Estimating Average and Peak Concurrent Facebook Users Worldwide
Company: Amazon
Role: Product Manager
Category: Product / Decision Making
Difficulty: medium
Interview Round: Other
Estimate the number of people using Facebook at a single moment in time.
```hint Turn a day into a minute
Start from a daily figure you can reason about, then ask what fraction of the day the typical user actually spends in the product.
```
### Clarifying Questions
- Does "on Facebook" mean actively using the app or website in that minute, or anyone signed in with a live session?
- Should Messenger count as Facebook, with Instagram and WhatsApp left out?
- Do you want a typical moment across the day, the busiest moment, or both?
- Is this worldwide, and should one person on two devices count once?
### What a Strong Answer Covers
- A precise definition of being "on" Facebook, and why that choice changes the answer by orders of magnitude
- An equation linking daily users, time spent per user and the minutes in a day
- Round-number assumptions labeled as assumptions, with arithmetic that is easy to follow
- How time zones and daily usage patterns turn an average into a peak and a trough
- A sanity check, and how sensitive the result is to the weakest assumption
### Follow-up Questions
- If you were sizing servers for the busiest minute of the year, how would your number and your safety margin change?
- Which assumption would you check first with internal data, and what signal in the logs would measure it?
- How would the estimate change for the United States alone at 9 p.m. Eastern time?
Overview: A product manager estimation question asking how many people are actively using Facebook at any single moment. It tests defining the metric precisely, building an equation from daily users and time spent, labeling round-number assumptions, reasoning about peak versus average across time zones, and sanity-checking the result.
