TAM, SAM, SOM: the number nobody checks

Every deck has the slide. Three circles, a big number on the outside, a smaller one inside, and a modest slice at the centre labelled "our target." It's been in every deck for fifteen years, and roughly nobody has ever asked where the outer number came from.

I'll tell you where: a headline from a market research report, found in the first ten seconds of googling, published by someone whose business model is selling that report for $4,000.

The number is decoration until someone asks one question

The question is: "what does the person paying you actually pay for today?"

If the market sizing was done top-down — take an industry figure, multiply by a percentage that feels reasonable — the answer collapses instantly. You end up defending a number you didn't build, sourced from a study you didn't read, about a market that isn't quite yours.

Bottom-up survives that question. It's slower, uglier, and it holds: how many people have the problem, how many of them are reachable through channels you can actually buy, what they pay now for the thing they use instead.

Tam sam som, in plain words

• TAM — everyone who could conceivably have this problem. Useful as context, useless as a target.
• SAM — the slice you can actually serve: your geo, your segment, your language, your price point.
• SOM — what you can realistically take in the next 12–18 months, given your channels and budget.

If those three numbers don't shrink dramatically between steps, you haven't done market sizing — you've done wishful arithmetic. (We keep the short definitions in the glossary if you want them handy.)

How to calculate TAM without lying to yourself

The version that takes an afternoon, not a quarter:

1. Count the demand you can observe. Search volumes, community sizes, review counts on competing products, app installs. Real signals, not "the industry is worth $X billion."
2. Price it against reality. What do the alternatives charge, and what do people actually pay — including the free thing they're using instead of paying.
3. Cut by reachability. Which of those people can you touch through a channel you can afford? That cut is usually brutal, and it's the whole point.
4. Sanity-check against a competitor. If a leader in the space is doing revenue X with their traffic, your slice has to make sense next to that.
5. Write the assumption next to the number. Every figure gets a source or a stated guess. Anything unlabelled is fiction by default.

When the number actually matters

Market sizing is decoration in most pitch decks and load-bearing in exactly two situations: choosing between markets, and choosing whether to enter one at all. That's when market entry strategy stops being a slide and becomes a spending decision — and when a wrong number costs a quarter of runway.

Market research for startups gets skipped because the proper version — six weeks, an agency, a five-figure invoice — is priced for companies that already have money. That's the actual reason founders copy a number off a chart: not laziness, arithmetic.

Doing it in hours instead of weeks

This is precisely the gap the product was built for. A single report covers demand, competitors, audience and channels with the sources attached — enough to size a market bottom-up and defend it. If the decision is about one specific country, geo research goes a level deeper: local demand, local competitors, local channels, in the local language.

Neither replaces judgement. Both replace the six weeks you don't have.

Bottom line

A market size you can't defend is a decoration you're paying for with credibility. Build it from demand you can observe, cut it by what you can reach, and write down every assumption. Then, when someone finally asks where the number came from, you get to answer.