PartnerRoundMarket sizing and estimation
Structure beats precision. Nobody checks your answer against reality: they check your path.
A sizing question asks you to estimate something unknowable from things you do know. The grade comes almost entirely from the path: a clear equation, stated assumptions, clean arithmetic, and a sanity check.
- 1
Write the equation first
Say the whole chain before computing anything: 'Population, times households, times penetration, times replacement rate.'
- 2
Choose top-down or bottom-up deliberately
Top-down from a population; bottom-up from a unit (one store, one plane, one customer). Name which you picked and why.
- 3
Use round anchors
330m US population, 2.5 people per household, 80 years of life. Round numbers you can defend beat precise ones you cannot.
- 4
Segment where it matters
Split only where the rate genuinely differs: urban vs rural, business vs consumer. Needless segmentation burns clock.
- 5
Sanity-check the result
'That gives $12bn, which is roughly a tenth of the US coffee market: plausible for this niche.' This sentence earns more than the arithmetic did.
Number of passenger lifts (elevators) in the UK
Estimate how many passenger lifts are installed in the UK.
Strong
'I'll go bottom-up by building type. UK population 67m, ~2.4 per household gives ~28m homes; say 20% are flats in blocks averaging 20 units, so ~280k blocks, and say 60% have a lift, ~170k. For commercial: ~30m workers at ~15 per office floor… I'll size offices next, then add hotels and hospitals. Let me total and sanity-check against lift-maintenance industry revenue.'
Weak
'Maybe a few hundred thousand? There are a lot of buildings in London and other cities, so possibly around 500,000: it's hard to say without data.'
'I'll assume 60% of blocks have a lift: happy to flex that.' This invites the interviewer to correct an input rather than mark down the whole answer, and it is exactly how a consultant handles an unknown in front of a client.