRs have quieter daily effects, grooming executives, giving quiet contributors a voice, turning leaders into better communicators and contributors into more rigorous thinkers. Structured goals teach managers to think like executives early, working within real constraints while pushing against them, before the company outgrows them.
David against three Goliaths
The $10 billion US pizza-delivery market belonged to three chains, Domino's, Pizza Hut and Papa John's: unremarkable pizza, but strong brands and economies of scale. When Zume opened in spring 2016 in an out-of-the-way concrete bunker in Silicon Valley, skeptics mocked "roboticized, artisanal pizza" as a West Coast gimmick.
Nearly two years on, Doerr says, Zume was beating the odds with world-class pizza at a competitive price. Machines take the rote tasks, freeing people for creative work, and labour savings go into ingredients: non-GMO flour, organically grown tomatoes, local vegetables, healthfully cured meats. The pie arrives hot as little as five minutes after the order.
- The line. Online and app orders feed a conveyor; robots stretch and shape the dough, apply sauce and slide pies into an 800-degree oven. Zume planned to automate everything, cheese and toppings included, through to loading the part-baked pies onto algorithmically run, baking-on-the-way trucks (perhaps one day driverless).
- Traction. 10 percent market share in its local area within three months; expansion across the Bay Area in 2018, then the West Coast, nationally, and overseas by 2019.
- Ambition. Garden, who met OKRs as president of Zynga Studios, said Zume would be "the Amazon of food."
For a David facing Goliaths, Doerr argues, there is no margin for unfocused work or misaligned staff.
Why a start-up needed OKRs (Collins)
At first Zume lived in the founders' two heads: ask either a question and you got the same answer. With the CTO they were still "the three cheeses" and it worked. Then came a fourth cheese, and by seven people a question about the day's main task produced eight different answers.
- Waterfall tool. LiquidPlanner was ideal for linear work like building the kitchen: pour concrete, let it dry, apply epoxy, let it dry, install the walk-in fridge.
- Complexity. By June 2016, before launch, Zume had 16 salaried staff plus about three dozen hourly kitchen workers and delivery "pilots", and was simultaneously doing large-scale manufacturing, robot integration, software and menu creation, with layers of interdependency. Engineers ran two-week sprints in JIRA each morning.
- The gap. Neither tool answered the big question: what is the most important thing to do? A talented team left alone chased good ideas that were not always in sync.
So Zume adopted OKRs three weeks after the first pizza went out. For the first two cycles the co-CEOs wrote all objectives and demanded 100 percent top-down alignment to get mission-critical work done, planning to loosen up as survival became less pressing.
Achieving what's real (Garden)
OKRs' explicit value is tying the organization to leadership's true ambitions. Garden stresses an implicit value start-ups overlook: OKRs are a superb training tool for managers, teaching them to run a business within existing limits.
Push the envelope, but the envelope is real: everyone is constrained by time, money and people, and the bigger the organization the more entropy, as in thermodynamics. At Microsoft's Xbox Live, as a general manager, Garden worked under visionary executives whose wishes outran the organization's capabilities; the hows and most of the whats were left to divisional foot soldiers executing an overscoped mandate. A good goal-setting process from the start would have saved much grief.
The old model says executives grow more abstract as they rise, buffered by middle managers. Garden thinks that fails at today's pace: OKRs work only when the top is unconditionally committed, like a religious calling. Proselytizing is thankless, and staff may dislike you through an adoption curve of up to a year, but it pays.
Seven internal benefits
The rest of the chapter is organized as a list of what OKRs did inside Zume.
1. Better discipline
Collins puts discipline first, for the co-CEOs themselves. Garden explains the hidden career trap: as an individual contributor you are graded on the volume and quality of your work; as a manager you are paid for the quality of your decisions, but nobody tells you the rules changed, so when you hit a wall you just work harder. The right, counterintuitive move is to stop, shut out the noise and choose the best path for your team relative to the organization.
- Formalized reflection. At least once a quarter, OKRs make people step into a quiet place and check how their decisions align with the company; they start thinking in the macro.
- Precision. You cannot write a ninety-page OKR; you must pick three to five things and how to measure each. By the time someone is made a manager, they already think like one.
- The start-up trap. Many start-ups say they move too fast for structure and will figure it out. Without early habits, unsuccessful companies outgrow their leaders and die, and successful ones outgrow their team, which gets replaced. Better to train people to lead when their department is a staff of one.
OKRs, in Garden's words, forge people: they make stronger executives, spare rookie mistakes and put a big company's rigor and rhythm into a small one. The first benefit was the process itself: forcing people to think about the business thoughtfully, transparently and interdependently accelerated performance.
2. Better engagement
Removing ambiguity makes some people decide this is not what they signed up for and leave, and others feel inspired because they finally know the aim. Either way there is clarity, and those who stay are bought in; team sports only work when the whole team plays.
The process grew more collaborative with familiarity. In Q3 2016 the founders wrote the company OKRs and department heads turned some company key results into their own objectives, a straight cascade. In Q4 the founders still wrote the objectives but the team "KRed" them from the top; with a more creative role, the OKRs improved, still stretched but more realistic.
Zume's core technology, baking on the way, is where it disrupts the industry. A top Q4 company objective was deploying the big trucks, 26 feet long with 56 ovens, tied to logistics and order-prediction software, so pizzas finish en route. Product manager Vaibhav Goel owned it. Objective: complete the truck delivery fleet for 250 Polaris (Mountain View HQ).
- Deliver 126 fully certified ovens by 11/30.
- Deliver 11 fully certified racks by 11/30.
- Deliver 2 fully certified full-format delivery vehicles by 11/30.
Airtight: hit the three KRs and the objective is reached.
- Equal voice. Loud people repeat their point until they win; quieter people get neglected. OKRs give every department equal weight: each team's objectives come up on the screen for comment and support, so nobody can suffer in silence.
- Mining dissent. Collins: a good company values different opinions and digs dissent out into the open; that is how a meritocracy forms.
- Execs first. Zume ran OKRs for two full quarters at executive level before individual contributors, to set the culture. Oddly, the biggest early skeptics became the most active users.
Marketing director Joseph Suzuki expected another diet program, follow the process and be thin and beautiful, or more bookkeeping. Instead his biweekly check-ins gave him a few minutes to see how his goals rolled up to the company's quarter. At a start-up, especially in marketing, people wear many hats and drown in tactical minutiae, losing sight of land; the check-ins reset his compass, so each report, campaign or event connected to something bigger.
3. Better transparency
When a fly ball drops between two outfielders, somebody must call for it, or it falls in or they collide. Zume's outfielders were marketing and product, both leads a month into the job, new to OKRs, at a company new to itself: who owned revenue targets? Seeing the confusion, the founders split the objective into new revenue (marketing) and repeat revenue (product), and the heads drilled down from there. The conversation was a by-product of the OKR process: anything not clearly delineated shows up at once.
4. Better teamwork
In eight months Zume launched a food, logistics, robotics and manufacturing company from a standing start. Garden used OKRs to teach a culture of consideration: reflexively asking how your work affects others and how you depend on them. The team was eclectic: executive chef Aaron Butkus came up through New York mom-and-pop restaurants, fleet manager Mike Bessoni from film production, plus a product wonk and software engineers. OKRs were their Esperanto. The seven-member leadership team lunches every Monday and discusses OKRs about every other week, with phrases like "who owns the customer?" and "how would you key-result that goal?" understood by all.
The best pie fails if it arrives cold, so Bessoni and Butkus share a customer-satisfaction objective. If manufacturing is late getting a vehicle online, Bessoni's KR to expand the delivery radius is threatened, which also hits service area, revenue, and Suzuki's top-line revenue OKR. In another life Bessoni might have yelled at the manufacturing lead. Saying "my KR is at risk" is less charged and more constructive: everyone already agreed the KR and its dependency, so there is no judgment, just a problem. The leads then advocate for each other for more resources from the founders.
Butkus adds the kitchen view: a new seasonal pie cannot be improvised. Marketing needs at least a week's notice, photography and design must shoot it, and the website and mobile app teams depend on it. With his deadline built into a key result, he delivers on time and sees the bigger picture; OKRs remind people they are one small, odd community working together on everything.
5. Better conversations
Every two weeks, each Zume employee has a one-hour one-on-one with their manager (the co-CEOs meet each other). It is sacred: no lateness, no cancelling, and one rule, no talking about work. The agenda is the person's goals for the next two to three years, broken into a two-week plan.
- Three opening questions. What makes you very happy? What saps your energy? How would you describe your dream job?
- Two expectations. Always tell the truth; always do the right thing.
- The promise. Meet those and you get unconditional backing, 100 percent of the time, plus Garden's personal guarantee you will reach your next personal and professional goals within three years.
It looks like altruism but connects people to the business and prevents churn. Paradoxically, these non-work meetings are a forum for ongoing performance feedback: a leader can ask why a goal that clearly matters made no progress in two weeks, and learn what drives or blocks the person's career. Regular deep talks also show when people need recharging: after an all-out sprint, raise time for personal-development goals from 5 percent to 15 or 20 percent for a quarter. It sounds like a tax, but it sets up the next two or three quarters of execution.
6. Better culture
For Collins, culture is the common language that ensures people mean the same thing and that it has meaning; it supplies a shared decision framework, without which functions cannot be made replicable and scalable. Above that sits the values layer: who do we want to be, how should people feel about their work and product, what impact do we want on the world?
Zume's two founding principles came from Collins's first phone call with Garden and hang on a giant poster on the kitchen wall:
- Serving food to people is a sacred trust.
- Every American has a right to delicious, affordable, healthy food.
Objective: delight customers. Detail: feeding people is a sacred trust, kept only through the best service and food quality; customers must be so happy they order more and rave to friends.
- Net Promoter Score of 42 or better.
- Order rating of 4.6/5.0 or better.
- 75 percent of customers prefer Zume to a competitor in a blind taste test.
The mission governs daily choices: it would be easy to add a little more salt, or sugar to the sauce instead of hunting for the freshest tomatoes, the small, creeping compromises that erode identity. Every hire gets mission and values training at onboarding; clear expectations make the founders accountable too. In a best-idea-wins culture anyone may call out anyone, especially the CEO; when someone challenges the founders in an open forum, they deliberately make a huge fuss of it to give others permission.
7. Better leaders
The great bosses Collins worked for shared a cold, sober focus: twenty minutes with them showed uncluttered thinking and clear priorities. Fund-raising, robot pizza and kitchen building mean rapid context switching that can feel frenetic, but knowing the company objectives like your last name is calming; however crazy things get, she can default back to what matters.
Doerr's firm backed Zume and he sat on its board, so this is an insider's exhibit written barely two years after launch. Zume later raised a reported $375 million from SoftBank (2018), abandoned pizza in 2020 with mass layoffs to pivot to compostable packaging, and shut down in 2023. Good internal practices did not rescue a weak business model: OKRs measure execution, not whether the strategy is right.
Culture
Goals need a medium, as sound waves do, and for OKRs and CFRs that medium is culture: an organization's living values and beliefs. Doerr argues that healthy culture and structured goal setting are interdependent, natural partners in pursuit of operating excellence. OKRs bring purpose and clarity to culture change, CFRs supply the energy for the journey, and together they build transparency and accountability, the tent poles of sustained high performance.
Why culture
Doerr opens with Jeff Bezos (you need a culture that applauds small, innovative ideas) and the saying that culture eats strategy for breakfast. Founders worry about protecting culture as they grow, heads of large companies use OKRs and CFRs as tools for culture change, and job seekers increasingly rank cultural fit first. OKRs carry leaders' priorities; CFRs make sure they are transmitted; culture is the medium both travel through. Doerr admits he has no simple answer to building a positive culture, but offers OKRs and CFRs as a blueprint: align teams on a handful of common objectives, then unite them with lightweight, goal-oriented communication.
Grove: culture as efficiency
In High Output Management, Grove defined culture as a set of values and beliefs plus familiarity with how things are and should be done, and called a strong, positive culture absolutely essential. As an engineer he equated culture with efficiency, a manual for faster, more reliable decisions. An intelligent corporate citizen who shares the values behaves consistently in similar conditions, so managers are spared the inefficiency of formal rules, procedures and regulations. Management must nurture the shared values, objectives and methods on which trust depends, in two ways: by articulation, spelling them out, and, more important, by example. Grove role-modelled Intel's standards and taught them to new hires in his iOPEC seminars.
- People oriented. Strong mutual commitment, respect for all work, challenge and opportunity.
- Openness. Highlighting problems is expected.
- Issue resolution. Clean and crisp; confrontation must be constructive.
- Results. Output orientation in all work; superficiality not respected; reward success with positive feedback.
- Discipline. Excellence in a competitive, complex environment demands it.
- Risk taking. Required by high technology; low fear of failing and of self-exposure; champions.
- Trust and integrity.
Project Aristotle: what makes Google teams good
Grove's prized qualities (collective accountability, fearless risk taking, measurable achievement) are also valued at Google. In Project Aristotle, an internal study of 180 teams, standout performance correlated with "yes" answers to five questions.
- Structure and clarity. Are our team's goals, roles and execution plans clear?
- Psychological safety. Can we take risks on this team without feeling insecure or embarrassed?
- Meaning of work. Is the work personally important to each of us?
- Dependability. Can we count on each other for high-quality work on time?
- Impact of work. Do we fundamentally believe the work matters?
Structure and clarity is the very reason OKRs exist; the other four are facets of healthy culture tied to the OKR superpowers and CFRs. Take dependability: transparency and alignment make people more diligent about obligations. At Google, teams own goals collectively, successes and failures, while individuals own specific key results; peak performance comes from collaboration plus accountability.
In an OKR culture you push toward a goal not because the boss ordered it but because every OKR is visibly important to the company and to colleagues who rely on you. Nobody wants to be the one holding the team back; everyone takes pride in progress. It is a social contract, but a self-governed one.
Catalysts and nourishers
Teresa Amabile and Steven Kramer (The Progress Principle) analysed 26 project teams, 238 people and 12,000 diary entries and found high-motivation cultures combine two elements:
Catalysts (like OKRs). Actions that support the work: clear goals, autonomy, enough resources and time, help with the work, openly learning from problems and successes, and free exchange of ideas.
Nourishers (like CFRs). Acts of interpersonal support: respect and recognition, encouragement, emotional comfort, and chances for affiliation.
In culture change, OKRs give purpose and clarity, CFRs the energy. Where people have real conversations and get constructive feedback and recognition, enthusiasm spreads, as do stretch thinking and daily improvement. Doerr's claim: companies that treat people as valued partners have the best service, products and sales growth, and will win.
Pulsing: a listening system
As continuous management spreads, the annual employee survey is giving way to real-time feedback. One frontier is pulsing, an online snapshot of culture: short, quick, wide-ranging questionnaires scheduled weekly or monthly by HR or run as an ongoing "drip" campaign. Sample questions:
- Are you getting enough sleep?
- Have you met your manager recently to discuss goals and expectations?
- Do you have a clear sense of your career path?
- Are you getting enough challenge, motivation and energy; do you feel "in the zone"?
Feedback is a listening system: leaders cannot wait for bad Glassdoor reviews or for valued people to quit; they must capture signals as they are emitted. Doerr imagines a goal platform that pulses two or three questions at each login and merges quantitative goal progress with qualitative conversation and pulse data, and predicts software that will prompt a manager: "Talk to Bob, something's going on with his team." OKRs build goal muscle; CFRs make it flexible; pulsing gauges real-time health, work and culture.
Coursera: OKRs tied to values
Online-learning leader Coursera adopted OKRs in 2013, a year after its founding, with input from then-president Lila Ibrahim, an Intel alumna who revered Grove. It did something rare: it tied OKRs explicitly to its values and mission, a world where anyone, anywhere can transform their life through the best learning. Team objectives rolled up to top-line strategic objectives, which rolled up to five core values, each mapped to its own OKRs.
| Core value | What it means |
|---|---|
| Students first | Engage and add value for students; extend reach to new ones |
| Great partners | Be a great partner to universities |
| Think big, advance pedagogy | Build an innovative, world-class education platform |
| Care for teammates; be human, be humble | Build a strong, healthy organization |
| Do good, do well | Experiment toward a sustainable business model |
Objective: extend Coursera's reach to new students.
- Run A/B tests, learn and iterate on ways to acquire new students and engage existing ones.
- Increase mobile monthly active users to 150k.
- Create internal tools to track key growth metrics.
- Launch features that help instructors make more engaging videos.
OKRs let teams state their goals and align with company objectives and values; years later, Doerr says, Coursera's friendly, inclusive culture still contrasts with the blustery, combative style of many Valley start-ups. Former CEO Rick Levin says he cannot imagine the company without them: each quarter they look back to hold themselves accountable and ahead to live their values better.
Seidman: out-behave to outperform
In 2007 business philosopher Dov Seidman published HOW: Why HOW We Do Anything Means Everything. His premise: culture guides behaviour, how things really happen, and in an open, hyperconnected world behaviour defines a company more than products or market share; it is the one thing that cannot be copied or commoditized.
- Out-behave to outperform. Companies that out-behave rivals also outperform them.
- Self-governing organizations. A values-driven model where long-term legacy beats next quarter's ROI; such firms inspire rather than merely engage, replace rules with shared principles, and swap carrots and sticks for common purpose.
- Next right thing. Once employees only had to do the next thing right, following orders; now they must do the next right thing. A rulebook says what you can do; culture says what you should do.
Proclaiming values such as courage or creativity is easy; scaling them needs a system with metrics. Seidman's view: what we choose to measure reveals our values, because measuring something tells people it matters. His firm LRN has tested the idea in annual HOW reports. Where Grove balanced quantitative goals with qualitative ones, Seidman quantifies abstract values: his trust index measures behaviours, not perceptions. He does not ask whether people feel the company is honest; he looks at information flows (hoarded, rationed on a need-to-know basis, or flowing freely?) and asks whether going around your boss to someone senior gets you punished or celebrated.
- Self-governing organizations: 8 percent of companies, up from 3 percent in 2012.
- 96 percent of them scored high on systematic innovation.
- 95 percent had superior employee engagement and loyalty.
- 94 percent reported increased market share.
Transparency and what comes next
Seidman names active transparency (people opening up, sharing the truth, bringing others in, being vulnerable) as the most powerful cultural force, and Doerr pictures Grove smiling: an OKR/CFR culture is above all transparent. His lessons from Intel, Google and dozens of others: vision-based leadership beats command-and-control, the flatter the org chart the more agile the organization, and two-way, networked performance management lets people grow into greatness. Seidman's closing point is that our ability to connect is itself an engine of growth and innovation.
OKRs and CFRs, Doerr concludes, build the pillars of a values-driven culture: top-down alignment, team-first networking, bottom-up autonomy and engagement. The two closing stories show the relationship from both sides: at Lumeris, culture had to change before OKRs could work; at ONE, a charismatic founder (a literal rock star, Bono) used OKRs to change culture from the top.
Aristotle, Amabile and the HOW report show association, not that OKRs cause good culture; the HOW figures come from Seidman's own consultancy, and "94 percent gained share" lacks a comparison group. The Coursera photo caption also misspells cofounder Daphne Koller as "Kohler."
Culture Change: The Lumeris Story
At Lumeris, a St. Louis health-care technology firm, OKRs first failed because the culture rejected them, in HR chief Andrew Cole's image, as antibodies reject a donor organ. Following Jim Collins (right people on the bus, in the right seats, before stepping on the gas), Lumeris fixed its culture first and relaunched OKRs second. When an organization isn't ready for openness and accountability, culture work must come before OKRs; then OKRs sustain and deepen the new culture.
The business
Lumeris supplies software, services and know-how to health-care providers and payers, from university hospital networks to traditional insurers. It began in 2006 by partnering with 200 St. Louis-area physicians through a federally regulated insurer, Essence Healthcare, to serve 65,000 Missouri seniors on a Medicare Advantage plan.
- The model. Mining a large trove of patient data, Lumeris helps partners turn fee-for-service, volume-based "sick care" into value-based care that rewards prevention and discourages needless tests and harmful hospital stays.
- Who is responsible. Primary-care doctors own their patients cradle to grave; the aim is better quality of life while saving resources and money, and Lumeris claims the two goals work together.
- The moonshot. CEO Mike Long wants to rationalize the US health supply chain. Every other industry runs on transparent cost, quality, service and choice; health care is opaque, so doctors barely know which services are ordered for you, let alone their cost, and cannot be held accountable for financial outcomes.
A data-transparent firm seemed a natural fit for OKRs. But Cole, the former head of HR and an experienced architect of organizational change, warns that unaddressed cultural barriers set the antibodies loose.
OKRs on paper
When Cole arrived, Lumeris had run OKRs for three quarterly cycles, with a reportedly outstanding participation rate. A deep dive showed it was superficial: at quarter end a lone HR person chased managers "like a Jack Russell" for numbers before the board meeting; people adjusted metrics after the fact, claimed completion, added a date and ticked the box. It looked great in PowerPoint but wasn't real.
- No rationale. Few understood the business case for OKRs.
- No buy-in. Executive leadership had not explicitly committed.
- No accountability. Nobody held anyone responsible for getting the system right.
- No connection. Objectives were unrelated to actual work; asked why something was in their OKRs, managers often had no idea how it linked to company aims. It was window dressing.
Cole tries to understand an organization before breaking things, but two quarters later he still doubted OKRs could be saved. In a closed board session he asked Doerr whether Lumeris could drop them if they did not fit; Doerr said "Absolutely." Cole's diagnosis: a passive-aggressive culture and an unanswered question everyone was asking, what's in this for me? People did not trust the program, however sincere, and it could not succeed unless the environment changed.
Two clashing cultures
The executive team had introduced OKRs partly to merge two cultures:
Essence. The insurer formed by the doctors' group: risk averse, in the Hippocratic spirit, nurturing a proprietary model in a hypercompetitive industry.
Lumeris. Pushed to the edge for the next big tech and data insights, and shared its learnings with the world.
As demand surged, the culture gap slowed everything. In May 2015, eleven weeks after Cole arrived, the company was reorganized entirely under the Lumeris name (one company, one name). OKRs might become the common language, but they had to wait: without cultural alignment the best operational strategy fails.
HR transformation
People watch what you do more than what you say.
- Leaders out, with dignity. Some senior leaders were old-school autocrats not living the core values (ownership, accountability, passion for the job, loyalty to the team). Nothing else would matter until they left; Lumeris made sure they went with dignity and respect intact, a telling moment in any transformation.
- An obligation to call out executives. At every culture meeting, staff were told they had the right, indeed the obligation, to hold executives to the stated culture: book a meeting, send an email, or stop them in the hallway.
- The inflection point. It took three months before anyone did. Then Long asked a lunch group why anyone would want to work somewhere afraid of mutual accountability, and people began to believe. Conversion still happened one conversation at a time, showing that collaboration, shared accountability and transparency would be rewarded.
- Hiring is culture. Culture is ultimately the people you recruit and their values. Middle management had A and B players but also C players and below, hired on wrong criteria and vague interviews; no tool works with the wrong instruction manual.
- Speed. Time is the enemy of transformation. In under 18 months Lumeris replaced 85 percent of its HR professionals; with senior leaders and front line on board, it tackled the harder task of strengthening middle management, typically three years to steady state.
Objective: institute a culture that attracts and retains A players.
- Focus on hiring A-player managers and leaders.
- Optimize the recruitment function to attract A-player talent.
- Scrub all job descriptions.
- Retrain everyone involved in interviewing.
- Ensure ongoing mentoring and coaching opportunities.
- Create a culture of learning for new and existing employees.
OKR resurrection
Late in 2015 Cole had HR dissect the failed first attempt. A second try would mean retraining everybody; there would be no third chance. In April 2016 Lumeris relaunched with a 60-day pilot for 100 people in operations, with sharper training and better software. The doubtful SVP of operations and delivery became an enthusiast within two weeks, emailing the group: why did you word this objective that way, what is the metric, this OKR does not match what I see in client feedback. His people concluded he's paying attention and looked harder.
Winning people over was slow. Transparency is scary; admitting failures publicly can terrify people raised since kindergarten to hide them. Cole's image is a first scuba dive to 35 feet: adrenaline and fear, then exhilaration and a new view of what lies beneath the surface.
- Honest two-way conversation. Managers began to see what made reports tick: their wish to connect to something larger and to be recognized for work that matters.
- Known weaknesses. Through OKRs each side learns the other's weaknesses without fear of being caught out; a side benefit is that managers learn to hire people who offset their own limitations.
- Failure without shame. People stopped dancing around setbacks; OKRs help you fail smart and fail fast.
The tide turned, with former naysayers saying they now saw how it could work for them.
- 98 percent of pilot members became active platform users.
- 72 percent set at least one objective aligned to company goals.
- 92 percent said they now understood what my manager expects of me.
- Q3 2016 rollout to all 800 employees; HR met all 250+ managers in five weeks; 75 percent of staff created at least one OKR in that first full quarter.
Transparency without judgment
Art Glasgow joined as president and COO in spring 2016. He and Cole agreed OKRs were pointless unless they went all the way; Glasgow became executive sponsor and told an all-hands that OKRs were how the company would be run and how bosses would be measured, the carrot that balanced the stick. His tone was "brutal transparency without judgment," and Cole says he made the job less lonely.
For the Q3 rollout Lumeris built its own coaches' training program. Over five weeks the rebuilt HR team met every one of 250+ managers in classroom-sized groups, held open houses for one-on-one talks, and promised there were no dumb questions; the sessions drove engagement. Goal setting being more art than science, training went beyond wording objectives and KRs to a cultural agenda:
- Transparency. Why does it matter, why should other departments know your goals, and why does our work matter?
- True accountability. The difference between accountability with respect (for others' failings) and accountability with vulnerability (for your own).
- Work through others. How OKRs help managers get work done through others, essential for scaling, and how to get other teams to adopt your objective as a priority.
- Timing. When to stretch a team's workload or ease off the throttle; when to move an objective to someone else, rewrite it for clarity, or drop it. Timing builds confidence.
No handbook answers these; the wisdom sits with leaders close to their teams who can show what success looks like and know when to declare victory (Cole: not too soon). The payoff: 75 percent created OKRs in Q3 2016, retention improved, involuntary exits fell, and Lumeris was hiring the right people and keeping those who thrive.
Selling your reds
Soon after arriving, Glasgow held a full-day off-site business review, now a monthly fixture. Top-line OKRs go on screen, so everyone sees who is hitting objectives. Glasgow dislikes yellows, so each OKR is green (on track) or red (at risk): no bell-curve ambiguity, nowhere for problems to hide. In three hours a dozen executives spend little time on greens and instead "sell" their reds; the team votes on the most important at-risk OKRs for the company, brainstorms as long as needed, and colleagues volunteer to "buy" someone else's red. "We're all in the same bathwater."
The new Lumeris values interdependence and intentional coordination. Jeff Smith, SVP for US markets, says OKRs make people work on the business, not just in it; regional heads quarterback opportunities instead of running solo, a move from hero culture to team culture. He was surprised to see operations and delivery tie their objectives to his sales goals: where people once said "I'm in delivery, you're in sales, do your job," they now step up like a wide receiver offering to run a play.
Doerr's summary of the sequence: Lumeris first needed the right culture for OKRs to take root, then needed OKRs to sustain and deepen it, a campaign that never ends.
Outcome
2017 was a banner year: Lumeris led the market for value-based care, and Glasgow said that as the market shifted his sales plan might become realistic enough to need stretch goals. At the time of writing it had partnerships with payers, provider groups and health systems in 18 states, covering more than one million lives. Doerr claims the Missouri model, adopted nationally, could save up to $800 billion a year in wasteful spending while improving quality and length of life. OKRs are now part of the wallpaper; in Cole's diving image, once people had seen the company beneath the surface they kept diving back in.
Doerr repeats that the Missouri model could save up to $800 billion a year nationally with no source or method. Pilot statistics come from the company's own platform, and "retention moving the right way" is unquantified.
Culture Change: Bono's ONE Campaign Story
Where Lumeris used OKRs to lock in culture change after the fact, Bono's story shows them springboarding a cultural reset. OKRs gave his advocacy groups, DATA and the ONE Campaign, a structure for passion: fewer goals, clearer ownership, and a framework that let ONE remake itself from working on Africa to working in and with Africa. If the heart doesn't rhyme with the head, passion means nothing; OKRs gave ONE's outrage walls and floors while still demanding more red.
An experiment in anti-apathy
For nearly two decades Bono has run what he calls a global experiment in anti-apathy. Doerr sets out the record.
2000 Bono's first BHAG, through Jubilee 2000: $100 billion of debt relief for the poorest countries.
2002 Cofounds DATA (Debt, AIDS, Trade, Africa) with a Gates Foundation start-up grant, to push policy on poverty, disease and development in Africa with governments and NGOs; Bill Gates later called it the best million dollars he ever spent.
2004 Launches the ONE Campaign, a nonpartisan, grassroots activist coalition, the outside game to DATA's inside game.
2005 Bono and Bill and Melinda Gates named Time's Persons of the Year.
2017 ONE reports 8.9 million members (December), over 3 million in Africa.
ONE has helped deliver nearly $50 billion for health initiatives and lobbied for anti-corruption transparency rules and for channelling African oil and gas revenue into the fight against extreme poverty. Doerr was struck by Bono's "factivism", fact-based activism; in ONE's hard-headed, results-oriented culture OKRs were an easy sell, and for ten years they have clarified priorities. Former CEO David Lane: ONE needed a discipline to keep from trying to do everything, and OKRs were central to its philosophical shift toward helping developing countries develop themselves.
The U2 template
U2 had big goals from the start (Bono jokes about early megalomania) and modest skill: Edge was already accomplished, Larry a fairly good drummer, Bono a poor singer, and Adam could barely play bass. Their logic: we are not as good as the other bands, so we'd better be better. What they had was chemistry; the others had everything, but U2 had it.
Their first measures of effectiveness were about usefulness: could music be useful, could art drive political change? In 1979, at 18, they played an anti-apartheid show and a pro-contraception show, a big deal in Ireland. In their early twenties they deliberately annoyed Irish paramilitary groups and their fellow travellers, insisting that bombing children in supermarkets can never be right, and gauged their impact by the bile that came back. Later, wanting to chart, they measured success by tickets, then albums: a live phenomenon whose singles had struggled.
Picking our fights
DATA was formed like a band: Lucy Matthew, Bobby Shriver, Jamie Drummond and Bono, nobody sure who played what. They were not hippies or wishful thinkers but "more punk rock", tough-minded opportunists working one idea at a time on a vicious schedule, first debt cancellation for the poorest countries.
Next came universal access to anti-AIDS drugs, and people laughed: impossible, and why fight the most expensive disease instead of malaria, river blindness or polio? Bono's answer: the pills (two then, one now) were a visible symbol of inequality. In Dublin or Palo Alto you get them; in Lilongwe, Malawi, you do not, so an accident of longitude and latitude decides who lives. Everyone knew that was wrong, so the argument was winnable. Years before OKRs he already used a goal method: picture Everest, describe how hard the climb is, then describe how to reach the summit. First describe it; then climb it.
- 21 million people on antiretroviral therapy.
- AIDS-related deaths down 45 percent in ten years.
- New HIV infections in children down by more than half; Bono expected to end mother-to-child transmission by 2020 and to see an AIDS-free world in his lifetime.
Growing up with OKRs
The NGO band was entrepreneurial and tracked goals internally, but process has limits. With real impact and access, DATA needed more measurable procedures and outcomes. Then eleven groups joined to form the ONE Campaign coalition, full of gifted people with far too many goals: a green revolution in Africa, girls' education, energy poverty, global warming. They were all over the map.
- Two cultures. DATA and ONE merged very different cultures, which was tricky.
- Too little transparency. Without clear sign-offs on goals you get overlap and dissonance, people confused about their jobs; for a time there was a real schism.
- Stretched too thin. They never thought small, but goals so gigantic wore people out.
Board chairman Tom Freston saw the value of OKRs and made them part of the operation. Bono says OKRs saved them: they forced clear thinking and agreement on what was achievable with the resources at hand, gave a frame to hang passion on (without one the brain is too abstract), and their traffic-light color coding transformed board meetings, sharpening strategy, execution and results.
The pivot: a seat at the table
At his first ONE board meeting Doerr asked who the client was. The world's poorest and most vulnerable, came the answer. Do they have a seat at the table? Of course, the whole table is for them. Doerr pressed: can you picture that person, and shouldn't they be physically at the table? That question seeded the pivot. It echoed a Senegalese man in Paris who told Bono a proverb: if you want to cut a man's hair, it is better if he is in the room. The loving message: be careful assuming you know what we want; you are not African, and the messiah complex has not always ended well.
In 2002 in southeast Africa Bono saw people with HIV queuing to die. He and other activists sounded dramatic alarms and he told staff never to say AIDS without adding "emergency." By 2009 there was backlash: some well-off Africans objected to the framing, and economist Dambisa Moyo's Dead Aid argued that aid did more harm than good and undermined efforts to rebrand the continent as a place to invest, live and work. Bono saw ONE's credibility under threat. It had focused on northern governments (Washington, London, Berlin), whose decisions mattered greatly; but Drummond and activists John Githongo, Ory Okolloh and Rakesh Rajani kept insisting that Africa's future must be decided by Africans. An organization called ONE was only half the people needed; ending extreme poverty without full partnership with the global South was fantasy.
ONE committed to organizational and cultural change, still expanding work with African leaders "grassroots, grass tops, and all in between," and built a growing office in Johannesburg and across the continent. OKRs kept it focused on concrete steps: hiring staff in Africa, expanding the board, reconnecting with old Jubilee partners, finding new advisory networks. ONE, Bono says, became a better listener, and could not have done it without OKRs.
Objective: proactively integrate a broad range of African perspectives into ONE's work, align more closely with African priorities, and share and use ONE's political capital for specific policy changes in and toward Africa.
- Three Africa-based hires complete and onboard by April; two African board members approved by July.
- African Advisory Board in place by July and convened twice by December.
- Relationships fully developed with at least 10 to 15 leading African thinkers who regularly challenge and guide ONE's policy and external work.
- Four participatory trips to Africa during 2010.
Measuring passion
Sudanese businessman and philanthropist Mo Ibrahim, with his daughter Hadeel, gave ONE the African "intellectual static" it lacked. Before joining he was "properly rude" about some of its objectives, and he steered it toward transparency as a central goal in Africa, Europe and America. Research showed corruption drains about $1 trillion a year from developing countries; Ibrahim argued fighting it would save more lives than fighting HIV/AIDS.
- Publish What You Pay. Lobbying alongside this collective, ONE helped win rules that Bono says make it illegal for companies listed in New York or the EU to conceal payments for mining rights (see the critical lens below).
- Dangote. Nigerian industrialist Aliko Dangote, sometimes called Africa's Bill Gates, joined the board.
Then the facts. ONE's 8.9 million members are people who signed up online or took at least one action. Bono imagines Gates rolling his eyes: signers are not members. Gates is right, and it raised the questions of how to measure engagement and whether it could grow. ONE set out to move people from signers to members to activists to catalysts, thanking and rewarding those who acted more than once, and flooding the districts of target legislators. Texas Republican congresswoman Kay Granger probably thinks ONE T-shirts are everywhere; she was a strategic target, and she came through. Nobody had measured activists' passion before, but the question is pure OKR: how passionate are you, and what does your passion make you do? Now when Gates asks hard questions at board meetings, ONE shows its OKRs: what it did and what impact it had.
An OKR framework
The downside Bono sees is getting too organized: ONE must stay disruptive, not institutional, and he fears it will go corporate and chase every quarterly green. Doerr's reminder, counterintuitive for a now well-funded team of high achievers, was that if everything is green you failed, and he kept urging "More red!" ONE is good at big ambitions and less good at the incremental.
ONE does not stand on passion or moral outrage alone but on principles, with walls and floors, a structure of thinking from OKRs. Change takes intellectual rigor and serious strategy. The framework contains the madness and chemistry inside it, and gives a place for risk and trust where failing is not a fireable offense; with that structure and the right people, magic is close. Bono closes where he began: a talented guitarist, an average singer, a weak bassist, a drummer getting there, but goals and a rough idea of how to reach them: they wanted to be the best band in the world.
Bono says US-listed and EU companies must disclose resource payments; in fact Congress repealed the SEC's Dodd-Frank 1504 rule in February 2017, before the book appeared (a weaker rule followed in 2020). Mother-to-child HIV transmission was not ended by 2020. And ONE's results mix its own advocacy with work by governments and many NGOs, so attribution to OKRs is loose.
The Goals to Come
Doerr closes with his own stretch goal: to see Andy Grove's system spread from companies to schools, governments, health care and social causes. His summary is six words: ideas are easy; execution is everything. OKRs are adaptable, with no dogma and no single right way, so each organization must find its own points of emphasis and make the tool its own.
What keeps me going is goals.
What the book has shown
Doerr recaps in one sweep: OKRs and CFRs have helped organizations of every type and size move mountains, inspiring workers, developing leaders and uniting teams. Measuring what matters helps Bono and the Gates Foundation fight poverty and disease in Africa, drives Google's 10x quest to make the world's information freely accessible, and lets Zume deliver a robot-assembled artisanal pizza hot to your door. His verdict: we're just getting started.
| Organization | What OKRs did there |
|---|---|
| Intel | Turned a chip war (Crush) into one company-wide priority |
| Drove 10x stretch goals: Chrome, YouTube | |
| Gates Foundation, ONE | Mobilized against poverty and disease in Africa |
| Zume, Remind, MyFitnessPal | Gave start-ups focus, alignment and discipline |
| Adobe, Lumeris, Intuit | Reset performance management and culture |
| Khan Lab School | Taught children to set their own learning goals |
A launch pad, not a protocol
OKRs can be called a tool, a protocol or a process, but Doerr's preferred image is a launch pad: a liftoff point for the next wave of entrepreneurs and intrapreneurs. His dream is for Grove's invention to change every walk of life, and he believes it could have a large effect on six fields:
- GDP growth.
- Health-care outcomes.
- School success.
- Government performance.
- Business results.
- Social progress.
His glimpse of that future is Orly Friedman, who introduced OKRs to every elementary pupil at the Khan Lab School in Mountain View, California: five- and six-year-olds setting their own objectives and key results while learning to read and reason. If structured goal setting and continuous communication were widely deployed "with rigor and imagination," Doerr predicts exponentially greater productivity and innovation across society.
No dogma
The tool's power, Doerr says, lies in its adaptability. There is no single right way to use it, and an organization's needs shift across its life cycle, so the right first move differs:
- For some, transparency. Simply making goals open and visible is a big leap.
- For others, cadence. A quarterly planning rhythm alone changes the game.
- For all, choice. Readers must pick their own points of emphasis among the superpowers (focus, align, track, stretch) and CFRs, and make the tool their own.
The whole book rests on one premise: good ideas are plentiful, but disciplined execution against a few measurable goals is rare. OKRs exist to close that gap.
The conversation continues
The book, Doerr admits, tells only a handful of behind-the-scenes stories; thousands more organizations are just starting with OKRs and CFRs or have yet to tell their stories. He moves the conversation online to whatmatters.com and invites readers to write to him directly. His ultimate stretch OKR has three parts: help people achieve the seemingly impossible together, build durable cultures for success and significance, and prime the pump of inspiration for the goals that matter most to each reader.
The claim that OKRs could raise GDP or school results is offered as conviction, not evidence; the book presents no failed adoption that wasn't later rescued. Readers should treat the closing as advocacy from the system's most prominent evangelist.
Coach Bill Campbell
Doerr dedicates the book to two men who died within four weeks of each other in 2016: Andy Grove, the inventor of OKRs, and "Coach" Bill Campbell, Silicon Valley's great executive mentor. Grove fills the book; Campbell gets this closing tribute because, through honest, open communication and a zeal for data-driven operating excellence, he embodied the OKR spirit. His motto, "be better every day," and his insistence on the unglamorous committed goals made him the human side of the system.
The biggest tent
Campbell's funeral mass was held on a clear April morning in Atherton, California, in a tent on the Sacred Heart playing fields where he had spent many Saturdays coaching eighth-graders in flag football and softball. More than 3,000 mourners came, from Larry Page and Jeff Bezos to generations of his former young players. He gave everyone bear hugs and selfless advice, and each believed Campbell was their best friend.
The son of a gym teacher who worked nights at the Homestead, Pennsylvania steel mill, he earned the nickname in the 1970s as football coach at his beloved alma mater, Columbia. He became the Coach in Silicon Valley's boardrooms: a world-class listener, hall-of-fame mentor and, to Doerr, the wisest man he had met, whose ambitious, caring, accountable, transparent and profane humanity shaped Google's culture and dozens of others. Ken Auletta wrote in The New Yorker that in the engineering capital Campbell taught founders to look up from their screens, and deserved front-page obituaries.
1970s Head football coach at Columbia.
Late 1980s Runs Apple's software subsidiary Claris; when Apple and John Sculley refuse the IPO spin-off he believed was promised, Doerr recruits him as CEO of pen-computing start-up GO Corporation.
1994 Joins Kleiner Perkins as executive in residence; one walk with Scott Cook makes him Intuit CEO for four years. Doerr also recruits him to Netscape's board.
1997 Steve Jobs returns to Apple, asks every director but one to resign, and invites Campbell onto the new board; he refuses pay.
2001 Becomes coach to new Google CEO Eric Schmidt, then to the whole executive team.
2016 Dies; Doerr signs this tribute in April 2018.
GO: no more voting
They met in the late 1980s when Doerr was hiring a CEO for GO, a pen-based tablet company and one of his most famous failures (Campbell's joke: call it "GO, Going, Gone"). Campbell came recommended by top recruiter Debra Radabaugh and by his former Apple marketing boss Floyd Kvamme, whom Doerr had brought to Kleiner. A visit to Campbell's team at Claris sealed it: its esprit de corps and evident esteem for him sold Doerr on the spot.
The business model failed, but Campbell changed how GO decided things.
- Before. Every executive vote was preceded by a heated strategy fight that left winners, losers and hard feelings.
- After. Campbell sat with each executive, asked about their families, told a story or two, and learned where they stood; he built agreement before the meeting, and soon GO stopped voting. He had no private agenda; the team and mission came first.
- Leaders who made leaders. Five of his GO reports became CEOs or business heads of their own ventures, all backed by Doerr and all profitable.
- Dignity in failure. When GO sold to AT&T, they made sure laid-off staff had strong references and good new jobs.
Intuit: go with the Coach
Back at Kleiner in 1994, Campbell sat in the corner office beside Doerr, who promised him another company. Intuit founder Scott Cook wanted a CEO; after one walk around Campbell's Palo Alto neighbourhood he had the job, and the two built a spectacular company.
Early in his four-year Intuit tenure revenue lagged and the quarter would be missed. A blue-sky board meeting in a Las Vegas hotel suite wanted to invest more and spend through the shortfall. Campbell refused: Intuit would cut back, lay people off, get leaner and make the numbers, because that was the discipline and culture he wanted, for shareholders, team and customers. As the poll went against him he looked more and more distraught. Doerr, unsure whether he was right, said go with the Coach because it was his call. The vote turned; Campbell later said he might have resigned otherwise.
From then they had an unbreakable bond: they could say harsh things, and the next day one would call to apologize, because loyalty to the relationship and the team outweighed any difference. Their routine became: Kleiner invests, Doerr sponsors, Doerr calls Campbell, Campbell coaches the team.
Apple and Google
When Jobs returned to Apple in 1997, in what Doerr calls the most amazing non-hostile takeover of a public company, achieved without a penny, Campbell joined the board. He took no pay, seeing it as giving back to the Valley; when some companies insisted he accept stock, he passed the proceeds to his philanthropy.
In 2001, after Doerr helped persuade Google's founders to hire Eric Schmidt, he told Schmidt he needed Campbell as a coach. Schmidt, already a former CEO and chairman of Novell, was offended ("I know what I'm doing"), but within a year his self-review said Campbell's role had been needed from the beginning and he should have set it up from his first day.
- Open mandate. He coached Page, Sergey Brin, Susan Wojcicki, Sheryl Sandberg, Jonathan Rosenberg and the whole team, in a style "one part Zen and one part Bud Light."
- Listen, then decide. Little direction, a few right questions, mostly listening. There were usually several right answers and the leader's job was to pick one: just make a decision; are you moving forward, breaking ties, rolling on?
- Committed OKRs first. He watched the less glamorous committed objectives: you have to make the trains run on time (in saltier words). Sundar Pichai: he cared about operating excellence day in, day out, under the motto "Be better every day."
- People over process. He never big-footed engineers (he knew more tech than he admitted); his insight was into what makes teams tick and how to stop people being steamrolled by process. Seeing someone treated unfairly, he called the CEO, called it "a process error" and fixed it.
- Love plus fearless feedback. Faces lit up when he entered Intuit meetings; affection often came as faux insult (an ugly sweater drew a crack about mugging someone for it), but people knew he had their back: a tough coach and always a players' coach.
- Family first. Happiest coaching his daughter Maggie (and Doerr's daughter Mary) in softball, on the field at 3:20 sharp whatever meeting was starting elsewhere, never checking his phone in the sixth inning.
Campbell was the éminence grise at Google's Monday staff meetings, in effect its unofficial chairman, and also lead outside director at Apple. After Android challenged the iPhone, Jobs badgered him to choose. Campbell said he could not "even spell HTML" and was only helping Google be a better business, then warned Jobs not to make him choose because he would not like the answer. Jobs backed down; Campbell was his one true confidant, and Schmidt told Forbes he "kept Steve Jobs going."
The coaching goes on
Even ill, Campbell kept coaching. His advice weighed heavily when Doerr took the Kleiner Perkins chairmanship, timed as his two daughters left for college: Campbell knew Doerr would not slow down or move "upstairs" but would use the role to do more of what he loved, finding and funding the best entrepreneurs and helping them build teams, a player-coach for the next generation. Months before his death, on a podcast with Doerr's partner Randy Komisar, Campbell said he always wanted to be part of the solution, that people are the most important thing and the job is to make them better. His hundreds of disciples, Doerr writes, are still trying to get better every day.
Sitting on Apple's board while advising Google's leadership through the Android years was a genuine governance conflict, not just something that "drove Steve Jobs crazy"; the tribute treats it as a charm story. Campbell's own coaching methods are set out at length in Trillion Dollar Coach (Schmidt, Rosenberg and Eagle, 2019).
Resource 1: Google's OKR Playbook
Resource 1 reprints, with Google's permission, excerpts from the company's internal OKR guidelines and templates, issued periodically as it scaled; nobody has more collective OKR experience (Google itself says your approach may and should differ). The premise: a team of a few percent of Google's workforce, acting together toward one ambitious goal, has repeatedly remade a mature industry in under two years (Search, Chrome, Android). Poorly managed OKRs are an empty management gesture; well-done OKRs make clear what is important, what to optimize and which trade-offs to make every day.
Purpose and scoring
Google likes to think big, and OKRs are how it communicates, measures and reaches those goals. Because a small team can transform an industry, every employee and manager must make conscious, careful, informed choices about time and energy; OKRs are those choices made visible, and the means of coordinating individuals toward great collective goals. They are used to:
- Plan what people will produce.
- Track progress against the plan.
- Coordinate priorities and milestones between people and teams.
- Focus people on the most important goals rather than urgent but less important ones.
OKRs are big, not incremental, and Google does not expect to hit them all; if it does, they were not aggressive enough. Grading uses a colour scale:
- 0.0 to 0.3 = red.
- 0.4 to 0.6 = yellow.
- 0.7 to 1.0 = green.
The colour bands apply to the whole set, but the playbook's real discipline is the split below: a committed OKR is a promise graded against 1.0, an aspirational one a bet graded against 0.7. Most of the six traps come from blurring the two.
Writing effective OKRs
Good OKRs are hard but not impossible to write, if you follow a few simple rules.
Objectives are the "whats." They express goals and intents; are aggressive yet realistic; are tangible, objective and unambiguous, so a rational observer can tell whether they were achieved; and their achievement must give clear value to Google.
Key results are the "hows." They are measurable milestones that, if achieved, advance the objective usefully for its constituents; describe outcomes, not activities (words like "consult," "help," "analyze," "participate" signal activities); and include evidence of completion that is available, credible and easy to find: change lists, doc links, notes, published metrics reports.
The outcome rule in practice: not "assess Colossus latency" but "publish average and tail latency measurements from six Colossus cells by March 7."
Cross-team OKRs
Many important projects need several groups, and OKRs suit that coordination. A cross-team OKR should include every group that must materially contribute, and each group's commitment should appear explicitly in its own OKRs. If Ads Development, Ads SRE and Network Deployment must all deliver for a new ads service, all three carry OKRs describing their part.
- Committed OKRs. Agreed to be achieved; Google will adjust schedules and resources to deliver them. Expected score 1.0; anything less requires an explanation, since it shows errors in planning or execution.
- Aspirational OKRs. Express how we would like the world to look, with no clear idea how to get there and/or without the resources to deliver. Expected average score 0.7, with high variance.
Six classic traps
| Trap | What goes wrong | Litmus test or fix |
|---|---|---|
| 1. Mixing committed and aspirational | Committed marked aspirational raises failure risk: teams don't take it seriously or reorder priorities. Aspirational marked committed breeds defensiveness and priority inversion, as committed work is de-staffed for it | Label every OKR deliberately |
| 2. Business as usual | Written from what the team can do without changing anything, not what it or its customers really want | Ask what the team and customers really want |
| 3. Timid aspirations | Start from today and ask what extra staff and a bit of luck would buy | Start from the world in several years freed of most constraints; you won't know how at first (that is what aspirational means), but without stating the end state you guarantee missing it. Test: does it meet or exceed what customers say they want? |
| 4. Sandbagging | Team hits everything without using all headcount or capital: hoarding resources, not pushing, or both | Committed should credibly use most, not all resources; committed + aspirational somewhat more than all (else they are really commits). Otherwise management reassigns resources |
| 5. Low-value ("Who cares?") | A Low Value Objective: even a 1.0 that nobody notices, e.g. raise task CPU use 3% | Could it score 1.0 without user or economic benefit? Reword: cut cores for peak queries 3% at equal quality and latency and return them to the free pool. Not "Launch X" but "double fleet-wide Y by launching X to 90%+ of borg cells" |
| 6. Insufficient KRs for committed Os | KRs necessary but not sufficient, tempting because they dodge hard resource, priority and risk commitments; especially pernicious because they delay discovering resource needs and slippage | KRs at 1.0 must yield an objective at 1.0. Could all KRs score 1.0 while the objective's intent fails? Then add or rework KRs |
Reading, interpreting and acting on OKRs
Committed OKRs
- Rearrange priorities. Teams reshuffle other work to deliver an on-schedule 1.0.
- Escalate early. A team that cannot credibly promise 1.0 must escalate promptly, whether the cause is disagreement about the OKR, about its priority, or lack of time, people or resources. In this common situation escalation is required, not merely OK: it lets management develop options and resolve conflicts.
- Escalation is normal. Nearly every new OKR involves some escalation, because it changes existing priorities and commitments; one that changes nobody's activities is business as usual, unlikely to be new even if never written down before.
- Postmortem every miss. A committed OKR below 1.0 at its due date gets a postmortem, not to punish but to understand what happened in planning or execution and improve the team's ability to reliably hit 1.0.
- Typical examples. Meeting a service's SLA for the quarter; delivering a defined infrastructure feature or improvement by a set date; manufacturing and delivering a quantity of servers at a cost point.
Aspirational OKRs
- Deliberately too many. The set exceeds a quarter's capacity by design; priority decides where remaining time goes after commitments are met, higher priorities first.
- Carry them forward. Keep them on the list, quarter to quarter, until done; dropping them for lack of progress disguises persistent problems of prioritization, resources or understanding. Moving one to a team with more expertise and bandwidth is good.
- Ask for resources. Managers estimate what aspirational OKRs need and request it each quarter, fulfilling their duty to express known demand, without expecting full funding unless the OKR is the company's top priority after commitments.
- Five minutes? OKRs written in five minutes probably aren't good. Think.
- One line. An objective that doesn't fit on one line isn't crisp enough.
- Impact, not internal jargon. Not "Launch Foo 4.1" (what matters is its impact) but "launch Foo 4.1 to improve sign-ups by 25 percent," or simply "improve sign-ups by 25 percent."
- Real dates. If every KR lands on the quarter's last day, there is probably no real plan.
- Measurable. It must be possible to grade objectively at quarter end: not "improve sign-ups" but "improve daily sign-ups by 25 percent by May 1."
- Unambiguous metrics. Is "1 million users" all-time users or seven-day actives?
- Full coverage. If important activities, or a big share of the team's effort, are not covered by OKRs, add more.
- Hierarchy. Larger groups need high-level team-wide OKRs and more detailed subteam ones, with horizontal projects (needing several teams) supported by KRs in each subteam.
Resources 2-4: The OKR Cycle, Performance Conversations, In Sum
The last three resources turn the book into working tools: a typical quarterly OKR cycle across company, team and contributor levels; a bank of questions for performance conversations; and In Sum, Doerr's checklist of the whole method. Continuous performance management is two interwoven parts: setting OKRs, and holding regular conversations tailored to each person's needs.
Resource 2: a typical OKR cycle
Assuming three levels (larger companies may add more), the cycle for Q1 runs:
- 4-6 weeks before. Senior leaders brainstorm company OKRs; for Q1 this is also when the annual plan is set.
- 2 weeks before. Finalize and communicate company OKRs for the year and Q1 to everyone.
- Start of quarter. Teams derive team OKRs from the company's and share them at their meetings.
- 1 week after start. Contributors share individual OKRs, often negotiated with managers in one-on-ones.
- Throughout. Employees track progress and check in with managers; periodically they assess the odds of full attainment and recalibrate if needed.
- Near quarter end. Contributors score their OKRs, self-assess and reflect.
The chart shows that planning never stops: brainstorming for Q2 begins in mid-February, while Q1 is only half run, and Q1 is scored as Q2 goals are being finalized. A company running OKRs is always executing one quarter and designing the next.
Resource 3: performance conversations
Goal planning and reflection
- What OKRs will you focus on to drive the greatest value for your role, team or company?
- Which of these OKRs align to the organization's key initiatives?
Progress updates
- How are your OKRs coming along?
- What critical capabilities do you need to succeed?
- Is anything stopping you from attaining your objectives?
- Which OKRs need to be adjusted, added or eliminated given shifting priorities?
Manager-led coaching
Before the conversation, the manager asks:
- What behaviors or values do I want my report to continue exhibiting?
- What behaviors or values do I want them to start or stop?
- What coaching can I give to help them realize their potential?
During it: what part of your job most excites you? What aspect of your role would you like to change?
Upward feedback
- What are you getting from me that you find helpful?
- What are you getting from me that impedes your effectiveness?
- What could I do to help you be more successful?
Career growth
- What skills would you like to develop to improve in your current role?
- In what areas do you want to grow to reach your career goals?
- What skills would you like to develop for a future role?
- From a learning and development standpoint, how can I and the company help you get there?
The leader considers:
- The contributor's main objectives and responsibilities in the period, and how they performed.
- If underperforming, how to course-correct; if performing well, how to sustain it without burnout.
- When they are most and least engaged; what strengths they bring; what learning experiences would help.
- Their focus for the next six months: meeting expectations, maximizing the current role, or preparing for the next opportunity (project, responsibility or role).
The contributor asks:
- Am I on track to meet my objectives?
- Have I identified areas of opportunity?
- Do I understand how my work connects to broader milestones?
- What feedback can I give my manager?
Resource 4: In Sum
Doerr's closing checklist restates the four superpowers (focus and commit, align and connect, track, stretch), continuous performance management and culture as working rules.
| Area | Core rule in one line |
|---|---|
| Focus and commit | 3-5 objectives, up to 5 KRs each, quarterly and annual in parallel, leaders first |
| Align and connect | Public goals to the CEO, about half bottom-up, horizontal OKRs across silos |
| Track | Honest grading from the top, weekly check-ins, revise mid-cycle when needed |
| Stretch | Label committed vs aspirational, aim for 10x, fail without judgment, roll over |
| Continuous performance management | Divorce OKRs from pay, CFRs alongside OKRs, contributor sets the agenda |
| Culture | Align OKRs to values, act by deed, fix trust before rollout |
Focus and commit to priorities
- Dual-track cadence. Run quarterly OKRs for short-term goals and annual OKRs for longer-term strategy in parallel.
- Top first. Phase the rollout in with upper management to work out kinks and harden leaders' commitment; let momentum build before individual contributors join.
- A shepherd. Name an OKR shepherd who makes sure everyone spends time each cycle choosing what matters most.
- Three to five objectives. More dilutes and scatters effort; expand capacity by deciding what not to do, then discard, defer or de-emphasize.
- Leverage. Pick the objectives with the most leverage for outstanding performance, mining the mission, strategic plan or a leadership theme for top-line OKRs.
- Elevate for support. To stress a department goal and enlist lateral help, raise it to a company OKR.
- Up to five KRs. Each objective gets no more than five measurable, unambiguous, time-bound key results; completing them all, by definition, achieves the objective.
- Pairing. Pair qualitative and quantitative KRs for balance and quality.
- Promote a KR. A key result that needs extra attention can become an objective for a cycle or more.
- Leaders' conviction. The single most important factor is buy-in by the organization's leaders.
Align and connect for teamwork
- Line of sight. Show people how their objectives relate to the leader's vision and top priorities; the route to excellence is lined with public goals, up to the CEO.
- Repeat the why. Explain at all-hands why an OKR matters, then repeat it until you are tired of hearing it.
- Give-and-take. With cascaded objectives from the top, welcome frontline input on key results: innovation lives at the edges more than the centre.
- About half bottom-up. Encourage a healthy share, roughly 50 percent, of bottom-up OKRs.
- Smash silos. Connect teams with horizontally shared OKRs; cross-functional work enables fast, coordinated decisions and competitive advantage.
- Explicit dependencies. Make every lateral dependency explicit, and tell all stakeholders when an OKR is revised or dropped.
Track for accountability
- Honest grading from the top. Continuous reassessment and objective grading build accountability; when leaders admit missteps, others feel freer to take healthy risks.
- Visible achievement. Motivate less with extrinsic rewards and more with open, tangible measures of achievement.
- Check-ins. The shepherd rides herd on regular check-ins and updates, so teams can course-correct or fail fast.
- Rhythm. Encourage weekly one-on-one OKR meetings and monthly departmental meetings.
- Not in stone. Revise, add or delete OKRs as conditions change, even mid-cycle; clinging to irrelevant or unattainable goals is counterproductive.
- Close the cycle. Use grades plus self-assessment to evaluate, celebrate and improve, and pause to savor the last cycle before the next.
- Software. Invest in a dedicated, automated cloud platform; public, collaborative, real-time systems work best.
Stretch for amazing
- Label up front. Each cycle, separate committed OKRs (100 percent) from aspirational BHAGs.
- Safe to fail. Let people fail without judgment.
- Ambitious, not absurd. Set goals high enough that some quarterly targets will be missed, but not obviously unrealistic; morale suffers when success is impossible.
- Gospel of 10x. Replace incremental OKRs with exponential ones for leaps in productivity and innovation; that is how industries are disrupted.
- Fit the culture. The optimal stretch varies by organization and over time with each cycle's needs.
- Roll over. A missed stretch OKR can carry to the next cycle if still relevant.
Continuous performance management
- Continuous, not annual. Catch issues before they become problems and support strugglers in time.
- Divorce OKRs from bonuses. Separate forward-looking OKRs from backward-looking reviews; tying attainment to pay invites sandbagging and risk aversion.
- Better criteria. Replace ratings and stack rankings with transparent, strength-based, multidimensional criteria, including team play, communication and ambition in goal setting.
- Intrinsic motivation. Purposeful work and growth beat financial incentives.
- CFRs with OKRs. Transparent OKRs make coaching concrete; continuous CFRs keep daily work on point and collaborative.
- Contributor's agenda. In performance conversations the contributor sets the agenda; the manager learns and coaches.
- Feedback everywhere. Make feedback two-way, ad hoc and multidirectional, unconstrained by the org chart; use anonymous pulse surveys for real-time reads on operations or morale; pair peer feedback with cross-functional OKRs.
- Recognition. Peer recognition should be frequent, specific, highly visible and tied to top-line OKRs.
The importance of culture
- Mission alignment. Align top-line OKRs with mission, vision and North Star values.
- Deeds over words. Convey values by word, but mostly by deed.
- Collective goals, individual KRs. Make OKRs collective but assign key results to individuals and hold them accountable.
- Catalysts and nourishers. Balance OKR catalysts (actions that support the work) with CFR nourishers (interpersonal support, even random kindness).
- Division of labour. OKRs bring transparency, clarity, purpose and big-picture orientation; CFRs build positivity, enthusiasm, stretch thinking and daily improvement.
- Fix trust first. Address cultural barriers, especially accountability and trust, before implementing OKRs.
Of every item in the checklist, Doerr singles out one as the most important for success: conviction and buy-in by the organization's leaders.
The rules pull against each other in places Doerr leaves the reader to resolve. "About half bottom-up" sits awkwardly beside the top-down first cycles praised at Intel, Google and Zume; "revise even mid-cycle" can erode the commitment that committed OKRs depend on; and "divorce OKRs from pay" coexists with Google counting them as up to a third of a rating. Treat the list as defaults, not law.
Master Takeaways
Doerr's book reduces to one loop: choose a few objectives, define measurable key results, make them public, check in often, grade honestly, and reflect before the next cycle, with conversations and recognition keeping people engaged. OKRs work when leaders commit to them openly and keep them separate from pay; they fail as private, top-down paperwork.
- Execution beats ideas. Good ideas are plentiful; disciplined delivery against a few measurable goals is rare, and OKRs exist to close that gap.
- What plus how. An objective states what you want; key results state how you will know, with a number and a date, verifiable without argument.
- Fewer is stronger. Three to five objectives, each with five or fewer key results; every commitment forfeits another, so say no as well as yes.
- Keep OKRs off the paycheck. Tie scores to bonuses and people sandbag; Google counts OKRs as a third or less of ratings and wipes scores each cycle.
- Two kinds of goal. Committed OKRs must hit 1.0; aspirational OKRs average 0.7, and a team that hits everything aimed too low.
- Leaders decide adoption. Executive conviction and public commitment matter more than any software; without them OKRs become window dressing.
- Hard goals work. Locke's 1968 finding, confirmed in 1,000+ studies, is that specific, hard goals beat easy or vague ones, provided people believe they are attainable.
- Pair key results. Balance quantity with quality, or effect with counter-effect, so the measure cannot be gamed (the Pinto and Wells Fargo warnings).
- Make goals public. Transparency up to the CEO exposes duplication, links work to the mission and lets good ideas surface anywhere.
- Half from the bottom. Rote cascading costs agility, flexibility and frontline voice; roughly half of OKRs should be written by those doing the work.
- Name every dependency. MyFitnessPal's slipped email OKR and Under Armour off-site show unspoken cross-team dependencies, not weak goals, derail execution; give every OKR one owner.
- Leaders need their own goals. Intuit's CIO learned that if his goals simply equalled the department's eighteen-month objectives, staff asked what he was doing.
- Give goals a home. Shared, searchable software makes transparency real; 328,000 Word files a year would not.
- Track weekly, grade honestly. Check in, then continue, update, start or stop; at the end average the KRs on 0.0 to 1.0, add self-assessment and reflect.
- Goals can change. Dropping an irrelevant OKR mid-cycle, as Remind did with its failed payments product, is discipline, provided everyone affected is told.
- A mission is not an objective. Gates's rule: missions give direction; objectives are concrete, measurable steps you are actually pursuing, and you must know how to scale and measure them.
- Measure the right thing. The Gates yam seed doubled yields but nobody would plant it; YouTube's switch from views to watch time mattered more than any target set on top.
- 10x forces reinvention. Gmail's gigabyte and Crush's 2,000 design wins show that exponential targets change the method, not just the effort.
- A miss can be a plan. Chrome missed 20 million and 50 million; each miss produced a new lever (distribution, TV, Mac and Linux, re-engagement) until it hit 111 million.
- Frame the stretch. YouTube's billion hours was under 20 percent of world TV viewing, which made a 10x goal believable.
- Keep the goal through a handover. YouTube made its target because a new CEO kept one stretch objective rather than diluting it with rival priorities.
- CFRs give OKRs a voice. Conversations, feedback and recognition replace the annual review; the one-on-one belongs to the employee, built on five standing questions.
- The annual review fails on its own terms. 7.5 manager hours per report, 6 percent of HR leaders think it worth it; Adobe's version cost 80,000 hours a year.
- Feedback must be specific and two-way. Name the behaviour and its effect, invite employee-to-manager and peer feedback, and route 360 input by purpose.
- Recognition is underrated. Peer recognition that is frequent, specific and tied to goals raises engagement; high-recognition firms show 31 percent lower turnover.
- OKRs train managers. Zume's founders used them to teach staff that managers are paid for decisions, not volume, before the company scaled.
- Fix culture before tools. Where trust and accountability are missing, OKRs get rejected like a donor organ; clean up leadership and HR first, as Lumeris did.
- Sell your reds. Lumeris reviews only green or red, spends the time on at-risk goals, and colleagues volunteer to "buy" them: transparency without judgment.
- Catalysts plus nourishers. Amabile's motivation research pairs work support (OKRs) with interpersonal support (CFRs); Aristotle's top team factor is structure and clarity.
- Measure behaviour to scale values. Seidman's trust index counts information flows and how skipping your boss is treated, not how people say they feel.
- Start small, be patient. Pilot with leaders or one department; expect four or five quarters, and up to a year of resistance, before the system takes hold.
- Coach the committed goals. Bill Campbell's lesson: stretch is glamorous, but someone must make the trains run on time and get better every day.
Frameworks & mental models
| Name | What it says | Use it when |
|---|---|---|
| OKR | Objective (what) plus 3-5 measurable key results (how) | Setting any quarterly or annual goal |
| Grove's hygiene (7 rules) | Less is more; bottom-up; no dictating; flexible; dare to fail; tool not weapon; patience | Designing or auditing an OKR program |
| Four superpowers | Focus, align, track, stretch | Explaining why OKRs work, diagnosing weak spots |
| Paired key results | Each quantity measure has a quality or counter-effect check | Preventing gaming and side effects |
| Four costs of cascading | Lost agility, lost flexibility, marginalized contributors, one-way links | Deciding how much to cascade |
| Continue / update / start / stop | Four options at any check-in | Mid-cycle reviews |
| Google 0-1 scoring | Average of KRs; 0.7-1.0 green, 0.4-0.6 yellow, 0.0-0.3 red | Grading at cycle end |
| Committed vs aspirational | 1.0 expected vs 0.7 average with high variance | Labeling goals and setting expectations |
| Google's six traps | Mixed types, business as usual, timidity, sandbagging, low value, insufficient KRs | Reviewing a draft OKR set |
| Stretch variables | Margin for error, morale, all-committed risk, no magic number | Calibrating ambition to the culture |
| Big rocks (Covey) | Put the most important things in the jar first | Cutting a long goal list |
| CFRs | Conversations, feedback, recognition | Replacing annual reviews |
| Five conversation questions | Working on? OKRs? Blockers? Needs from me? Growth? | Every manager one-on-one |
| Check-in (Adobe) | Goals, feedback, growth; no ratings; pay set separately by managers | Abolishing the annual review |
| Five ways to recognize | Peer-to-peer, clear criteria, stories, frequent, tied to strategy | Building a recognition habit |
| Garden's one-on-one | Happy? Drains? Dream job? Truth, right thing, three-year promise | Non-work coaching conversations |
| Project Aristotle | Clarity, safety, meaning, dependability, impact | Diagnosing team health |
| Catalysts and nourishers | Work support (OKRs) plus interpersonal support (CFRs) | Building a high-motivation culture |
| Selling your reds | Green or red only; peers "buy" at-risk goals | Running leadership OKR reviews |
| OKR cycle (Resource 2) | Brainstorm 4-6 weeks out, publish 2 weeks out, teams at start, people a week in | Planning the quarterly calendar |
Who's who
| Name | Role | Why they matter |
|---|---|---|
| Andy Grove | Intel CEO | Invented OKRs; seven hygiene rules |
| John Doerr | Kleiner Perkins investor | Carried OKRs to Google and 50+ firms |
| Larry Page, Sergey Brin | Google cofounders | Adopted OKRs in 1999; 10x thinking |
| Bill Davidow | Intel marketing chief | Led Operation Crush |
| Sundar Pichai | Google CEO | Chrome's stretch OKRs |
| Susan Wojcicki | YouTube CEO | Kept the billion-hour goal |
| Cristos Goodrow | YouTube engineering VP | Argued for watch time; guarded the OKR |
| Bill and Melinda Gates, Patty Stonesifer | Gates Foundation | Tracked giving with OKRs |
| Donna Morris | Adobe HR head | Replaced annual reviews with Check-in |
| Julia Collins, Alex Garden | Zume co-CEOs | OKRs as executive training |
| Andrew Cole, Art Glasgow | Lumeris HR chief, COO | Culture first; "selling your reds" |
| Bono | ONE cofounder | OKRs for activism and Africa pivot |
| Brett Kopf | Remind cofounder | One metric, focus amid ADHD |
| Jini Kim | Nuna cofounder | OKRs fail without leaders' commitment |
| Mike Lee | MyFitnessPal cofounder | Exposed hidden dependencies |
| Atticus Tysen | Intuit CIO | Transparent OKRs connect IT teams |
| Shishir Mehrotra | YouTube tech lead | Big rocks; set the billion-hour goal |
| Dov Seidman | LRN founder | Out-behave to outperform; trust index |
| Bill Campbell | Executive coach | "Better every day"; committed goals |
- Evangelist and investor. Doerr wrote from inside: Google, Zume, BetterWorks and others are Kleiner-backed (his BetterWorks stake goes undisclosed in the tracking chapter), and contributors are admirers. Later setbacks postdate the book: Zume's 2023 shutdown, Under Armour's 2020 sale of MyFitnessPal at a loss.
- Survivorship. Every story ends well; there is no sample of firms that adopted OKRs and stalled, so the causal credit given to OKRs is unproven.
- Internal tensions. The book says OKRs must stay off the paycheck yet celebrates Crush's Tahiti trip; it praises stretch while admitting hard grading suits few cultures.
- Metric risk underplayed. YouTube's watch-time optimization later drew heavy criticism for amplifying extreme content, the Goodhart problem Doerr's paired-KR rule only partly answers.
- Execution, not strategy. OKRs organize a chosen strategy; they do not choose it. Grove's PC misjudgment and Zume's business model show good goal hygiene cannot rescue a wrong bet.
- Thin evidence. Survey figures (Deloitte, Gallup, LRN) are cited without methods, and several factual slips (Moore's law, AdSense's launch year, disclosure rules) go uncorrected.
Connect it
- AI teams. Model work invites vanity metrics (benchmark scores, tokens served); Google's "outcomes, not activities" rule and paired KRs (accuracy with cost or harm rates) fit directly.
- Enterprise architecture. Cross-team OKRs, where every contributing team carries its own commitment, mirror dependency mapping in architecture roadmaps; Intuit's IT pilot is the template.
- Banking. Wells Fargo's fake-accounts scandal is the book's own warning: sales-volume targets without counter-measures and with pay attached produce exactly the "goals gone wild" failure.
About this book
Measure What Matters: How Google, Bono, and the Gates Foundation Rock the World with OKRs by John Doerr, 2018. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions