The metrics that measure success are the ones tied directly to a decision you can act on. A earns its place only if a change in the number would change what you do next week — everything else is a vanity metric wearing a suit. This guide shows how to choose those decision-driving KPIs, set honest baselines, and read the numbers without fooling yourself.
Key Takeaways
- A real KPI changes a decision. If the number moves and your plan doesn’t, it’s a vanity metric — track it, but don’t manage by it.
- Pair every outcome metric with a driver metric. Revenue tells you what happened; the drivers (, activation, retention) tell you why and what to pull next.
- Baselines beat targets. You cannot judge “good” without knowing your own last-90-days number and the direction of travel.
- Fewer, better. Most teams track too many KPIs and act on too few. Pick 3–5 per goal and defend them ruthlessly.
- Segment or stay blind. A flat aggregate hides the mobile-vs-desktop or new-vs-returning gap that actually explains the trend.
What counts as a real KPI — and what’s just a number?
A key performance indicator is a metric you’ve agreed to manage a specific objective by. The test is simple: if the number changed materially next month, would anyone do something different? “Total pageviews” usually fails that test — it goes up and nobody changes course. “Trial-to-paid conversion rate” passes: a two-point drop triggers an investigation. Vanity metrics feel good and read well in a board deck, but they don’t point at an action. Reserve the KPI label for the handful of numbers you’d actually re-plan around, and let the rest live as context.
Which KPIs should you actually track?
Start from the objective, not the dashboard. Map each goal to one outcome metric plus the one or two drivers that move it:
- Grow revenue: outcome = revenue or MRR; drivers = conversion rate, average order value, customer acquisition cost.
- Keep customers: outcome = retention or ; drivers = product activation rate, support resolution time, Net Promoter Score.
- Improve the funnel: outcome = overall conversion rate; drivers = step-by-step drop-off, page load time, add-to-cart rate.
Net Promoter Score and CSAT are useful for satisfaction, but treat them as leading indicators, not the scoreboard — they predict retention rather than pay the bills. The point is coverage: one number that tells you if you’re winning, and one or two you can actually pull.
Why outcome metrics alone will mislead you
Outcome metrics — revenue, signups, churn — are lagging. By the time they move, the cause is weeks in the past. If you manage only by outcomes, you’re driving by the rear-view mirror. Leading indicators (activation rate, trial starts, sales-qualified leads) move first and give you time to react. The discipline is to hold both: report the outcome to prove you’re winning, and watch the leading driver to steer. A month where revenue holds but activation quietly drops is a month you’ve already lost — you just won’t feel it until next quarter.
How to set a baseline and a target that isn’t made up
Before you set any target, establish the baseline: your own trailing 30-, 60-, and 90-day figure for the metric. That trailing number, plus its direction, is the only honest starting point. From there:
- Anchor to your own history. A target should be a defensible improvement on your baseline, not a round number someone liked.
- Add a public benchmark for sanity, not as gospel. Industry averages vary wildly by sector and source, so use them to check whether your number is roughly sane, not to set the goal.
- Set a threshold that triggers action. Decide in advance the level at which you’ll investigate or change course.
Without a baseline you have no way to call a result good or bad — you’re just reacting to whatever the number happens to be that day.
How to read the numbers without fooling yourself
Aggregates lie by omission. A flat conversion rate can hide mobile cratering while desktop climbs; a steady retention number can mask a great new cohort papering over a churning old one. Three habits keep you honest: segment every KPI by the dimension that matters (device, channel, new-vs-returning, plan tier); compare like periods to strip out seasonality; and watch the trend line, not the single point — one data point is noise, a direction is signal. When a KPI moves, your first question is always “for whom?” — the answer is where the action is.
Common KPI mistakes to avoid
- Tracking everything. A 40-metric dashboard is a place numbers go to be ignored. Cut to what drives decisions.
- Confusing correlation with cause. A metric moving alongside revenue isn’t proof it moves revenue — validate before you optimize for it.
- Gaming the number. The moment a metric becomes a target, people optimize the metric instead of the outcome. Watch for that.
- No owner. A KPI without a named person responsible is a KPI nobody will fix.
Alternatives: when a KPI dashboard isn’t the right tool
KPIs are for steering an ongoing operation. When you’re trying to understand why something happened, reach for qualitative tools instead — session recordings, customer interviews, and support-ticket themes explain the “why” that a number can’t. For one-off decisions (should we launch this feature?), a focused analysis or experiment beats a standing dashboard. Use KPIs to know something changed; use research and testing to know what to do about it.
How to build a KPI dashboard people actually use
A dashboard nobody opens is just decorated data. To make KPIs drive behavior, design the dashboard around decisions, not around everything you can measure. Lead with the three-to-five KPIs that actually change what the team does, show each against its baseline and trend so “good or bad?” is answerable at a glance, and give every KPI a named owner responsible for acting on it. Cut the vanity metrics that pad the view and dilute attention. The test of a good KPI dashboard is simple: when a number moves, does someone notice and do something? If your dashboard doesn’t pass that test, it’s reporting, not managing. Fewer metrics, clearer baselines, and clear ownership turn a passive dashboard into an instrument the team steers by.
Frequently Asked Questions
How many KPIs should a business track?
Three to five per objective. Enough to see the outcome and its main drivers, few enough that every one gets attention. Beyond that, metrics become wallpaper.
What’s the difference between a KPI and a metric?
Every KPI is a metric, but most metrics aren’t KPIs. A KPI is a metric you’ve committed to manage a specific goal by — one that would change a decision if it moved. The rest are supporting context.
How often should KPIs be reviewed?
Match the cadence to how fast the metric can move. Operational drivers (traffic, conversion) suit weekly review; outcome metrics like retention or revenue read better monthly or quarterly, where the signal outweighs the noise.
What is a leading versus a lagging indicator?
A leading indicator moves before the result (trial starts, activation rate) and lets you react early. A lagging indicator confirms the result after the fact (revenue, churn). You need both — one to steer, one to score.