Metrics question

You are the product manager of ratings in an ecommerce website say amazon / flipkart. On one day, the total number of ratings dropped to 8000 from a general average of 10000 / day. Identify the root cause.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Root cause diagnosis for a metric drop in a quality signal on a large e-commerce platform.

Confirm the drop is real, rule out a tracking or reporting bug before investigating a genuine behavioral cause.
Segment the drop by category, device, and user cohort to see if it is broad or concentrated in a specific area.
Check for a recent product change, e.g. a new rating prompt flow, UI change, or a bug in the review submission form.
Check for external factors, a site outage, a payment issue reducing completed orders, which are a prerequisite for most ratings, or a seasonal dip.
Form and test a hypothesis, e.g. if mobile app ratings dropped specifically, check for a recent app update or crash affecting the review flow.
Define a fix and a monitoring metric, resolving the identified cause and tracking daily rating volume going forward to confirm recovery.

Starts by validating the number rather than assuming the drop reflects real user behavior change.
Segments methodically, category, device, cohort, to localize the cause instead of guessing broadly.
Connects the rating drop to its dependency, completed orders, recognizing ratings cannot happen without a purchase first.
Proposes a concrete monitoring metric post-fix to confirm the issue is actually resolved, not just diagnosed.

Jumping to conclusions about user sentiment without first checking for a technical or tracking issue.
Not considering that a drop in orders could be the true root cause behind fewer ratings.
Stopping at diagnosis without defining how to confirm the fix worked.

Chapters of the AI PM course that teach what this question tests.