Ace Market Sizing Interviews
590 words · Reviewed for accuracy
"How many piano tuners are in Chicago?" "Estimate the market for electric scooters in Germany." Market sizing questions feel absurd until you realise what they're testing: not whether you know the answer (nobody does), but whether you can build a structured, assumption-driven estimate out of thin air and defend every step. It's arithmetic plus composure.
The core skill: decompose an unanswerable question into a chain of small, estimable numbers, multiply them together, and sanity-check the result. The interviewer grades your structure and your assumptions, never the final digit.
The five-step method
- Clarify the question. "Are we sizing revenue or units? Annual? Which geography?" Market sizing prompts are deliberately vague — pinning them down is part of the test.
- Pick a structure: top-down or bottom-up. Top-down starts from population and filters down; bottom-up starts from one unit (one store, one customer) and scales up. Say which and why.
- Draw the tree. Break the market into 3–5 multiplicative factors. Fewer than three is hand-waving; more than five is a mess.
- Fill in round numbers. Choose figures you can compute with — 320 million, not 331.9 million — and label every one as an assumption out loud.
- Sanity-check. Compare the result against something you can anchor to. Off by 10x? Find which factor did it.
A worked example, narrated properly
Prompt: "Estimate annual revenue from coffee shops in a mid-sized US city." Everything below is a labelled assumption for the exercise. Bottom-up feels right, so: assume the city has 500,000 residents. Assume 60% drink coffee and, of those, a third buy from a shop on a typical day — that's 500,000 × 0.6 × 0.33 ≈ 100,000 daily transactions. Assume an average ticket of $5. That's $500,000 a day; assume shops operate 350 days a year and you land near $175 million annually. Now the sanity check: that implies roughly one daily coffee-buyer per five residents — plausible for a commuting city? Defensible? Say why or adjust the factors. The check is worth as many points as the math.
Top-down vs bottom-up: choosing on purpose
| Approach | Best when | Risk |
|---|---|---|
| Top-down (population → segments → share) | Consumer markets, broad geographies | Segment percentages multiply your errors |
| Bottom-up (one unit → scale) | Location-bound businesses (shops, gyms, chargers) | The "typical unit" may not be typical |
Strong candidates offer both: "I'll go bottom-up, then cross-check top-down from population." Two independent paths to the same ballpark is how consultants build confidence in an estimate. And if the two disagree wildly, don't average them and move on — find which assumption is doing the damage. That reconciliation is where the real insight lives.
Common mistakes
- Jumping to numbers with no structure. A tree with shaky leaves beats confident guessing every time.
- Precision theatre. Carrying 7.3% through three multiplications signals you don't know what an estimate is for.
- Skipping the sanity check — or worse, doing it and ignoring a 10x miss.
- Freezing on an unknown figure. Estimate it, label it, flag it for follow-up: "I'd want real data here, but let's proceed."
FAQ
How close does my answer need to be? Within an order of magnitude with clean logic is a pass; a precise wrong number with no structure is not. Structure is the product.
Will I get these in tech interviews too? Yes — product manager and strategy roles love them, and the same decomposition skill powers capacity estimation in system design rounds.
Market sizing is one slice of case prep — pair it with the case study guide and rehearse out loud with Aissence practice sessions.