VI — Contemporary (2000 – today) · Chapter 28
OKRs and the Goal-Setting Trap
When the metric eats the mission
Objectives and Key Results, the goal-setting framework now usually abbreviated as OKRs, was developed by Andy Grove at Intel in the 1970s, taught to John Doerr in the 1980s, transmitted by Doerr to a young Google in 1999, and propagated from Google through the technology industry over the next two decades. The framework is now corporate boilerplate. Most companies use it; most have not read Grove's original treatment in High Output Management; many have implemented variants that contradict Grove's principles in important ways. The result is one of the most widely adopted and most widely misused management practices of the last forty years. The OKR backlash that has gathered force in the 2020s is not, on close inspection, a backlash against OKRs as Grove conceived them. It is a backlash against OKRs as most organizations actually run them.
What Grove Actually Wrote
Andy Grove's High Output Management (1983) describes OKRs in a single short chapter as one tool among several for what Grove called 'managerial leverage.' His treatment is notably modest. The objective is the qualitative direction the team is trying to move in. The key results are three to five concrete, measurable, time-bound milestones that, if achieved, would constitute meaningful progress toward the objective. Grove emphasizes that the framework should be reviewed and adjusted on a quarterly cadence, that key results should be ambitious enough that achieving 70 percent represents good performance, and — critically — that OKRs should not be used as the basis for compensation decisions.
The last point is the one most modern adopters have lost. Grove's argument was that compensation tied to OKRs would corrupt the goal-setting process. People would set goals they were certain to hit; they would lower the ambition of the goals to game the bonus; they would resist updating the goals in light of new information. The whole point of OKRs, Grove argued, was to be a forcing function for honest conversation about what the team was trying to do and how it was going. Tying that conversation to compensation made honest conversation impossible.
Doerr, Google, and the Great Diffusion
John Doerr learned OKRs as a young salesman at Intel under Grove. When he became a partner at Kleiner Perkins and invested in Google in 1999, he taught the framework to Larry Page and Sergey Brin. Google's early adoption of OKRs is one of the most-cited cases in modern management literature, and its actual implementation was relatively close to Grove's original. Goals were set quarterly, scored on a 0-1 scale, and not used for compensation. Underperformance was treated as information rather than as a failing. The framework gave Google a transparent shared language for prioritization across what eventually became a 100,000-person organization, and it served the company well through its growth.
Doerr's 2018 book Measure What Matters propagated the framework to a much wider audience. The book is sincere, well-written, and accurate to Grove's original. The audience that bought it, however, included thousands of CEOs who skimmed the case studies, picked up the OKR vocabulary, and proceeded to implement the framework in ways that violated its underlying principles in nearly every important respect. The diffusion of OKRs after 2018 is a textbook case of how good ideas degrade as they propagate.
Goodhart's Law and the Bonus Problem
Charles Goodhart, a British economist, observed in 1975 that 'when a measure becomes a target, it ceases to be a good measure.' The phenomenon is now called Goodhart's law, and it is the single most reliable empirical regularity in the literature on incentive design. Once compensation, promotion, or status is tied to a metric, the metric will be gamed — sometimes consciously, often unconsciously — and within a few cycles will no longer reflect what it was originally measuring.
The most common modern misuse of OKRs is to tie them to performance reviews and bonuses. The cycle is predictable. In the first quarter, OKRs are set ambitiously. The team underperforms; bonuses are reduced. In the second quarter, OKRs are set conservatively to ensure achievability. The team hits all of them; the company concludes that OKRs work and continues the practice. By the fourth quarter, the framework has degenerated into a forecasting exercise — predict what you can deliver, deliver it, get paid. The company has lost the function the framework was supposed to provide (forcing honest conversation about ambition) and has gained nothing in return except a slightly more elaborate way of distributing bonuses. Several years of this dynamic typically produce a thoroughly corrupted goal-setting process and a workforce that has stopped taking the framework seriously.
The Research on When OKRs Help
Empirical research on OKRs is thinner than on DORA-style operational metrics, but a consistent pattern emerges. OKRs help when they are used as a transparent prioritization tool, when they are reviewed honestly, when underperformance is treated as data, and when they are decoupled from compensation. They harm when they are used as a top-down accountability mechanism, when underperformance is punished, when the cadence is too frequent for the actual feedback loops of the work, or when the cycle is treated as ceremonial.
The broader research on goal-setting — Edwin Locke and Gary Latham's decades of work on the topic — supports a more complex picture. Specific, ambitious goals do produce higher performance than vague or easy ones, but only when the work is straightforward, the path is clear, and the goal-setter has accurate information about what is achievable. For complex, exploratory, novel work — which describes most of modern knowledge work — goal-setting can backfire by causing the worker to optimize for the wrong dimension, to ignore opportunities not captured in the goal, or to engage in unethical behavior to hit the target. The Sears auto-repair scandal of the 1990s, in which mechanics on aggressive sales-target compensation systematically overcharged customers and replaced parts that did not need replacement, is a famous example of goal-setting gone wrong in exactly this way.
How to Run OKRs Well
If you are using OKRs and want them to actually work, the operational answers are straightforward and unfashionable. Decouple them from compensation. Use them as a forcing function for prioritization conversations, not as a performance-evaluation instrument. Set fewer of them — three to five objectives is plenty for a senior team; ten is too many. Score 0.7 as a successful outcome and treat 1.0 as suspicious (it usually means the goal was set too low). Review them honestly every quarter; if a goal turns out to have been wrong, change it rather than pretending it was right. Be willing to abandon the framework entirely if it stops producing the conversation it was supposed to produce.
The deeper move is to remember what Grove was trying to do. OKRs were not the point. The point was that managers and their teams needed to have an explicit, periodic, structured conversation about what they were trying to accomplish and how they would know they had accomplished it. The framework is a scaffolding for that conversation. If your scaffolding has become a load-bearing wall — if the OKRs themselves have become more important than the work they were meant to coordinate — the scaffolding has failed. Take it down and have the conversation directly.
OKRs are a useful tool that has been widely misused. The Andy Grove who wrote about them in 1983 was a working CEO trying to give his middle managers a way to align their teams' work with the company's direction without micromanaging. The framework that propagated from Intel through Google through Measure What Matters into the modern enterprise is, in many companies, exactly the kind of compensation-tied, top-down, theatrically-rigorous performance apparatus that Grove warned against. Returning to Grove's original — fewer goals, decoupled from compensation, used to provoke honest conversation rather than to enforce compliance — is one of the cheaper improvements available to most organizations. It costs no software, no consultants, and no change-management initiative. It costs only the willingness to read a chapter of a 1983 book and follow what it says.
Sources
- 1.OKR — Objectives and Key Results · Wikipedia
- 2.High Output Management · Wikipedia
- 3.Measure What Matters — John Doerr · Wikipedia
- 4.Goodhart's law · Wikipedia
- 5.Edwin Locke — goal-setting theory · Wikipedia
- 6.Sears auto-center scandal of the 1990s · Wikipedia