Estimation Questions and How to Structure the Answer Out Loud

Estimation questions are scored on stated assumptions and a narrated chain of factors, not on how close the final number lands to the true figure.

Estimation questions in trading interviews are scored on the structure of the reasoning, not on how close the final number lands to the true figure, and structuring the answer out loud means stating assumptions before calculating, breaking the problem into a short chain of multiplications or divisions, and checking the result against a rough bound before presenting it as final. An interviewer asking you to estimate a quantity you cannot look up wants to hear each assumption named as you make it, since a silent calculation followed by a single number gives them nothing to evaluate.

State the assumptions first, out loud, before any arithmetic begins. Break the quantity into a short chain of factors that are each individually easier to estimate than the original question. Multiply or divide the chain step by step, narrating each intermediate result rather than going silent until the end. Check the final number against an independent rough bound, and adjust it if the two disagree by an order of magnitude.

The order matters because an interviewer who hears the assumptions first can interrupt and correct one that is clearly wrong before the rest of the calculation is built on it. Presenting the final number first and the reasoning afterward removes that opportunity and makes a single bad assumption invisible until the interviewer asks where it came from.

A classic version of this question asks for the number of piano tuners working in a city of a given size. Assume a population of 2,700,000 and roughly 2.5 people per household, giving about 1,080,000 households. Assume 1 household in 20 owns a piano that gets tuned, giving 54,000 pianos. Assume each piano is tuned once a year, so 54,000 tunings are needed annually.

Assume a tuner works 250 days a year and completes 4 tunings a day, giving 1,000 tunings per tuner per year. Dividing the 54,000 annual tunings needed by the 1,000 a single tuner can complete gives an estimate of 54 piano tuners in the city. The exact figure is not the point: every assumption was stated, the chain of five factors was narrated, and the result is defensible even if any single assumption is later challenged and revised.