87. Analytics show that nobody uses a control that research participants rely on. How would you investigate?
MediumEvaluation
Case
Show model answer
Check whether the event actually fires, how eligibility and exposure are defined and whether the dashboard includes the relevant users and versions. Observed use is a reason to inspect measurement, not to dismiss either evidence source.
Compare permitted event traces with representative task observations. Correct the measurement and preserve any break in the time series before recommending removal based on an apparent absence of use.
What interviewers listen for
- Verify event and population definitions
- Compare traces with observed tasks
- Correct measurement before removal decisions