Diagnosing a 20% Decline in Facebook Live Usage
Company: Meta
Role: Product Manager
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Other
You are the product manager for Facebook Live, the feature that lets people, Pages and Groups stream video in real time while viewers watch, comment and react. You notice that Live usage has fallen 20%. Walk through how you would find the root cause.
```hint Pin down the number first
Live has broadcasters and viewers, and "usage" could mean either side's activity. Decide which metric fell and confirm it is measured correctly before you look for reasons.
```
### Clarifying Questions
- Which metric fell 20%: live broadcasts started, unique broadcasters, unique viewers, minutes watched, or comments and reactions on live videos?
- What is the comparison window (week over week, month over month, year over year), and did the drop appear on one day or build up gradually?
- Is the decline spread evenly, or concentrated in particular countries, platforms (iOS, Android, web), app versions or kinds of broadcaster such as individual creators, Pages and Groups?
- Did overall Facebook usage, or usage of other video formats in the app, change over the same period?
### What a Strong Answer Covers
- A precise definition of "usage" and a check that the drop is not a logging, definition or pipeline artifact.
- The size, timing and shape of the drop against seasonality and normal variation, and whether it is specific to Live or follows Facebook as a whole.
- Segmentation and a decomposition across broadcasters and viewers that localize where in the Live experience the decline happens.
- Internal and external hypotheses tested in a sensible order, each with the evidence that would confirm or rule it out.
- A clear next step once the cause is found, including whether the decline needs a fix.
### Follow-up Questions
- The number of broadcasts is flat, but viewers per broadcast fell, and only on Android. What do you check next?
- Live watch time fell, while watch time on short-form video in the same app rose by about the same amount. Is that a problem, and what would you recommend?
- Once you have found two causes, how would you estimate how much of the 20% each one explains?
Overview: A product manager interview question about Facebook Live, where usage has dropped 20% and the candidate must find the root cause. It tests defining an ambiguous usage metric, validating the data, segmenting the decline, and ranking internal and external causes across broadcasters and viewers.