Estimation question

Estimate the daily watch time of cooking videos on YouTube.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests structured top-down estimation, decomposing a global platform number into population, adoption and duration assumptions.

Frame the formula: daily watch time equals YouTube's daily active viewers of cooking content times the average minutes watched per viewer.
Assume YouTube has roughly 2 billion monthly active users, with daily actives around 40%, roughly 800 million daily active users.
Assume 10% of daily actives watch some cooking content on a given day, roughly 80 million viewers, since cooking is a popular but not universal category.
Assume an average of 15 minutes of cooking content watched per viewer on days they watch any.
Multiply: 80 million viewers times 15 minutes gives roughly 1.2 billion minutes, about 20 million hours, of daily cooking-video watch time, stated as an assumption-based estimate.

Builds the estimate from YouTube's known rough scale, 2 billion MAU, rather than guessing a number from nothing.
States every intermediate assumption, daily active share, cooking-viewer share, minutes per viewer, clearly and separately.
Shows the full multiplication so the final number can be sanity-checked or challenged step by step.

Jumping to a final number without decomposing it into population and adoption assumptions.
Not stating that these are assumptions, presenting the estimate as if it were a known fact.

How would you adjust this for regional differences in cooking-content popularity?
How would you sanity-check this against YouTube's public total watch-time disclosures?