Market sizing questions: the method, US numbers and a worked example

A market sizing question, also called a guesstimate or estimation question, asks for a number with no data given: how many, how much, how often. Consulting, product management, strategy and some finance interviews use them because they show, in five minutes, whether you can structure a problem, make reasonable assumptions and do clean math under pressure. The interviewer cares about your logic far more than the exact number.

Choose the side where you have better anchors. Use the other side, briefly, as your sanity check.

1. 1

   Clarify the scope

   Units or revenue? New sales or the installed base? Which geography and time period? Write the definition down.

2. 2

   Lay out the structure

   Say the equation before any numbers: "Market = adults × share who buy × purchases per year." Segment the biggest driver (by age, income or region).

3. 3

   Assume, out loud

   Give each assumption a number and a one-line reason. Round to one or two significant figures.

4. 4

   Calculate

   Multiply through step by step, keeping units attached. Say intermediate results.

5. 5

   Sanity-check

   Compare against a per-person figure, a second method or a number you know.

6. 6

   Conclude

   State the answer, the one or two assumptions that move it most, and what it means for the client.

Scope: new pairs bought by US consumers in a year, all channels, all ages. We go top-down and segment by age, because children outgrow clothes and adults replace them. Every assumption below is illustrative, chosen to show the method, not taken from industry data.

About one new pair a year for most people, and a bit more for growing children, feels right. The answer is most sensitive to the adult replacement rate: at one pair a year instead of 1.5, the total falls to about 346 million. Say that out loud. It shows the interviewer you know which assumption matters.

You do not need many. These cover most US questions. Each is from a government or research source, with the date it describes.

• Per person or per household. Divide your total by 342 million people or 135 million households. Does the result feel plausible for one person?
• Share of wallet. If your estimate implies a household spends $5,000 a year on coffee, something is off: that is about 6% of median household income.
• A second method. Rebuild the number from the supply side: stores × sales per store.
• A known anchor. Compare with a figure you know, such as the number of licensed drivers (240 million) for anything car-related.

• Starting the math before defining the unit, then answering a different question.
• Assuming without saying why. Every number needs a one-line reason.
• Not segmenting the biggest driver. Usage differs by age, income and region; one average hides that.
• False precision. "453,712,000" signals you do not understand estimation. Say "about 450 million".
• Skipping the sanity check, which is where most arithmetic slips are caught.