KPIs monitor. OKRs move.

A KPI is a permanent instrument. It watches the health of business-as-usual, it has a threshold, it lives on the KPI Scorecard, and its job is to catch drift before drift becomes a fire. Nobody celebrates a KPI. That is the point of it: a KPI at threshold is a machine running as designed.

An OKR is a time-bound bet. It targets a step change, not a steady state: a from-X-to-Y Key Result that would be worth a quarter of a team's focused effort. It lives in the quarterly cycle, and it is worked, deliberately, through initiatives, process commitments and experiments. When the bet lands, the OKR retires.

The two costumes

Confusion between the two wears one of two costumes, and both are easy to spot once named.

The first is a KPI dressed as a Key Result: a "Key Result" that restates a dashboard number at roughly its current level. "Maintain uptime at 99.9%" is not a bet, it is a threshold. Nothing about it deserves a quarter of anyone's attention, and putting it in an OKR teaches teams that OKRs are paperwork. (The other costume in this wardrobe is the task, which we deal with separately in Task-Based Key Results.)

The second is an ambition dressed as a KPI: a scorecard metric with no threshold, which nobody would act on if it moved. That is not monitoring, it is an orphaned ambition, and it belongs either in an OKR with a real target or off the dashboard entirely.

The membrane between them

The useful insight, and the one most OKR content misses entirely, is that the boundary between KPIs and OKRs is a membrane, not a wall. Metrics cross it in both directions, on purpose.

A KPI promotes to an Operational Excellence OKR in two situations: when it is broken beyond what firefighting can fix, or when you deliberately decide to push it to a new level. Onboarding time has crept from three days to nine? That is no longer monitoring, that is a quarter's work. So it is promoted, worked as an OKR with real initiatives behind it, and when the step change lands it demotes back to the scorecard with a new threshold to hold.

This promotion and demotion cycle is what makes the pair a system rather than two competing lists. You need both, and you need to know which is which at any given moment, which is the argument of BAU Has Its Own Approach.

How this connects

Underneath the pair sits a measurement hierarchy worth knowing: raw measurements feed metrics, and a few metrics are designated KPIs, the critical few that drive success. Choosing which few is where leading and lagging indicators earn their keep: lagging indicators confirm the right things happened, leading indicators buy you time to act. And a Key Result targeting a metric that moves too slowly to respond within a quarter is a bet you cannot learn from, which is why the from-X-to-Y discipline matters as much as the number itself.

The test in practice

Take any metric your leadership team looked at this week and ask two questions. Would we act if it moved? And are we currently trying to change its level? No to the first: take it off the dashboard. Yes to the second: it should be in an OKR, with a team and a deadline, not sitting on the scorecard pretending to be weather. Ten minutes of this sorting usually redraws half a company's "OKRs" into KPIs, and the honesty is the beginning of the fix.