Validation demo · simulated data
Investigation: mobile onboarding
The recommendation is grounded in a small evidence bundle: current value, comparison segment, baseline and funnel change.
Mobile D1
8.7%
-9.2 pp vs desktop
Desktop D1
17.9%
Stable
Step 3 churn
67%
+25 pp vs baseline
Affected users
3,420/wk
Estimated
Evidence chain
1
Segment divergence
Mobile retention underperforms desktop materially.
2
Funnel localization
The largest new drop appears at Tutorial Step 3.
3
Timing
The shift started after Thursday's content update.
4
Guardrail check
Desktop retention and payer conversion did not decline similarly.
AI analyst
Likely friction in Step 3
The strongest supported hypothesis is increased mobile friction inside Step 3. The data supports testing a shorter step; it does not prove a causal UI defect.
Recommended next action
Test 3 actions instead of 5 for new mobile players while keeping desktop unchanged.