OKR

The original OKR, in brief

Created by Andy Grove at Intel in the 1970s, adapted from Peter Drucker's management by objectives, and described in "High Output Management" (1983). John Doerr brought it to Google in 1999 and wrote "Measure What Matters" (2018).

Andy Grove built OKRs at Intel in the 1970s and described them in High Output Management (1983)¹. Intel called the system iMBOs, for Intel Management by Objectives². It adapted Peter Drucker's management by objectives, with two changes that stuck: goals set every quarter instead of every year, and kept apart from pay².

Grove's version answers two questions. Where do I want to go? That is the objective. How will I pace myself to see if I am getting there? Those are the key results, and he treated them as milestones you can check without argument¹. He wanted few objectives, because the point of the system is focus. He also warned that it is "not a legal document upon which to base a performance review"¹.

John Doerr took Grove's course at Intel in 1975 and brought the method to Google's founders in 1999²³. His book Measure What Matters (2018) made it famous³. Google's public guide suggests three to five objectives with about three key results each, and holds company-wide meetings each quarter to share and grade them⁴.

Common misreadings

• "Key results are the to-do list." A key result shows that the objective moved. Google's playbook defines key results as measurable milestones that advance the objective⁵. "Launch the new welcome email" is a task. "Raise 30-day activation from 48% to 65%" is a key result.
• "Tie them to bonuses." Grove said the opposite¹, and Google's guide says OKRs are not the same as performance evaluation⁴. Once pay rides on the score, people set targets they know they can hit.
• "More OKRs, more ambition." Grove built the system for focus¹. A team with twelve objectives has no priorities. Google's own ceiling is five⁴.
• "Hitting 70% is always good." Only for aspirational OKRs. Google's playbook splits OKRs in two. Committed OKRs are promises, expected to score 1.0, and a miss needs an explanation. Aspirational OKRs are stretch goals, expected to average about 0.7⁵. Mixing the two up is the first trap the playbook lists⁵.

OKR examples

When OKR fits, where it struggles, and what it does not answer

How to implement OKR

1. Start from the strategy. OKRs turn choices you have already made into goals for the next quarter. If the strategy is still open, settle it first.
2. Write three to five objectives. Plain words, no numbers, each one worth a quarter of effort. If you cannot remember them all without looking, there are too many⁴.
3. Give each about three key results. Each is a number with a start and a target that shows the objective moved. If it describes work to be done, it is a task; move it to the team's plan⁵.
4. Mark each OKR committed or aspirational. Committed OKRs are promises, expected to be met in full. Aspirational ones are stretch goals, where about 70% is a good result⁵.
5. Check in often, score at the end, keep it away from pay. Check key results every week or two. Score each from 0 to 1.0 at the end of the quarter and share the scores⁴. Keep scores out of bonuses and ratings¹.

Adopting OKR: the first cycle and the rhythm

First cycle. Plan on one quarter. Set the company objectives and key results in a half-day or full-day session. Then give teams about two weeks to draft their own under them and agree them with their manager. Google sets annual OKRs as well as quarterly ones⁴; a smaller company can start with the quarter alone.

Rhythm. Check key results every week or every other week, in a few minutes, by number. Score each one at the end of the quarter, on a scale from 0 to 1.0, and share the scores openly⁴. Then set the next quarter's OKRs with what you learned.

How long until it runs by itself. Expect the first two or three quarters to feel clumsy. Teams tend to write too many OKRs, set targets too safe or too wild, and forget to check in mid-quarter. The habit usually settles once people have seen a few end-of-quarter scores and what came of them.

What derails it. Writing tasks as key results. Setting more objectives than anyone can remember. Linking scores to bonuses. Setting OKRs in January and looking at them again in March. Not saying which OKRs are promises and which are stretch goals⁵.

Common OKR mistakes

• Writing tasks as key results. "Launch the new welcome email" is work. "Raise activation from 48% to 65%" is a result. A key result should show the objective moved⁵.
• Too many OKRs. Grove built the system for focus¹. Ten objectives means no priorities. Stay at five or fewer⁴.
• Tying scores to pay. Once a bonus rides on the score, people set targets they know they can hit. Grove and Google both keep OKRs out of performance reviews¹⁴.
• Not saying which OKRs are promises. A team that treats a commitment as a stretch goal lets others down; one that treats a stretch goal as a commitment sets it too low. Google's playbook lists this as the first trap⁵.
• Business as usual dressed up. An OKR that describes what the team would do anyway adds nothing. Ask what would change if you hit it⁵.
• Set and forget. OKRs set in the first week and read again in the last week are a wish list. Check key results every week or two.

Does OKR work? The evidence

On OKRs themselves, there is little. A systematic mapping study published in 2024 found 47 studies on OKRs and concluded that the method is under-documented, with few academic studies in depth⁹. Most published cases describe one company and rely on interviews. No controlled study shows that companies using OKRs outperform those that do not.

The parts rest on strong research. Edwin Locke and Gary Latham summarized 35 years of goal-setting studies in the American Psychologist in 2002⁸. Their core finding: specific, difficult goals lead to higher performance than easy goals or a vague "do your best," as long as people are committed, have the ability and get feedback⁸. Effects are weaker on complex tasks that people are still learning⁸. That is the case for measurable key results and regular check-ins.

Team goals work; the wrong personal goals do harm. A 2011 meta-analysis of group goal setting in the Journal of Applied Psychology found that specific, difficult group goals raised group performance well above vague ones. Individual goals aimed only at personal results lowered it when people depended on each other⁷. The authors also noted few recent field studies inside real organizations⁷. OKRs' ancestor, management by objectives, showed productivity gains in 68 of 70 studies in a 1991 meta-analysis in the same journal, with the largest gains where top management was committed¹⁰.

The strongest criticism. In "Goals Gone Wild," published in Academy of Management Perspectives in 2009, Lisa Ordóñez and three colleagues argued that goal setting is oversold⁶. They listed side effects: neglect of what the goal leaves out, riskier choices, more cheating, less learning, and less intrinsic interest⁶. Locke and Latham replied in the same issue that the critics overstated the case and that these risks come from badly designed goals¹¹. Both sides agree on the practical point: set goals with care, watch what they leave out, and do not tie stretch goals to pay.

OKR compared

OKR glossary

Objective

What you want to achieve, in plain words. Short, motivating, no numbers.

Key result

A measurable result, with a start and a target, that shows the objective moved.

Committed OKR

A promise the team will deliver in full, moving time and people to do it. Expected to score 1.0⁵.

Aspirational OKR

A stretch goal with no sure path. Expected to average about 0.7, with wide swings⁵.

Score

How much of a key result was achieved, from 0 to 1.0, graded at the end of the period⁴.

iMBO

Intel Management by Objectives: the name OKRs had at Intel under Andy Grove².

MBO

Management by objectives, Peter Drucker's idea from the 1950s that OKRs grew from².

Sandbagging

Setting targets you know you can beat, to look good at the end of the quarter⁵.

CFRs

Conversations, feedback and recognition: the ongoing talk between managers and staff that Doerr pairs with OKRs³.

Free tools that help