What Are OKRs?

What OKRs actually mean

Open any modern product-team handbook and you will find a section on OKRs within the first twenty pages. The structure is always the same. An Objective is a short sentence stating what you want to accomplish in a fixed window, usually a quarter. Below it sit three to five Key Results, each numeric, time-bound, and pass-or-fail. The OKR either says "ship mobile checkout in Q2" or it says "get three thousand new paid subscribers by June 30." It does not say "improve user experience" or "grow the business." Those are wishes, not OKRs.

A well-known example comes from Google's early years. Larry Page's first Google OKR was, roughly, to build the world's best search engine, with Key Results tied to query speed, crawl size, and index freshness. There was no "try our best" in the Key Results. A number was either hit or missed at quarter-end.

That is the whole technique in one paragraph. An objective states intent. Key Results define what would count as done. Grade it publicly at the end of the period. Set the next one.

It sounds almost too simple to have been a significant contribution. That is itself a feature, not a bug, and a sign that the idea has survived long enough to be polished down to its useful shape. The more interesting question is what Grove, or Drucker before him, was reacting against when they formulated it.

Why it matters in practice

Most teams do not suffer from a shortage of goals. They suffer from a surplus of vaguely-phrased ones. "Become more customer-centric" is an ambition, not a goal. You cannot tell at the end of Q2 whether it was achieved, because the thing being measured was never defined. OKRs force a much harder conversation up front: if we are serious about customer-centricity, what number would we see move?

The shift is from intent to evidence. A good OKR is one where reasonable people can disagree about the objective but cannot disagree about whether the Key Results were hit.

Two short examples make this concrete. At Spotify in its early growth years, OKRs were used to sequence country expansion: one quarterly objective on launching outside Sweden, with Key Results tied to specific country launches and free-to-paid conversion ratios in each. At Intuit around 2005, an OKR drove TurboTax from CD-ROM to web, with Key Results pinned to online-completion rates rather than revenue. Revenue numbers would have been satisfied by promoting harder on CD-ROM for another year. The Key Result named the thing that actually had to change.

One warning. OKRs are not a substitute for strategy. They assume you already know what matters. If the objective is wrong, a beautifully graded 100% attainment on the Key Results just accelerates you in the wrong direction. Grove himself said the framework is only as good as the thinking that precedes it.

Now, the interesting bit.

OKRs are 45 years older than you think

Intel, Google, Doerr. That is how OKRs reached the product-management world. It is not how the idea reached the world.

The earliest clear version is Peter Drucker's 1954 book The Practice of Management, where he introduced Management by Objectives. Drucker's argument was almost the OKR argument exactly. Managers, he wrote, fall into an "activity trap," where the busyness of present work drowns out the question of what the work is for. His fix was to have every manager, individually, write down the specific objectives they were pursuing for the next period, and have those objectives roll up to company objectives. Measurable, written, reviewed. The 1954 text uses the word "measurable" repeatedly. It is not a precursor of OKRs; it is the thing.

The other root is older and less celebrated. In 1926 Mary Parker Follett published "The Giving of Orders," where she argued that a command works only when worker and manager agree on the situation that calls for it. The "law of the situation" becomes the shared authority rather than either person. That is structurally the alignment logic sitting under Drucker's MBO thirty years later, which is sitting under Grove's OKRs fifty years later still.

Andy Grove joined Intel in 1968 and began running a version of Drucker's MBO on the factory floor almost immediately. By 1975 he was teaching it to new hires under the name iMBO, Intel Management by Objectives. One of those new hires was a 23-year-old sales engineer named John Doerr, who sat through the class and kept the notes. Grove eventually wrote the full method down in 1983 in High Output Management, still the best one-volume presentation of the technique. The book rarely uses the acronym OKR. The structure is there, the ritual is there, the compensation warning is there. The rebrand is Doerr's.

Twenty-four years after that Intel class, Doerr walked into Google's office with a PowerPoint deck. The slides still exist online and are worth looking at: thirteen pages, mostly examples from Grove's Intel. Larry Page and Sergey Brin adopted the system the following quarter. Google scaled, OKRs became "the Google method," and in 2018 Doerr published Measure What Matters, which canonised the origin story around Intel and the ping-pong table.

What Grove genuinely added to Drucker was real. Quarterly rhythm rather than annual. Public visibility rather than one-to-one. A deliberate 70% attainment target, so people would set stretch goals without being punished for missing them. And the rule that gets broken most often today: OKRs are not linked to compensation, because the moment they are, people stop setting ambitious ones. None of that is in the 1954 text. All of it is in the 1983 one. But the theory of goals-as-measurable-intent is already fully formed in the first.

Why the history is interesting

The point is not that Grove owes Drucker a royalty, or that Doerr plagiarised Grove. The point is that frameworks in management are almost never inventions. They are re-expressions of older ideas, packaged for a new audience by someone with distribution. Doerr's distribution was Silicon Valley capital and a bestselling book. Grove's was Intel's headcount and a 1983 paperback. Drucker's was a 1954 book that redefined the word "manager." Each re-expression was genuinely useful to its audience. None was a theoretical leap over the prior one.

Three practical implications follow.

First, if you want to learn OKRs properly, do not begin with Measure What Matters. Begin with Drucker's 1954 book, then read High Output Management, then skim Doerr. Going in reverse order takes a weekend and saves you from three hundred blog posts written by people who only read Doerr. You learn where the thinking is load-bearing versus where it is ceremony.

Second, a useful test when you see a new management framework with a founder-CEO's name on it: search for the same structure thirty years earlier. The Lean Startup of 2011 is the Toyota Production System of the 1950s, which is Frederick Taylor's scientific management of 1911. Design Thinking is problem-structuring from 1969-era IDEO plus a pretty typeface. The acronym rebrand is marketing. The underlying idea is almost always in a yellowed paperback someone's father owned.

Third, and this is where it gets personal. I have sat in interview rooms where a candidate, asked about goal-setting, recited the four superpowers of OKRs verbatim from chapter two of Doerr's book. They did not know what Grove's compensation rule was, and did not know Drucker had the framework before Grove did. That is the difference between having read a book and having read the literature. In a serious interview, the gap is visible within thirty seconds.

The neatest compression of all of this is a line attributed to Drucker himself: "what gets measured gets managed." Doerr put a version of it on the cover of his 2018 book. The original shape of the idea sits inside Drucker's 1954 text, in the chapter where MBO is introduced. OKRs are that sentence, at company scale, rediscovered every generation.

One last thing

Understanding where a framework comes from is the difference between using OKRs in an interview and performing them. Rehearsal's briefs work that muscle: fifteen minutes on one concept, taken seriously. That is all.