You have reached a number for your market size. How do you check that it is reasonable?

Run three checks. Shrink it to one household or store and ask if that looks right, rebuild it by an independent route, and compare it with a hard ceiling like total spending. If a check fails, fix that input instead of averaging.

See this answer go from 50 to 90

Example answer

50/100

I'd step back and ask whether the number makes sense. I'd compare it to something I know and go over my math again.

Say I've sized US haircuts at $66 billion a year. $66 billion over 130 million households is about $510 a year, or about $42 a month. That's the size of a phone bill, not a rent payment, which is easy to believe for a home of two or three people.

If something looked off, I'd go back and find the assumption that caused the gap.

What holds it back

"Ask whether it makes sense" and "go over my math" name no test. The one real check, dollars per household, is good, but nobody knows where the $66 billion came from.

The numbers, step by step

An example estimate of US haircuts at $66 billion a year, tested per household, rebuilt from the number of stylists, and set against a ceiling.

Cuts a year330 million people × 8
2.64 billion
The estimate2.64 billion × $25
$66 billion
Per household a year$66 billion ÷ 130 million
about $510
Per household a month$510 ÷ 12
about $42
Stylists330 million ÷ 10,000 × 25
825,000
Second route, cuts a year825,000 × 12 a day × 250 days
about 2.5 billion
Second route, dollars825,000 × 3,000 cuts × $25
about $62 billion
Revenue per stylist3,000 × $25
$75,000
Ceiling330 million × 12 × $100
$396 billion

What the interviewer listens for

  • Do you name specific tests, not just say you'd see if it feels right?
  • Do you bring a huge number down to one person or one household?
  • Is your second route truly independent of the first, and do you test any input the two share?
  • Do you know an upper bound the answer cannot pass?
  • Do you say what you'd do when a test fails?

Common mistakes

  • Saying "that seems reasonable" with nothing to compare against.A number in the billions feels neither right nor wrong until it's divided down to a household.
  • Checking by repeating the same chain a second time.That catches a slip in arithmetic but not a wrong assumption, because both runs share it.
  • Running a per-person check that just divides back to an input.If you assumed 8 cuts a person, getting 8 cuts a person proves nothing.
  • Averaging two estimates that disagree.The gap is telling you an assumption is off, and the job is to find which one.

More market sizing

Consulting interview questionsAll interview questionsMarket sizing practice

Ready for an interviewer that talks back?

Start the 7-day free trial and run your first full case with the AI interviewer today.

Start 7-day free trial