Operating

KPI (Key Performance Indicator)

KPI (Key Performance Indicator) is a single number you commit to tracking because it tells you whether a specific part of your business is working or breaking. A good KPI is tied to a decision: when it moves, you know what to do next.

Also known as: KPI, key performance indicator, performance metric

RawactivityKPIthresholdKeep goingKill or pivot
A KPI sits between raw activity and a clear decision, with a threshold that tells you to keep going or kill it.

Why it matters

Before product-market fit, the wrong KPIs let you feel busy while learning nothing. A founder can ship features, post on social, and watch signups climb, and still have a dead business if those signups never activate, pay, or stick around. KPIs force you to name the few numbers that actually predict survival, usually activation, retention, and revenue, instead of vanity counts like page views or total users. The validation lens is simple: pick KPIs that, if they stall, tell you to kill or pivot, not ones that always look fine. Each KPI should map to a real decision, so when it crosses a threshold you act instead of debate. Track three to five, not twenty, because a dashboard nobody reads is worse than a sticky note you check daily. The best early KPI is the one closest to a customer paying or coming back, because that is the only signal that money and habit are real.

Worked example

A solo founder running a meal-planning app picks one KPI: week-4 retention, the share of new signups still active 28 days later. It sits at 9 percent. She sets a kill-or-fix line at 25 percent, ships a faster onboarding, and reruns the cohort. Retention moves to 11 percent, well short of the line, so she pivots the offer instead of pouring more money into ads that just fill a leaky bucket.

Common mistakes

  • Tracking vanity KPIs like total signups, pageviews, or social followers that go up no matter what and never trigger a decision.
  • Picking too many KPIs. If you watch twenty numbers you watch none. Choose three to five that map to activation, retention, and revenue.
  • Setting a KPI with no threshold. A number without a 'good vs kill' line is a fact, not a decision tool, so define the cutoff before you read the result.

Frequently asked questions

What is a good KPI (Key Performance Indicator)?

A good KPI is specific, tied to a decision, and movable by your team. It sits close to real value, such as activation rate, week-4 retention, or paying-customer growth, rather than counting raw signups or views. The test is simple: if the number drops, do you know what to do? If not, it is the wrong KPI.

KPI vs metric: what is the difference?

Every KPI is a metric, but most metrics are not KPIs. A metric is anything you can measure. A KPI is the small set of metrics you have decided actually predict success and that you will act on. Calling everything a KPI is how dashboards turn into noise that nobody reads.

KPI vs OKR: how do they relate?

An OKR is a goal with an objective and key results for a quarter or year. A KPI is a standing health metric you watch continuously whether or not it is a current goal. Key results are often expressed as a target on a KPI, so they overlap, but KPIs keep running after the OKR cycle ends.

How many KPIs should a startup track?

Three to five at the company level for an early-stage startup. More than that and no single number gets attention, so nobody acts when one moves. Most pre-PMF founders are best served by one north-star metric plus two or three supporting KPIs covering activation, retention, and revenue.

What KPIs matter most before product-market fit?

Retention and activation matter most, because they show whether the product delivers real value, not just curiosity. Acquisition counts and pageviews look good early but predict little. Watch the share of new users who reach the core action, then the share still active weeks later, since those signals tell you if there is a business to scale.

How do you choose a KPI for your business?

Start from the one outcome that means you are winning, usually customers paying and coming back, then pick the metric closest to it. Add a threshold that splits 'on track' from 'kill or pivot' before you look at the data. Drop any candidate KPI that does not change what you would do, because it is just trivia.

Deep-dive guideHow to Validate a Startup Idea

Related terms

More in Operating

Stop reading definitions. Pressure-test your idea.

Knowing the terms is the easy part. Olune runs your actual idea against live Reddit signals, competitor data, and real search demand, then gives you an honest GO / NO-GO verdict in about eight minutes. Free, no card.

Last updated 2026-06-09 · Back to the glossary