TAM, SAM, and SOM answer three different questions: how big the market could be, how much of it your product can actually serve, and how much you can realistically win in the next few years. Most B2B decks get the first number impressively wrong and the third one embarrassingly vague. This guide gives the formulas, the bottom-up method that survives due diligence, a fully worked B2B example, and the data sources to build each number from.

The numbers downstream of market sizing live in our B2B SaaS pipeline benchmarks and CAC payback benchmarks.

TL;DR

Calculate all three bottom-up: TAM = total target accounts x achievable annual contract value, SAM = the subset your product, language, geography, and channel can actually serve x ACV, and SOM = the share your sales capacity can win in 2-3 years, built from rep math rather than a plucked percentage. Worked example below: 40,000 target agencies x $3,600 ACV gives a $144M TAM, ICP filters cut it to a $43M SAM, and a two-rep capacity model yields roughly $1.3M ARR in year three, about 3% of SAM. Top-down sizing from analyst reports is context, not evidence: investors and operators both trust the count-times-price build, and account counts are pullable in an afternoon from LinkedIn and B2B databases.

The three numbers, defined properly

TAM, total addressable market. Everyone who could conceivably buy the category, times what they would pay annually: total accounts x ACV. It sizes the prize if you had no constraints, and its job is to prove the ceiling is high enough to matter.

SAM, serviceable addressable market. The TAM after your real constraints: segment, geography, language, compliance, deployment model, channel reach. This is the market your current product can sell into, and it is the number your go-to-market plan should be built against.

SOM, serviceable obtainable market. What you can win in a defined window, usually 2-3 years. The credible version is built from sales capacity and win rates, not from "1% of the market": capacity-based SOM is a plan, percentage-based SOM is a wish.

Top-down vs bottom-up: use both, trust one

Top-down starts from an analyst market figure and slices downward ("the market is $12B, our segment is 8%..."). Fast, citable, and structurally flattering, because every slice is an assumption stacked on someone else's assumption. Use analyst data from Statista, Gartner, or vendor intelligence platforms like HG Insights as context and sanity check.

Bottom-up counts the actual buyers: how many companies fit the profile, times what each pays. Slower, defensible, and the version that survives investor diligence and board scrutiny, because every input can be audited. The account counts are an afternoon of work: LinkedIn Sales Navigator filters give company counts by industry, headcount, and geography, Apollo and Crunchbase give firmographic counts and funding filters, and your own closed-won data gives the honest ACV.

Value-theory sizing (what would the solved problem be worth?) is useful for pricing strategy and category creation, less so for sizing; our B2B SaaS pricing strategy guide covers that lens.

The worked example, end to end

Product: a deliverability monitoring platform sold to B2B agencies at $300/month, $3,600 ACV.

Step 1, TAM. Database and LinkedIn counts find roughly 40,000 marketing and lead-gen agencies in the target regions that run email as a service. TAM = 40,000 x $3,600 = $144M. Honest, and honestly the least useful number of the three.

Step 2, SAM. Apply the real filters: 10-200 employees (below that they will not pay, above that they build in-house), English-speaking markets, actively running cold outbound (observable via job posts and tech signals). The count drops to ~12,000 accounts. SAM = 12,000 x $3,600 = $43M.

Step 3, SOM from capacity, not vibes. Two AEs, each closing five deals a month at steady state, is ~120 new customers a year; with ramp and churn, call it 360 customers by end of year three. SOM = 360 x $3,600 = ~$1.3M ARR, roughly 3% of SAM. That number connects directly to pipeline math: at a 25% win rate it implies ~1,440 opportunities over three years, which your MQL-to-SQL conversion rates translate into a required lead volume. If the required lead volume is implausible, the SOM was too.

The sanity checks. ACV comes from closed-won data, not aspiration. The account count excludes companies that structurally cannot buy (wrong stack, wrong compliance regime). And the three numbers should be far apart: a SAM that is 90% of TAM means the filters were cosmetic, a SOM above ~10% of SAM in three years needs an extraordinary distribution story to defend.

Where each number actually gets used

TAM belongs in fundraising narratives and category bets: it answers "is the ceiling high enough". SAM drives go-to-market design: ICP definition, territory planning, channel budgets, and whether the pipeline math in your plan is physically possible. SOM sets the revenue plan, headcount model, and quota capacity, and connects sizing to strategy choices like ABM tiers (a small SAM of high-ACV accounts wants ABM; a wide SAM of small accounts wants volume motions per our GTM motion comparison). Re-run the build yearly: ACV moves, filters sharpen, and a sizing model that never changes is a sizing model nobody uses.

FAQ

What is the difference between TAM, SAM and SOM?

TAM is total demand for the category (all possible buyers x ACV), SAM is the portion your product and channels can actually serve after segment, geography, and capability filters, and SOM is the share you can realistically win in 2-3 years given sales capacity and win rates.

How do you calculate TAM bottom-up?

Count the companies that fit your buyer profile using database and LinkedIn filters, then multiply by an achievable annual contract value taken from real closed-won data. Bottom-up TAM = target accounts x ACV, with every input auditable.

What percentage of SAM should SOM be?

There is no magic percentage: build SOM from sales capacity (reps x deals per rep x timeframe, adjusted for ramp and churn) and then express it as a share of SAM. In practice credible three-year SOMs often land in the low single digits of SAM; above ~10% requires an unusual distribution advantage to defend.

What data sources work for B2B market sizing?

LinkedIn Sales Navigator for company counts by size, industry, and geography; Apollo or similar B2B databases for firmographic and technographic filters; Crunchbase for funding-stage cuts; Statista, Gartner, and market intelligence platforms for top-down context; and your own CRM for the only ACV that matters.

Is top-down or bottom-up market sizing better?

Bottom-up is the number people trust, because accounts-times-price can be audited input by input. Top-down is useful as a sanity check and for citing market momentum. Present bottom-up as the case and top-down as the context, never the reverse.

How often should you recalculate TAM SAM SOM?

Yearly as a rhythm, and immediately after anything that changes the filters: a new product line, a new geography, a pricing change, or a pivot in ICP. SAM and SOM move much faster than TAM, and they are the two numbers your plan actually runs on.

Bottom line

TAM SAM SOM done right is one multiplication done three times with progressively honest filters: all accounts x ACV, serviceable accounts x ACV, winnable accounts x ACV. Build it bottom-up, source the counts from databases you can re-run, derive SOM from rep capacity, and let the three numbers drive fundraising, go-to-market design, and the revenue plan respectively.

Want the sizing wired into an actual pipeline plan, from account counts to quota math? Book a call with GROU. We build go-to-market engines for B2B companies across verticals.

We are GROU, a B2B pipeline agency that runs lead generation, outbound, and LinkedIn content for clients across manufacturing, fintech, iGaming, software, and professional services. The sizing method reflects the market builds we run for client go-to-market plans between 2024 and 2026, anonymized to protect client confidentiality.

Some links in this article are affiliate. We may earn a small commission at no extra cost to you. We only recommend tools we've deployed for clients.