How to Calculate TAM, SAM and SOM (Bottom-Up)

Investors see the same slide dozens of times a week: three concentric circles, a billion on the outside, a few million in the middle. Most of them skip straight past the numbers to the footnote. If the outer circle came from a report about “the global HR software market”, the slide has told them nothing.

For a founder, though, market sizing is not decoration. It answers a blunt question. If everything goes right, is this worth years of your life?

TAM, SAM and SOM in one paragraph each

TAM, total addressable market, is the annual revenue available if every customer who has the problem bought your product. You choose the boundary, a country, a language area or the world, but you must say which.

SAM, serviceable available market, is the part of TAM your product can actually serve today. An English-only tool that only integrates with UK payroll systems does not serve a German bakery, however much the bakery shares the problem.

SOM, serviceable obtainable market, is the slice of SAM you can realistically win over a stated period, usually three to five years. This is the number you will be held to.

Why bottom-up beats top-down

Top-down sizing starts with an analyst figure and carves off percentages. “The market is worth X, we only need two percent.” It sounds modest. It is usually meaningless, because the report’s definition of the market rarely matches what you sell, and the two percent is a guess wearing a suit.

Bottom-up sizing counts customers and multiplies by price. That is the entire formula: number of customers × annual revenue per customer. Its strength is that every assumption sits in plain view, where you and any sceptical reader can poke at it.

Top-down still has one use. If your bottom-up TAM comes out ten times larger than total industry spending, one of your inputs is wrong.

Worked example: a UK onboarding tool for employers

Say you are building software that handles new-hire paperwork and checklists. It is useless to sole traders, so your customers are businesses that employ people.

The UK government’s business population estimates give you a clean starting point. At the start of 2025 there were about 5.69 million private sector businesses, of which roughly 4.27 million had no employees. That leaves about 1.42 million employers.

Private sector businesses, UK

5,690,265

minus businesses with no employees

4,272,535

Assumed price per customer per year

£600

TAM: 1,417,730 employers × £600

£851m

Firms with 10 to 249 employees (220,085 + 38,435)

258,520

SAM: 258,520 × £600

£155m

Assumed market share after five years: 1%

2,585 customers

SOM: 2,585 × £600

about £1.55m a year

Business counts from the UK Business Population Estimates 2025 (start of 2025). Price and market share are assumptions for this example.

Why cut SAM to firms with ten to 249 staff? Because in this example the product only earns its price once onboarding happens more than a few times a year, and the largest firms already run enterprise HR suites. Your own reasoning will differ. What matters is that it is written down.

The result is sobering and useful. Around £1.55 million in annual revenue after five years is a solid business for a small, profitable team. For a venture fund hoping for a company with nine-figure revenue, it is too small unless you can show a credible route into further countries or products.

Sanity-check SOM against your sales capacity

“One percent market share” is an assumption that feels safe and means little. The better check is how many customers your go-to-market can physically close.

Here is a rough version with assumed inputs. One salesperson runs 15 first calls a week and closes one in five. That is three customers a week, or about 135 across 45 working weeks. With a second salesperson from year two, you reach roughly 1,200 customers after five years, before churn.

That lands at under half the 2,585 from the market-share route. When the two methods disagree this much, trust the capacity number and either lower your SOM or explain the channel that closes the gap: self-serve sign-ups, partners, resellers. Each of those comes with its own assumptions to test.

Marketplaces and consumer apps need different units

Customers times price fits subscription software neatly. For other models you swap the building blocks and keep the logic.

Marketplaces earn a take rate. Their TAM is the annual transaction volume in your segment multiplied by your commission, not the volume itself. A platform matching households with cleaners would count households that buy cleaning, times annual spend, times take rate. Quoting gross volume as TAM is a common way to inflate the slide tenfold.

Consumer apps usually count paying users. TAM equals people with the problem times average annual revenue per paying user. With freemium, only a minority pays, so the expected conversion to paid belongs in the model as an explicit assumption or a range.

Common sizing mistakes

The “one percent of a huge market” pitch tops the list. It says nothing about how you reach that one percent.

Confusing an industry’s revenue with its spending on your problem is close behind. Total UK hospitality turnover tells you almost nothing about the budget for rota software.

Pricing comes third. Founders plug in the price they hope for rather than what interviewees said they pay today. Your customer discovery interviews should give you that figure.

Last, time. Any SOM without a year attached is not a forecast.

When market size matters less

For a bootstrapped company, a small market can be ideal: fewer competitors, higher prices, customers who know each other and refer. Whether you are building that kind of business or a venture-scale one also shapes your company structure and funding route, which our section on incorporation and funding covers.

The reverse holds too. Even a huge TAM does not rescue an idea that nobody responds to. The order from our guide on how to validate a business idea still applies: problem first, willingness to pay second, market size third.

Most sizing work fits in a spreadsheet. If you end up combining larger datasets, such as company registers with scraped competitor data, that is the kind of analytics work studios like OnlineBuilders build tooling for.

Next step: write down the four inputs of your model, customer count, price, serviceable share and market share, and next to each note where it came from. Every line that says “guess” is your next research task.

Questions people ask

What is the difference between TAM, SAM and SOM?

TAM is total annual revenue if every possible customer bought from you. SAM is the part your product and go-to-market can actually reach. SOM is the share you can realistically win in the next few years.

Should I use top-down or bottom-up market sizing?

Bottom-up, built from customer counts and prices, is far more convincing and more useful to you. Top-down figures from analyst reports work as a cross-check at most, because their market definitions rarely match your product.

What TAM do investors want to see?

Venture funds look for markets where one company could become very large, so they want a credible path to a big outcome. A profitable niche business can thrive on a modest SOM. Either way, the derivation matters more than the headline number.

Where can I find reliable numbers?

National statistics offices publish business counts by size and sector: the ONS and the Department for Business and Trade in the UK, the Census Bureau in the US, Eurostat for the EU. Prices come from your interviews and from competitors who publish theirs.