Market sizing
How to Size a Market You Can Actually Reach
Most market-size numbers founders put in a deck are made up. Not because founders are careless, but because the top-down method — start with a global figure, cut it down — almost always over-counts for an early-stage company. This page is the bottom-up alternative: a method you can run in an afternoon, on a single spreadsheet, that produces a number you can defend in a customer interview and an investor meeting.
Last updated · September 8, 2026
Quick answer
How do you size a market you can actually reach?
Start from the buyer you can name — a specific persona, in a specific industry, in a specific geography — and count how many of those buyers are out there. Multiply by the realistic percentage you can reach in year one, then by the realistic revenue per buyer per year. That is your Serviceable Obtainable Market (SOM). The Total Addressable Market (TAM) and Serviceable Addressable Market (SAM) sit above it as theoretical ceilings; the number that matters for a startup is SOM, because SOM is the only number that constrains the next twelve months of decisions. The five-step bottom-up method below produces all three numbers from the same inputs.
Key takeaways
The page in 30 seconds
- TAM is the theoretical ceiling — every possible buyer of every possible alternative. SAM is the slice you can address with your business model. SOM is the slice you can realistically reach in the next year. The only number that matters for an early-stage decision is SOM.
- Top-down sizing — starting with a global market figure and cutting it down — almost always over-counts. The bottom-up method starts with the buyer you can name and counts outward.
- The five-step bottom-up method: name the buyer, count the buyers, apply a realistic reach rate, apply a realistic revenue per buyer, then sanity-check against public data.
- Every market-size number must answer four questions: which buyer, which geography, which price point, and which year. A number that does not answer all four is not a number; it is a guess.
- Common mistakes: doubling on every step, using national totals for a regional product, equating 'addressable' with 'reachable', and forgetting that the price point constrains the buyer as much as the buyer constrains the price.
- A market-size number is a hypothesis, not a fact. The same exercise produces a different number twelve months later, after the first customers have actually bought — and that revision is the most informative market research a startup does.
Definitions
What TAM, SAM, and SOM actually mean
The three-tier framework is the one used in McKinsey, BCG, and most investor-grade market models. The definitions vary slightly across sources, but the order of the tiers does not.
TAM — Total Addressable Market
The total annual revenue the market would generate if every potential buyer of every possible alternative bought from a single vendor at the prevailing price. TAM is a theoretical ceiling and is the number most often quoted in pitch decks, because it is the largest of the three. It is also the least useful for an early-stage decision, because no realistic startup addresses TAM in the first five years.
SAM — Serviceable Addressable Market
The slice of TAM you can actually address with your current product, business model, and go-to-market motion. A US-only SaaS for accounting firms does not address TAM for 'all accounting software' — it addresses SAM, the US accounting-firm segment. SAM is the number you can grow into over three to seven years with the same product line.
SOM — Serviceable Obtainable Market
The slice of SAM you can realistically reach in the next twelve months, given your distribution channels, your team's capacity, and your price point. SOM is the number that constrains the early-stage decision: how much revenue can the company plausibly book in year one, and is that number large enough to be worth the next six months?
The terms are not regulated. Different sources draw the lines in slightly different places — sometimes SAM and SOM are used interchangeably, and sometimes 'TAM' is used loosely to mean any of the three. The order is the same in every standard reference.
Top-down vs bottom-up
Why top-down sizing is almost always wrong for an early-stage startup
There are two ways to produce a market-size number. The first starts from a global figure and cuts down. The second starts from a buyer you can name and counts outward. They produce very different numbers, and for a startup the second is the one you can defend.
Top-down — start with a global figure
Take a global market research figure — for example, 'the global CRM market is $X billion' — and cut it down by geography, by vertical, by company size, and by product type until you reach the slice your startup addresses. The method is fast, and the TAM looks impressive. The problem is that the global figure is almost always wrong for the segment, and the cuts are guesses. Doubling a guess by a guess produces a guess with a power of two.
When it works
Bottom-up — start with the buyer you can name
Name the buyer — a specific persona, in a specific industry, in a specific geography. Count how many of those buyers exist (from a census, a registry, or a public dataset). Apply a realistic reach rate (the percentage you can plausibly reach in year one, given your channel). Multiply by the revenue per buyer per year at your price. The result is SOM, and it is the number you can defend.
When it works
Top-down produces a TAM that looks impressive. Bottom-up produces a SOM that you can defend. For an early-stage decision, the defendable number is the useful number.
The method
Bottom-up market sizing in five steps
Each step produces one input. The five inputs together produce TAM, SAM, and SOM in a single spreadsheet.
- 1
Name the buyer in one sentence
Write down exactly who buys. 'US accounting firms with 5–25 employees' is a buyer. 'Small businesses' is not. The sentence must specify the persona, the company type, the size band, and the geography. If you cannot write the sentence in one line, the buyer is not specific enough yet.
- 2
Count the buyers from a public source
Use a primary source to count the buyers. The US Census Bureau publishes establishment counts by NAICS code; the Bureau of Labor Statistics publishes industry employment counts; Eurostat publishes EU establishment counts; the Companies House (UK) and the State Administration for Market Regulation (China) publish equivalent datasets. The number must come from a primary source, not from a market-research aggregator that does not show its methodology.
- 3
Apply a realistic reach rate
Of the buyers who exist, what percentage can you plausibly reach in year one, given your distribution channel, your team's outbound capacity, and your existing audience? For a SaaS sold over SEO, the realistic reach rate in year one is usually well under 1% of the buyer population. For a SaaS sold over an existing B2B channel, it can be higher. Write down the rate and the reason.
- 4
Apply a realistic revenue per buyer
What does the buyer pay, per year, at your price point? The answer is constrained by the alternative they are using today (status quo, spreadsheet, agency, competitor). If the buyer is paying $200/month today, the price point is in that range; if the buyer is paying $0 today, the price point is constrained by the time the buyer saves. Use the lower end of the realistic range, not the upper end.
- 5
Sanity-check against public data
Cross-check the result against published sources — the Census Bureau, BLS, the Bureau of Economic Analysis, OECD, World Bank, or industry-specific regulators. If your number is an order of magnitude off from a published figure, the inputs are wrong. If your number is in the same order of magnitude, the inputs are defensible.
Worked example
US SaaS for accounting firms with 5–25 employees
Step 1: Buyer is 'US accounting firms with 5–25 employees'. Step 2: The US Census Bureau publishes establishment counts by NAICS code 541211 (Certified Public Accountants). The count of firms in the 5–25 employee band is the denominator. Step 3: Realistic reach rate in year one for a new SEO-led SaaS is ~0.3% of the addressable population (one new customer per ~330 firms). The numerator is that rate times the denominator. Step 4: Realistic revenue per buyer per year at a $1,500/year price point. Step 5: Cross-check against the Census Bureau's services-revenue totals for NAICS 5412 and the SBA's small-business statistics. The number should be within an order of magnitude of the published total, not within a factor of ten. The output is SOM — the year-one revenue ceiling the team can defend in a customer interview and an investor meeting. TAM and SAM sit above it as theoretical ceilings; the next step is to make the SOM number visible, and to revisit it every quarter once the first cohort of customers has actually bought.
Sanity checks
The four questions every market-size number must answer
A market-size number that does not answer all four is not a number — it is a guess with a dollar sign on it.
Which buyer?
The persona, the industry, the size band, and the geography. 'US accounting firms with 5–25 employees' is a buyer; 'small businesses' is not.
Which geography?
The countries, regions, or regulatory zones the product can be sold into. A US-only SaaS cannot address TAM for 'the global accounting market' in year one.
Which price point?
The annual revenue per buyer at the price the buyer actually pays, not the price the founder wishes the buyer paid. If the realistic price is lower than the modeled price, the modeled revenue is wrong.
Which year?
The year the number applies to. Market-size numbers change year over year; a number labeled '2026' cannot be used as a '2030' projection without an explicit growth assumption.
Common mistakes
Five ways founders produce a market-size number that does not survive a customer interview
Each of these is a way to produce a TAM that is large on a slide and indefensible in a meeting. None of them are reasons to ship the number; they are reasons to redo the math.
Doubling on every step
'Global CRM is $X, US is 1/3, mid-market is 1/4, our slice is 1/2 of that — so the market is $Y.' Four cuts, each one plausible, each one a guess. The product is a guess with a power of four. Replace one of the cuts with a primary-source count and the chain collapses.
Using national totals for a regional product
A SaaS that only sells in California cannot address a TAM derived from 'all US small businesses'. The geography cuts apply multiplicatively with the other cuts, and the cumulative effect is to shrink the addressable slice by an order of magnitude or more.
Equating 'addressable' with 'reachable'
'Addressable' means a buyer who could plausibly buy. 'Reachable' means a buyer the team can actually reach, in year one, with the channel they have. The gap between the two is usually one to two orders of magnitude, and it is the gap between SAM and SOM.
Forgetting that the price point constrains the buyer
A $50,000/year SaaS is not bought by the same buyer as a $500/year SaaS. The price point is part of the buyer definition, not a separate input. Re-segment the buyer population by price band before counting.
Treating the market-size number as a fact
The number is a hypothesis. It is refined by the first cohort of customers, and again by the second. The exercise that produces the number is more useful than the number itself — because the exercise forces the founder to write down which buyer, which geography, and which price, and that sentence is what the next experiment tests.
Revisiting
When to redo the market-size exercise
The number is a hypothesis. These are the moments when the hypothesis should be re-tested.
- After the first ten customers have actually bought — the cohort's actual price, retention, and segment changes SOM in a way no model predicted.
- After a major change in the buyer — a new regulation, a new buyer persona, a new geography opens or closes.
- After a major change in the channel — a new distribution partner, a new SEO position, or a new outbound motion changes the realistic reach rate.
- Before raising — the deck's market-size number should be the most-recent defensible number, not the number from a year-old model.
Connecting to Yibud
How the Yibud tools help
Yibud does not produce a market-size number — the bottom-up method above does. What the tools do is connect that number to the same idea evaluation so the size and the score move together.
Run the idea evaluator
The Startup MRI evaluator scores the idea across eight dimensions — market, competition, distribution, monetization, build, founder fit, opportunity, and validation. The market dimension is the one the bottom-up sizing exercise most directly informs; the other seven are independent of the market-size number.
Open the evaluator →Use the score calculator
The score calculator is a transparent version of the same eight-dimension model. The market dimension is the input the bottom-up sizing exercise most directly produces; the other seven are independent and use the same rules the evaluator uses.
Open the score calculator →Carry the number into the validation checklist
The validation checklist treats the market-size number as one of seven inputs that go into the continue-or-pivot decision. The number is not the decision; it is one piece of evidence, and the checklist is the place where it sits alongside the other six.
Open the validation checklist →FAQ
Frequently asked questions
What is TAM in simple terms?
TAM stands for Total Addressable Market. It is the theoretical annual revenue ceiling if every potential buyer of every possible alternative bought from a single vendor at the prevailing price. TAM is the largest of the three numbers and the least useful for an early-stage decision. McKinsey's market-sizing framework uses the same TAM / SAM / SOM terminology.
What is SAM vs SOM?
SAM (Serviceable Addressable Market) is the slice of TAM you can address with your current product, business model, and go-to-market. SOM (Serviceable Obtainable Market) is the slice of SAM you can realistically reach in the next twelve months. For a startup, SOM is the number that constrains the early-stage decision; SAM is the ceiling for three-to-seven-year growth; TAM is a theoretical ceiling for the whole category.
How do you estimate market size without making up numbers?
Use a primary source for the buyer count. The US Census Bureau publishes establishment counts by NAICS code; the Bureau of Labor Statistics publishes industry employment counts; Eurostat and the OECD publish equivalent datasets for other regions. The number must come from a source that shows its methodology, not from a market-research aggregator.
Is top-down or bottom-up better for a startup?
Bottom-up is more defensible for an early-stage startup, because every input is traceable to a primary source. Top-down is faster and produces a larger TAM, but the cuts are guesses and the chain is fragile. The two methods produce different numbers; the bottom-up number is the one you can defend in a customer interview.
How often should I redo the market-size exercise?
After the first ten customers have actually bought (the cohort's actual price and segment change SOM in a way no model predicted), after any major change in the buyer or the channel, and before raising. The exercise is a hypothesis, and the hypothesis should be re-tested when the inputs change.
What if the SOM is too small?
A small SOM is honest information. If the realistic year-one revenue ceiling is below the threshold the founder needs (a salary, a runway extension, a venture-scale outcome), the right response is to revisit the buyer definition — usually by widening the persona, broadening the geography, or adjusting the price point. The right response is rarely to inflate the number.
Where does market sizing fit in the validation sequence?
Market sizing sits between the idea evaluation and the customer-discovery phase. The evaluation surfaces the per-dimension evidence to collect; market sizing produces the year-one revenue ceiling; customer discovery tests whether the buyer is real. None of the three is a substitute for the other two.
Does Yibud produce a market-size number?
No — the bottom-up method on this page is the one that produces the number. Yibud's evaluator and score calculator use a market dimension as one of eight inputs, and that dimension is informed by the bottom-up exercise but is not the exercise itself.
Run the idea evaluator
Generate a Startup MRI report. The market dimension is informed by the bottom-up sizing exercise; the other seven dimensions — competition, distribution, monetization, build, founder fit, opportunity, and validation — are independent of the market-size number and feed into the same report.
Analyze my idea →