Estimation question

Estimate the number of videos uploaded to Youtube per day.

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

Fermi estimation using a top-down channel or creator-based approach.

State you'll estimate via number of active uploading channels and average uploads per channel per day.
Estimate active channels that upload regularly, e.g. tens of millions globally, though only a fraction upload daily.
Assume a low average uploads-per-channel-per-day when averaged across all channels, e.g. a small fraction.
Multiply to get a daily total.
Sanity-check against general knowledge that the platform's scale is measured in hundreds of hours uploaded per minute.
State a wide, clearly-labeled range.

Builds from channel count and per-channel upload frequency rather than guessing a raw daily total.
States the low per-channel-per-day upload rate explicitly, since most channels don't upload daily.
Cross-checks against the well-known 'hours uploaded per minute' framing to sanity-check scale.
Presents a wide, clearly-labeled range given the huge uncertainty in a global platform estimate.

Assuming a high average upload rate per channel, ignoring that most channels are inactive most days.
Giving a single overprecise number instead of an order-of-magnitude range.

Chapters of the AI PM course that teach what this question tests.