A PM at Meta asked you to describe the distribution of daily minutes spent on Facebook per user. How would you describe it?

First, it is critical to clarify whether we’re measuring active time (e.g., clicks, scrolls) versus idle time (e.g., background tab open), and whether Daily Active Users (DAUs) are included.

The distribution of daily minutes spent on Facebook per user is likely a right-skewed (positively skewed) distribution. This means that a large number of users spend a relatively short amount of time on the platform daily, while a smaller portion of users spend a much longer time, creating a tail that extends to the right. There may also be multiple peaks (i.e., multimodal distribution), representing different user engagement types, such as casual versus highly engaged users.

Metrics to describe this distribution would include:

• Mean: Average time spent, sensitive to outliers.
• Median: More robust to skewness.
• Mode: Helps identify the most frequent behavior.
• Standard deviation and variance: Measures of spread.
• Percentiles (25th, 50th, 75th): Help capture distribution shape.
• Kurtosis: Indicates the presence of heavy tails or outliers.