HOW TO ANSWER ESTIMATION QUESTIONS IN QUANT INTERVIEWS

You are not expected to know the exact number. The interviewer wants to see how you build and check a rough estimate.

UPDATED 2026-08-12 · 6 MIN READ · BY THE VARIANCE TEAM

The five-step framework

  1. Define the target: clarify the geography, time period, and unit.
  2. Build an equation: split the target into quantities you can estimate.
  3. State assumptions: choose round, defensible numbers and say them aloud.
  4. Calculate cleanly: keep units attached and use scientific notation when helpful.
  5. Sanity-check: test whether the result is plausible and name the most sensitive assumption.

Worked example: daily coffees sold in a city

Suppose the interviewer asks: “How many takeaway coffees are sold in a city of one million people each day?” First define the target as takeaway cups sold within the city on a typical weekday.

Build the estimate from demand: population × adult share × coffee-drinker share × takeaway purchases per drinker.

  • 1,000,000 residents
  • 75% adults
  • 60% drink coffee
  • About 1/3 buy a takeaway coffee on a typical day

That gives 1,000,000 × 0.75 × 0.60 × 1/3 ≈ 150,000takeaway coffees per day.

Cross-check from the supply side

Now estimate another way. If the city has roughly 1,000 coffee-selling locations and each sells 150 takeaway cups per day, the supply-side estimate is also 150,000. The agreement does not prove the answer, but it shows the assumptions are mutually plausible.

What the interviewer is listening for

  • A clear equation before arithmetic begins.
  • Assumptions that are explicit and easy to revise.
  • Correct units and order of magnitude.
  • A quick sensitivity check: which assumption moves the answer most?
  • Comfort correcting course when challenged.

Common mistakes

Do not hunt for a secretly exact answer, bury the interviewer in tiny demographic segments, or use precise-looking assumptions you cannot defend. Two or three meaningful drivers with round numbers usually reveal more judgment than ten fragile assumptions. Finish with a range. Here, 100,000 to 250,000 is defensible. Then say which assumption could push the estimate toward either end.