~/learn/Product Metrics · Diagnosing a metric drop
Internal vs external causes
Sort hypotheses into things your company did and things the world did.
why this matters
Sorting hypotheses into buckets keeps your answer structured and complete.
Once you've confirmed the drop is real, hypotheses fall into two buckets. Internal causes are things your company changed: a product launch, a bug, a pricing change, an experiment ramp, a marketing campaign ending, an infrastructure outage, or a tracking change.
External causes come from outside: seasonality (holidays, school terms, weekends), a competitor launch, news events, an app-store policy change, a regional internet outage, or a change in a partner platform like Google search ranking.
Hypothesis buckets
INTERNAL EXTERNAL release / bug seasonality / holiday logging or pipeline change competitor launch experiment ramp news / regulation marketing spend paused platform change (OS, app store) pricing change regional outage
Interview trap
Seasonality is the explanation candidates forget most. Always compare to the same period last year (year-over-year) as well as last week.
Pro tip
Use timing as evidence: a cliff on a single day points to a release, outage or logging change; a slow slide over weeks points to competition, product quality or market shifts.
key takeaways
- Internal: releases, bugs, logging, campaigns.
- External: seasonality, competitors, platforms.
- Timing hints at the cause.
check your understanding
Bookings slid gradually over six weeks in one country. Which cause fits best?