Why a checklist prevents the most expensive mistakes when choosing KPIs
Choosing KPIs seems simple until the dashboard is three months old and nobody opens it. Most problems don’t come from the analytics tool, but from decisions made during the definition phase. This checklist brings together the mistakes that come up again and again in measurement audits, with the practical nuance of what to do in each case.
Checklist of mistakes when choosing KPIs (and how to spot them in time)
1. Confusing a metric with an indicator
A metric is a data point (sessions, impressions, leads). A KPI is a metric tied to an objective with a success threshold. If “unique visitors” appears on your list without an associated goal, it isn’t a KPI: it’s a number that bloats the dashboard.
Check: for each row of the dashboard, write in one sentence what decision you would make if that value rose or fell by 20%. If there’s no decision, it’s unnecessary.
2. Choosing KPIs you don’t control
A typical case: a content team measures “market share” or “total revenue” as its main KPI. These are business outcomes, not operational levers. The team can’t act on them directly and ends up ignoring them.
It’s worth separating two layers: business KPIs (visibility for management) and operational KPIs (the ones the team can move this week). Mixing them in the same table breeds frustration.
3. Ignoring data quality before setting the target
Setting a conversion goal when event tracking is poorly implemented is a mistake you discover late. According to the official documentation of analytics platforms, changes in the consent model and in cookie blocking affect the comparability of historical series.
Practical check: before committing to a KPI, validate for two weeks that the data arrives consistently and that two different sources (for example, the analytics platform and the CRM) tell a coherent story. If they differ by more than a reasonable margin, fix the measurement first.
4. Measuring only what’s easy to measure
Availability bias leads people to choose KPIs because the tool provides them for free. That’s how vanity metrics appear (page views, followers, time on page) and the ones that matter disappear (activation rate, cost per qualified customer, 90-day retention).
This doesn’t work the same way in every sector: in a local business with a phone-call sales cycle, the form-fill KPI may be noise; in a B2B SaaS, time on page rarely predicts anything. Recognize which metrics don’t serve your model before including them.
5. Defining the KPI without a formula or source
“Conversion rate” can mean five different things depending on who calculates it. If the formula, the period, the source and the exclusions (internal traffic, bots, countries) aren’t documented, every report will tell a different version.
Check: every KPI needs a spec sheet with name, formula, source, frequency, owner and alert threshold. Without a spec sheet, it doesn’t go on the dashboard.
6. Accumulating too many KPIs
A panel with 40 indicators doesn’t get read. Practical experience shows that teams go back to looking at two or three numbers when the list grows. Limit the main panel to a handful and leave the rest in detailed reports on demand.
7. Not linking the KPI to a threshold and an action
A KPI without a threshold is decoration. Define what value is considered acceptable, what the alert zone is and what is done when it’s crossed. For example: if cost per lead rises by more than an agreed percentage for two weeks, campaign segmentation is reviewed.
8. Forgetting periodic review
KPIs expire. An indicator that’s relevant in the launch phase stops being so when the business matures. Schedule a quarterly review to retire the ones that no longer guide decisions.
What this checklist doesn’t solve
No list replaces judgment about the business model. Some KPI decisions depend on context that isn’t in the data: management priorities, partner agreements or legal restrictions. Nor does it guarantee that the data is perfect; it only reduces the risk of choosing indicators that nobody will use. If your organization doesn’t have a data-driven decision culture, changing the KPIs won’t fix the underlying problem.