The metric you optimize is the behavior you get.

Metrics do not measure behavior neutrally. They shape it. Whatever you put on a dashboard becomes what people work toward, whether or not it is what you actually want.

We tracked support tickets resolved per day. The team got fast at closing tickets. Resolution time dropped. Then we noticed something: repeat contacts were going up. Tickets were being closed without being solved, because closed was what got measured.

We switched to a different metric: percentage of contacts that did not generate a follow-up within 7 days. That number was lower and harder to move. It also reflected what we actually wanted, which was problems that stayed solved.

Within 2 months, repeat contacts dropped 34%. Ticket volume dropped too, because fewer people were coming back.

What metric are you tracking that might be rewarding the wrong behavior?