KPIs: choosing the numbers that deserve a screen
Most dashboards we see hold many numbers and few decisions. The reason is that metric selection became collecting everything measurable instead of picking what changes behavior. Here is a simpler and harsher method.
The only test: what changes if the number moves?
Take any metric you are considering and ask: if it rose twenty percent, what would I do? If it fell twenty percent, what would I do? If both answers are nothing clear, the metric does not deserve screen space however easy it is to collect. This test alone usually deletes half of metric lists.
Separate outcome metrics from driver metrics
Monthly sales is an outcome: it tells you what happened after the fact. Proposals sent this week is a driver: you can act on it today. A useful dashboard shows both, but drivers take the prominent position because they are the only movable ones.
One audience per screen
An executive needs five numbers telling them whether the path is sound. A branch manager needs entirely different numbers about their own branch. Trying to satisfy both on one screen produces a screen serving neither. Split them; the extra cost is far below the cost of a dashboard everyone ignores.
Three recurring mistakes
First: a metric without comparison, since a number alone is meaningless unless set against last month or a target. Second: averages that hide detail, since a two-hour average delivery time may hide customers waiting six. Third: metrics measured because they are easy rather than important. Try the five dashboards in our demo and see how comparison and alerting sit beside each number.
Frequently asked questions
How many metrics per screen?
Four to seven for the main screen. Anything more moves to secondary screens opened by whoever needs the detail, since the eye cannot hold more in a single glance.
Our data is missing for some important metrics, what do we do?
Show them as a declared blank, not a zero. A zero reads as a real value and misleads; a declared blank reminds the team this data needs collecting.
Should metrics change over time?
Yes, and a quarterly review is a healthy practice. A metric that has not driven a decision in three months gets deleted without hesitation.