Market Sizing Drill
Nobody in the room knows how many cups of coffee the UK sells in a day, the interviewer included. What is being graded is whether you can take a question with no available data and turn it into four or five factors a stranger could follow. So this drill makes you build the chain under a clock, then puts it next to a reference chain so you can see which term carried your error.
Six cases across consumer markets, physical counts and platform volumes. Landing within a factor of three is a good answer.
How many cups of coffee are sold in UK coffee shops on a typical day?
Answer in cups per day
Build the chain
You are not being asked for the answer. Nobody in the room knows it. You are being asked whether you can turn a question with no available data into four or five factors that each sound defensible on their own.
Start the clock, then add one factor at a time. Name it, give it a number, and pick whether it multiplies or divides. The running product sits at the top of the page. When you lock it in, your chain lands next to a reference chain and you can compare them term by term — which is where the learning is.
What good sounds like
- Say the units at every step. The single most common way a chain lands twelve times out is multiplying by a monthly figure and then by twelve months, or dividing a stock by a rate. If you narrate the units, you catch it live.
- Anchor on something you genuinely know. Population, households, GDP, an average salary. Never anchor on a market size, because that is what you were asked to produce.
- Cross-check from the other direction. If you built it bottom-up from people, check it top-down from spend. Two routes agreeing is worth more than either route being precise.
- Flag your weakest term unprompted. Saying “the one I am least sure of is average spend, and if it is half my number the answer halves” is the sentence that separates the top decile.
- Round hard. 54m, not 54.3m. Precision you did not earn is a tell, and it slows the arithmetic down when the clock is running.