GuidanceSubmissions
30:00
Guidance

You've just launched Facebook Jobs, a marketplace for employees and employers.

Job listings are up 20%, but employer response time for applicants is down 20%. How would you approach this?

You've just launched Facebook Jobs, a marketplace for employees and employers.

NOT SAVED
0 words·0 / 8,000 chars
Product Execution

The Root Cause Analysis Playbook

Diagnosing a metric drop is a systematic investigation, not a guessing game. The key is to logically narrow down the problem space—from macro to micro—before jumping to conclusions.

Step 1

Clarify & Validate the Metric

Before panicking, ensure the data is real. Ask clarifying questions: Is this a sudden drop or a gradual decline? Is it a 5% drop or a 50% cliff? Most importantly, is the logging broken? A sudden 100% drop in an event usually means an engineering tracking bug, not user behavior.

Step 2

Analyze Time & Seasonality

Step 3

Slice and Isolate the Data

Step 4

Investigate Internal Factors

Step 5

Investigate External Factors

Step 6

Synthesize and Propose Next Steps

Additional Tips

  • Think MECE (Mutually Exclusive, Collectively Exhaustive): Show the interviewer you are highly structured. Group your hypotheses clearly into buckets (Internal vs. External, Technical vs. Behavioral) so you don't miss any blind spots.
  • The TPM Flex (3rd-Party Dependencies): Stand out by mentioning downstream services. A drop in e-commerce checkout conversion isn't always a UX issue; it could be the third-party payment gateway timing out.
  • Cannibalization is Normal: Remind the interviewer that sometimes a metric drop is expected. If you launch a hyper-efficient "Quick Reorder" button, your "Time Spent in App" metric will drop—but that's actually a UX win, not a problem.