What you are asking for
You want the team to reduce the problems that keep coming back. The review counts requests handled and tasks done.
What the system hears
“The more I handle, the better I look. If I fix the cause, there is nothing left to count.”
What that produces
You pay for hours, you get hours. If the review rewards volume handled, the team handles. Whoever fixes the cause sees their count drop, and their work disappear.
The support team rewrites the help page on refunds. Questions on the subject drop from forty to ten a week. It is the best work of the month. In a review that counts requests handled, it is a drop in activity.
Measure the effect: requests avoided, time saved, fewer errors. Separate what is observed, what is estimated, and the cost moved elsewhere. Compare it to the number of users, to be sure the drop comes from the fix and not from support becoming hard to reach.
Recognise prevention in the same terms as production: in reviews, in appraisals, in the means you give. One isolated compliment is not enough.
The decision
→ At the next review, present one improvement by its effect: what went down, by how much, and the work it took.
→ Credit every contribution: the people who built, the people who maintain, the people who serve day to day.
→ Bring these effects into the criteria for appraisals and means.
To discuss
Which useful contribution stays invisible in our reviews?