How to size a market for a startup

You cannot count "small businesses". You can count "UK-registered companies with 2–20 employees in professional services". The difference between those two sentences is the difference between a real estimate and a decorative one.

Write your customer definition with four filters:

• Who they are — role, company type, or a demographic that appears in some published dataset.
• Where — the countries you can actually sell and support in, in a language you speak.
• What makes them qualified — the condition that means they have the problem. "Runs paid ads" or "invoices more than 20 clients a year".
• What makes them reachable — do they gather anywhere, appear in any directory, or hold any searchable job title?

That last filter is the one founders leave out and it's the one that decides whether the number means anything. A market of two million people you have no way to reach is smaller in practice than a market of 8,000 who all read the same newsletter.

Search the phrase "how many X are there" and you'll get a number from a press release quoting a report nobody has read. Better sources, roughly in order of how much you should trust them:

Government statistics. Business counts by industry and size band, published free in most countries — the ONS in the UK, the Census Bureau's County Business Patterns in the US, Eurostat across the EU. Usually the best number available for B2B.

Platform counts you can observe. LinkedIn's result count for a job title filtered to a country. App store category counts. A marketplace's published seller number. Imperfect and observable, which beats precise and asserted.

Public competitor disclosures. A listed competitor's customer count and revenue gives you market size and average price in one document. Annual reports and S-1s are free and unusually candid about market definition.

Association membership. Trade bodies publish member counts, and their members are by definition reachable.

Write down the source and the date next to every number. Six months from now you'll need to know whether it was 2019 data, and so will whoever you show it to.

Bottom-up is the version that survives questioning. Chain the numbers:

Reachable people × the fraction with the problem × the fraction who'd switch × annual price.

Example: 340,000 UK freelance designers (ONS) × 30% who invoice enough clients to feel the pain (from your interviews) × 5% who'd switch tools in a given year (industry churn norms) × £180/year = about £918k of first-year addressable revenue.

That's small, specific, and defensible. Every factor is something you can be challenged on and answer.

Top-down is the sanity check, not the pitch. Take a published market-size figure, cut it to your geography and segment, and see whether it lands anywhere near the bottom-up number.

Within 2–3×: good, you have something real. 50× apart: one of them is wrong, and it's usually the top-down report measuring a much broader category than the one you're in. Reconciling the gap is where the actual understanding happens — do it before anyone else does it for you.

Use the calculator below to run the arithmetic and, more importantly, to print your assumptions back as a sentence you can put on a slide.

"Our TAM is $2.4 billion" invites exactly one question — where did that come from — and you'll have nothing.

"3.4 million UK SMEs (ONS 2024), 12% have this problem based on 40 interviews, £600 a year, 3% obtainable in three years" answers that question before it's asked, and moves the conversation to which assumption you're least sure about. That's the conversation worth having.

Three rules for presenting it:

• Show the chain. Every multiplication visible.
• Give a range, not a point. "Between £900k and £2.1m depending on switching rate" is more credible than a single figure, because it demonstrates you know which factor is uncertain.
• Name the weakest link yourself. Saying "the switching rate is our softest assumption and here's how we'd test it" builds more confidence than any number you could present.

And if the honest bottom-up number is small, say it and explain why it's still worth doing. A £900k addressable market is a fine business for one person. Inflating it to sound venture-scale converts a good small business into a bad big one on paper, and experienced readers spot it instantly.