Execution and measurement
KPI vs metric vs OKR: what the difference actually is
A metric is any number you can measure. A KPI is a metric you have decided is decisive, with a target and an owner. An OKR is a goal-setting format: an objective plus key results, usually quarterly. Metrics describe, KPIs judge, OKRs commit.
These three terms get used interchangeably in meetings, which is why reviews so often end with everyone agreeing on a number and disagreeing on what it meant.
Metric
Any quantity you can observe: page views, tickets closed, days to hire, cost per unit. A metric has no opinion attached. Most organizations have hundreds and should keep them — they are diagnostic material.
A metric answers: *what happened?*
KPI (key performance indicator)
A metric that has been promoted. To be a KPI it needs four things:
- A definition written down, including exclusions.
- A target and a baseline, each with the date it was set.
- An owner — a person, not a team.
- An update frequency and a named source.
The word doing the work is *key*. If you have thirty KPIs, you have thirty metrics and no key ones. One to three per goal is the working range.
A KPI answers: *are we succeeding or not?*
OKR (objective and key results)
A goal-setting format, not a measurement type. An objective states a qualitative outcome; three to five key results state how you will know it happened, each numeric. OKRs are usually set quarterly and often deliberately ambitious, with 70% attainment treated as good.
An OKR answers: *what are we committing to this quarter?*
Target
Simply the value you are aiming at, on a date. Both KPIs and key results carry targets. A KPI without a target is a dashboard tile.
How they fit together
- Your metric library is broad and mostly diagnostic.
- Your KPIs are the small promoted subset attached to goals, tracked continuously across years.
- Your OKRs (if you use them) are the current quarter's commitments, and their key results usually *reference* KPIs rather than inventing new numbers.
That last point prevents the most common duplication: a team writes quarterly key results with new definitions, so the quarterly number and the annual KPI disagree. Point key results at existing KPI definitions instead.
Choosing the right instrument
| Situation | Use |
|---|---|
| Multi-year outcome the board judges you on | KPI |
| Quarterly focus for a fast-moving team | OKR whose key results reference KPIs |
| Diagnosing why a KPI moved | Metrics |
| Regulatory or public reporting | KPI with a written definition and source |
Leading and lagging, briefly
A lagging KPI reports the outcome after the fact: retention, reserve ratio, permit processing time. A leading KPI moves earlier and is partly under your control: onboarding completion rate, share of applications submitted online, pipeline coverage.
Pair one of each per goal. Lagging alone tells you too late; leading alone lets you feel productive while the outcome does not move. The selection method is in how to choose the five KPIs that matter.
Three rules that prevent most arguments
- One definition per number, stored once. Two systems holding "active clients" differently will produce two truths.
- Every KPI has a person's name on it. Central ownership of all measures creates a bottleneck and stops updates.
- Never change a definition without re-baselining and dating the change. Silent redefinition is how a plan loses credibility in one meeting.
Common questions
- Is every KPI a metric?
- Yes. A KPI is a metric that has been given a definition, a baseline, a target, an owner and an update frequency because it is decisive for a goal.
- Can you use OKRs and KPIs together?
- Yes, and it is usually the cleanest arrangement: KPIs are the enduring measures attached to goals, and quarterly key results reference those same KPI definitions rather than introducing new numbers.