By Night Watcher · Examples are fictional and illustrate a research method.
Decide what the number is for
A market size answers a specific question. Is the market big enough to be worth a year of work? Which of two segments should you test first? What revenue could a small team plausibly reach in three years? Write the question down before you look for numbers, because it decides which customers you count and over what period.
Use three figures. TAM, the total addressable market, is the yearly revenue if every possible customer bought your product. SAM, the serviceable addressable market, is the part you can actually serve with your product, region and sales channels. SOM, the serviceable obtainable market, is the part of SAM you can realistically win in your planning period. Each is a revenue figure for one year, so they can be compared with each other and with your plan.
Count customers from the bottom up
The bottom-up method multiplies two things you can check: how many customers could buy, and what each would pay in a year. Count your buyer directly. For businesses, government counts of establishments by industry are a common starting point. Professional associations, licensing bodies, marketplace or platform listings and your own list of qualifying accounts are others. For consumers, start from the group that has the problem, not the whole population.
Price with what a customer would actually pay per year, not your highest list price. If you have not set a price, use a range and size the market at each end. Write the source, date and region next to every input. A count of businesses in one country is not a global figure, and a count of establishments is not a count of buyers if one owner runs several locations.
Narrow to what you can serve
SAM removes the customers you cannot reach or serve. Common reasons are language and region, a required integration you do not have, company size outside what your product handles, regulation you cannot yet meet, or a sales motion you cannot run, such as enterprise procurement for a solo founder.
Express SAM as a share of TAM and write down why each group is excluded. That list is useful later: each exclusion is a possible expansion, and you can see what it would take to add it.
Estimate what you can win from capacity
SOM is where most market sizes go wrong. Picking a round share, such as 1% of a very large market, skips the question of how you would actually get those customers. Work it out from capacity instead: how many qualified conversations your channels produce each month, how many of those become customers, and how many months are in your plan.
Divide the customers you can realistically sign by the customers in SAM to get your win share. If that share looks tiny, the market may still be worth it. If it looks large, check whether your channel assumptions are realistic before trusting the result. Churn matters too: customers you lose during the period do not count towards the SOM at the end of it.
Worked example: a scheduling tool for physiotherapy clinics
Suppose, as a fictional example, a founder finds 40,000 physiotherapy clinics in the regions they can sell to, and expects to charge about $1,800 a year. TAM is 40,000 × $1,800 = $72,000,000 a year.
The product only integrates with the two most common clinic-management systems, which the founder estimates cover 45% of clinics. SAM is $32,400,000. With one founder doing sales, they expect about 25 qualified demos a month and to close one in five, which is 5 new clinics a month, or 180 over three years before churn. That is 1% of the 18,000 clinics in SAM, a SOM of $324,000 a year before churn. The figure is small next to the TAM, and it is the one to plan around.
Cross-check with a top-down figure, and keep both
A top-down estimate starts from a published industry total and takes a share of it. It is quick, but you inherit someone else's definitions: what counts as the industry, which regions, which year, and whether the figure is spending, revenue or something else. Use it as a sanity check on your bottom-up TAM rather than as the answer.
If the two disagree by a wide margin, look for a difference in definitions before deciding one is wrong. Keep both figures, their sources and their dates. Do not average them into a single number.
Mistakes to avoid
Mixing units: searches, website visits, app downloads and paying customers measure different things and cannot be substituted for one another. Search demand is useful evidence that people are looking for a solution; it is not a customer count.
Counting everyone who could conceivably benefit, sizing only the most optimistic price, and presenting TAM as if it were reachable revenue all make a market look bigger than the business you can build in it. The TAM SAM SOM calculator shows all three figures side by side, with the customers your SOM implies, so the gap stays visible.
Your worksheet
Copy these prompts into your own notes. No signup is required.
Market-size worksheet Question this number answers: Buyer definition: Customer count, source, date and region: Annual price per customer (low / expected): TAM (customers × price): Groups excluded from SAM and why: SAM share and SAM: Qualified conversations per month: Close rate: Months in plan and expected churn: Customers you can realistically win: Win share and SOM: Top-down cross-check, source and date: Biggest assumption to test next:Download the text worksheet
Test your assumptions
Use the related free calculator, then compare the scenario with evidence from actual buyers.
For company and search-demand research, explore Night Watcher plans.