Root-causing a steep weekly decline in Facebook daily active users
Company: Meta
Role: Product Manager
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Other
You are a product manager at Meta. Facebook's daily active users (DAU) have dropped by 15% in each of the past three weeks. Find the cause: explain how you would investigate, which hypotheses you would check and in what order, and how you would confirm the root cause.
```hint Size it before you slice it
Before breaking the data down, work out what three consecutive weekly declines of this size add up to, and ask which kinds of causes could move a product of Facebook's scale that far that quickly.
```
### Clarifying Questions
- Is the 15% a week-over-week decline each week, or is each week 15% below the same baseline, such as the week before the drop began?
- How is a daily active user defined here: anyone who opens the app or website while logged in, or only users who post, react or comment?
- Is the decline worldwide and on every platform, or concentrated in certain countries, devices or app versions?
- Are monthly active users, and Meta's other apps such as Instagram, Messenger and WhatsApp, showing the same pattern?
### What a Strong Answer Covers
- The size and shape of the decline worked out first (cumulative effect, start date, sudden versus steady) and used to judge which causes are plausible
- A check that the drop is real, covering the metric definition, logging and the data pipeline
- Segmentation by cuts such as geography, platform, app version and user tenure, each with a stated reason
- Internal and external hypotheses, each paired with the specific data that would confirm or rule it out
- A way to confirm the root cause and size its share of the decline, plus what happens once it is found
### Follow-up Questions
- If the decline were concentrated in one app version on one operating system, what would you do in the first 24 hours?
- If Instagram's daily users rose while Facebook's fell, how would you tell whether people moved within Meta or left it?
- How would you tell a real loss of users from a change that only made fewer of them count as active?
Overview: A metric-diagnosis interview question in which Facebook's daily active users have fallen 15% in each of the past three weeks and the candidate must find out why. It tests sizing the decline, validating the data, segmenting the drop, ordering internal and external hypotheses and confirming the root cause.