← Meta Interview Insights

Meta·Software Engineer·Technical Phone Screen·Intermediate

Intermediate
Jun 2026

Summary

Interviewed for a data science role at Meta, got a product metrics question about Messenger that seemed straightforward but had a few layers worth thinking through.

Questions Asked (1)

Q1

What percentage of users who were active on Messenger yesterday also made a video call yesterday?

Product Analytics & MetricsData Modeling
Author's notes

Defining 'active' is where I got a little tangled.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Clarify the metric definition and assumptions, then break down the calculation into numerator (users who made a video call yesterday) and denominator (users active on Messenger yesterday). Discuss how you would obtain these counts from data, considering edge cases like time zones and user overlap.

Pro tip: Show awareness of potential data discrepancies, such as users active on Messenger but not eligible for video calls (e.g., desktop users), and propose how to handle them.

1. Clarify the metric

Define what 'active on Messenger' and 'made a video call' mean precisely, including time frame (yesterday in which time zone?) and platform (mobile, desktop, web).

2. Identify data sources

Determine which tables or logs contain user activity and video call events, such as Messenger activity logs and call logs.

3. Compute numerator and denominator

Calculate the number of distinct users who made a video call yesterday and the number of distinct users active on Messenger yesterday.

4. Calculate the percentage

Divide the numerator by the denominator and multiply by 100 to get the percentage.

5. Validate and consider edge cases

Check for data quality issues, such as duplicate events, time zone mismatches, and users who are active but cannot make video calls (e.g., on unsupported devices).

Key Points to Mention

  • Definition of 'active' (e.g., opened app, sent message) and 'video call' (initiated or received).
  • Time zone considerations: use a consistent time zone, likely UTC or user's local time, and be explicit.
  • Data sources: activity logs, call logs, and how to join them by user ID.
  • Distinct user counts to avoid double-counting users with multiple sessions.
  • Edge cases: users on desktop without camera, users who attempted but failed to connect, and bots.
  • Potential need for sampling or approximation if data is large, and how to ensure accuracy.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.