The Android-specific angle is where I almost tripped up.
Start by clarifying the metric definition and scope (e.g., 1% drop in what exactly—DAU, sessions, time spent? Over what period? Segmented by platform, region, user cohort?). Then systematically rule out data/measurement issues, external factors, and internal changes before diving into user behavior and product experience.
Pro tip: Always validate the data first—many 'drops' are instrumentation bugs or logging changes. Also, segment by new vs. existing users and by Android OS version, as these often reveal the root cause quickly.
Define exactly what 'usage' means (DAU, sessions, time spent, etc.), the time frame, and whether the drop is global or segmented. Ask if it's statistically significant.
Check for logging bugs, SDK changes, app version updates, or data pipeline issues that could cause a false drop. Compare with other platforms (iOS, web) to see if it's Android-specific.
Look for external events (holidays, competitor launches, OS updates) and internal changes (product updates, experiments, marketing campaigns) that coincide with the drop.
Break down the drop by user cohorts (new vs. existing, region, device type, Android version), feature usage, and funnel steps to isolate where the drop is concentrated.
Based on segmentation, generate hypotheses (e.g., a bug in a recent release, poor performance on a specific device) and validate with further data analysis, user feedback, or A/B tests.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.