Prove the opportunity
Use player evidence to size the problem and link the expected behaviour change to GGR.
PM sets baseline, target, guardrails and stop rule.7 Stopping low-value delivery
No value case, no commitment. No impact, no success claim.
Set the value test before build. Act on the result after release.
Test a small change
Test the riskiest assumption. Set up tracking and a measurement window.
Data checks the metric; Engineering ships safely.Check the outcome
Compare exposed players with baseline or control; check adoption and guardrails.
Review: proven value, early signal or no effect.Do not declare success. Diagnose, then decide.
1 · Verify
Check tracking, exposure, usage and the agreed window.
2 · Explain
Find the broken link: discovery, adoption, behaviour or value. Check segments and guardrails.
3 · Act
Try one time-boxed fix if justified; otherwise stop or roll back.
PM proposes the action; Product Lead decides on further capacity and records the learning.
Example: Loyalty points history launches, but repeat play is flat. Check whether players saw and understood it. If one focused fix has no effect, stop rollout and move capacity to a stronger retention opportunity.