Scoreboard Hijack

The number looks better, but the result you care about may not. Compare the rewarded proxy with an independent observation of the real outcome and trace what the reward makes people repeat.

Proxy. Real outcome. Independent check. Rewarded behavior.

Name the metric being optimized and the outcome it is meant to represent. Observe that outcome through a route independent of the scoreboard, then inspect whether rewards or allocation reinforce metric-seeking behavior. A mismatch raises a question; it does not prove gaming or justify silently changing the goal.

Try this when…

Our score rises while the underlying problem remains.

We are getting very good at what gets counted.

The easiest way to win the metric misses the point.

What would success look like away from this dashboard?

What you can use it for

Learning trackers

Compare streaks and completions with independent performance.

Service metrics

Inspect downstream resolution alongside speed or volume counts.

Creative reach

Check whether attention measures correspond to the intended audience experience.