Estimation questions show up in PM interviews at Google, Amazon, Microsoft, and many other companies. They sound intimidating, but they follow a predictable structure once you know the approach. The goal is never to get the exact right answer. Interviewers want to see how you think.

Here is a complete guide to nailing estimation questions in PM interviews.

Why Companies Ask Estimation Questions

Estimation questions test several things at once: your ability to structure a complex problem, your comfort with quantitative reasoning, the quality of your assumptions, and your ability to sanity-check results. These are the same skills you use every day as a PM when sizing markets, forecasting adoption, or estimating engineering effort.

The Three-Step Framework

Every estimation question can be answered using the same three-step approach: Break it down, estimate each component, and sanity-check.

Step 1: Break It Down

Take the big question and decompose it into smaller, estimable components. The key is choosing a decomposition that uses numbers you can reasonably estimate.

For example, "How many Google Searches happen per day?" can be broken down as: Number of internet users worldwide multiplied by the percentage who use Google multiplied by the average number of searches per user per day.

Each of these sub-estimates is something you can make a reasonable guess about, even if you do not know the exact number.

Step 2: Estimate Each Component

For each component, state your assumption clearly, explain your reasoning briefly, and give a number. Do not agonize over precision. Round numbers are fine and actually preferred because they make the math easier to follow.

For the Google Search example: "There are roughly 5 billion internet users worldwide. About 90% use Google as their primary search engine, so 4.5 billion Google users. The average user probably does about 3 to 4 searches per day. Let me use 3.5. So: 4.5 billion times 3.5 equals roughly 15.7 billion searches per day."

Step 3: Sanity-Check

After arriving at your answer, take a moment to check whether it makes sense. Compare it to reference points you know. "Google has publicly reported processing over 8.5 billion searches per day. My estimate of 15.7 billion is in the right order of magnitude, though a bit high, probably because not all internet users search every day. If I adjust the daily search rate to 2, I get about 9 billion, which is very close to the reported number."

This sanity-check step is where candidates really impress interviewers. It shows intellectual honesty and the ability to calibrate your thinking.

Common Estimation Question Types

Market sizing: "How big is the US market for electric scooters?" "How much revenue does Starbucks generate per store per day?"

Cost estimation: "How much does it cost Zoom to host an average video call?" "Estimate the storage cost of all photos on Instagram."

Usage estimation: "How many messages does WhatsApp process per day?" "How many rides does Uber complete per day in New York?"

Revenue estimation: "Estimate Apple Music's annual subscription revenue in the US." "How much does Google earn from YouTube ads per year?"

Tips for Standing Out

Talk through your reasoning out loud. The interviewer cannot evaluate your thinking if you do the math silently.

State your assumptions explicitly. "I am assuming the average American household has 2.5 people" is much better than just using 2.5 without context.

Use round numbers. Nobody expects you to multiply 4,327 by 6,891 in your head. Round to 4,000 and 7,000.

Show range awareness. After your calculation, mention that the true answer could be 2x higher or lower depending on which assumptions are off. This shows maturity.

Do not panic if the interviewer challenges an assumption. They are not trying to catch you. They want to see how you adapt. Adjust the number, redo the math, and move on.

Product Alliance's Hacking the PM Interview course includes a dedicated estimation section with video walkthroughs of real estimation questions, step-by-step frameworks, and practice problems. It is the most efficient way to build the quantitative muscle that PM interviews demand.

SPECIFIC COMPANY DEEP DIVES