Estimation question

Calculate the number of queries answered by Google per second.

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 estimation: breaking a huge number into defensible population and behavior assumptions rather than guessing a final figure.

State the funnel: world population, internet users, then daily Google searchers, then searches per user per day.
Assume roughly 5 billion internet users globally, and that Google captures the large majority of search traffic, say 90%.
Assume an active searcher does about 5 to 10 searches a day on average, factoring in casual and heavy users.
Multiply: 5 billion times 0.9 times ~7 searches equals about 31.5 billion searches per day.
Convert to per second: 31.5 billion divided by 86,400 seconds is roughly 365,000 queries per second.

Shows the full chain of assumptions explicitly rather than jumping to a memorized number, which is what this question is testing.
Labels every assumption clearly, e.g. 'assume 90% search-engine market share for Google', so the interviewer can push back on a specific link.
Sanity-checks the result against a known public figure (Google has publicly cited roughly 8.5 billion searches per day) and reconciles any gap.
Adjusts for daily and seasonal patterns, noting peak QPS during work hours is higher than the 24-hour average.

Quoting a memorized number without showing the derivation, which defeats the purpose of an estimation question.
Forgetting to convert from per-day to per-second, or using the wrong number of seconds in a day.

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