OKRs vs KPIs: which one a leadership team needs

An OKR pairs an objective, written in words, with three to five key results that say how you will know it happened. A KPI is a single measure with an owner, a goal and a fixed review slot. One is a container for ambition over a quarter. The other is an instrument you read on a schedule.

The OKR format came out of Intel under Andy Grove and reached most of the software industry through John Doerr's Measure What Matters, which is also where the habit of scoring key results from zero to one comes from. Doerr describes seven tenths as a respectable score on a stretch objective, which tells you a lot about what the format is for.

KPIs have no single author. The phrase comes out of management reporting, and its whole discipline is choosing few numbers rather than many. Nobody scores a KPI at seven tenths. A row is on track or it is not, and next Monday you look again.

OKRs are for change. They exist to point a team at something it is not doing yet, and they expire. KPIs are for health. They exist to tell you whether the ordinary work of the business is holding up, and they never expire. A company with only OKRs stops noticing the roof leaking while it builds the extension.

This is why an OKR that reads keep churn under three percent is a sign the format is being misused. Nothing about that is a change, and holding it steady is exactly the job a KPI does. If churn is already where you want it, it belongs on the weekly list rather than in a quarterly plan.

The reverse mistake is quieter. A KPI called new market entered is not a measure of anything. It is a project with a checkbox, and putting it on a weekly board gives you eleven weeks of no followed by one yes.

At the key result. A well-written key result is a number with a target and a date, which is a KPI wearing a quarterly deadline. That overlap is real and it is where most of the confusion comes from. The difference is that a key result retires at the end of the quarter and a KPI keeps being read afterwards.

A practical way through it: if a number matters only because of this quarter's plan, track it as a key result and let it go in thirteen weeks. If you would still want to know the number next year, it is a KPI and the quarterly plan is just changing its target for a while.

The teams that get tangled here are usually running both systems in two different tools, with two different owners and two different review meetings. That is a tooling problem dressed up as a philosophy problem. One list, one meeting, two columns solves most of it.

Almost always, once the company is past its first few people. Run a short weekly number list for health and a small quarterly set of priorities for change, and review them in the same meeting. What you should not run is two parallel systems with separate owners, separate tools and separate reviews, which is how both end up ignored.

Start with the weekly numbers if you have neither. They are cheaper to build, they teach the team what a target argument feels like, and they usually surface the two problems your first quarter of priorities should be aimed at anyway.

Add quarterly priorities once the weekly read is boring, which is the correct state for it. A five-minute scan where everything is on track is not a waste of time. It is the whole point, and it frees the rest of the meeting for the work that changes something.

Take a services firm losing money on delivery. The quarterly priority is to rebuild how jobs are scoped, owned by the operations lead, finished when every new job uses the new template. The weekly numbers underneath it are jobs scoped the new way, rework hours and gross margin, and only the first of those moves quickly.

Weeks one to three look like failure. Rework hours are unchanged, margin is unchanged, and the only row moving is the count of jobs scoped the new way. That is the leading measure doing its job, and a team without one would spend those three weeks arguing about whether the priority was working.

By week eight the rework hours start dropping, because the jobs scoped in week two are now being delivered. Margin follows in the month after that. When the quarter closes, the priority retires and two of the three numbers stay on the weekly board permanently, with the targets reset to the new normal.

None of that requires scoring anything out of one. It requires knowing which number was supposed to move first, and being honest in week three when it did not.