Worked Example: Build a Falling Activation Tree
Build a diagnostic issue tree for a falling activation metric and select a first test.
Activation fell from 46% to 35% after release 3.8, and the team is proposing an onboarding redesign before identifying the cause. Parent question -> MECE cause branches -> testable hypothesis -> cheapest discriminating test The common trap is treating users need a better checklist as the problem statement. That smuggles in a solution, hides rival causes like permissions or measurement changes, and can send a sprint toward the wrong branch. Parent question Write: What explains the 11-point activation decline after release 3.8, and which cause is controllable this sprint? The question names the metric, magnitude, timing, and action boundary without…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in