Skip to content

TAM / SAM / SOM

First PublishedLast UpdatedByAtif Alam

The decision this page enables: how big the opportunity actually is — credibly, defensibly, and at three useful levels of zoom.

Market sizing answers “how much revenue could this product generate, in theory?” The TAM / SAM / SOM framework breaks the answer into three nested estimates so you don’t confuse the size of a fantasy (the whole industry) with the size of what you can actually go after this year:

  • TAM (Total Addressable Market): total demand if every possible customer in the universe bought your category.
  • SAM (Serviceable Addressable Market): the slice of TAM you can realistically reach given your product, geography, distribution, and constraints.
  • SOM (Serviceable Obtainable Market): the share of SAM you can realistically win in a defined time horizon (usually 3-5 years).
flowchart TB
    TAM["TAM<br/>Total Addressable Market<br/>universe"] --> SAM["SAM<br/>Serviceable Addressable Market<br/>realistic reach"]
    SAM --> SOM["SOM<br/>Serviceable Obtainable Market<br/>3-5 year capture"]
  • It tells you whether the opportunity is big enough to be worth building — early-stage products under-size all the time, but they also occasionally over-size and chase a $50M opportunity dressed as a $5B one.
  • It’s a required artifact for investor and board conversations.
  • It anchors Workbook → Targets baseline — your year-1 revenue target sits inside SOM, not TAM.
  • It connects directly to Strategy: ICP — SAM is the ICP, sized in dollars.

The two methods (and why bottom-up wins at early stage)

Section titled “The two methods (and why bottom-up wins at early stage)”

There are two ways to estimate each layer. Investors and founders argue about which to use; the right answer is “use both, and trust bottom-up more.”

  • Start from an industry-analyst total (“project management software is a $X billion market”).
  • Apply percentages to narrow: ”% in your geography × % in your segment × % we can reach.”
  • Strength: fast; works when bottom-up data isn’t available.
  • Weakness: the analyst total is often wrong, the percentages are guesses, and the result feels precise without being grounded. Investors discount top-down heavily.

Bottom-up (build from your unit economics)

Section titled “Bottom-up (build from your unit economics)”
  • Count the actual addressable accounts/users.
  • Multiply by realistic ACV (annual contract value) at typical maturity.
  • Strength: grounded in your actual ICP definition and pricing; defensible line-by-line.
  • Weakness: requires you to have a credible ACV — which means you need real customer-discovery and a tested price, not a guess.

Rule of thumb: if your bottom-up SAM is more than 10% of your top-down SAM, one of your numbers is wrong. They should be in the same order of magnitude; if they diverge by more than 10×, you have a definition mismatch.

TAM = (total addressable accounts globally) × (average ACV at maturity)
or, top-down:
industry analyst's category total revenue
SAM = (addressable accounts in your reachable geo / segment) × (average ACV)
or, top-down:
TAM × (% in your reachable geo) × (% in your target segment)
SOM = (accounts you can realistically win in years 1-3) × (your ACV)
or, top-down:
SAM × (realistic market share in years 1-3, typically 1-5%)
  • SOM in years 1-3 is rarely more than 1% of SAM for a new entrant. >5% requires extraordinary distribution or category-creation momentum.
  • SAM is rarely more than 10% of TAM unless you’re explicitly geographically-narrow.
  • If TAM looks > $50B, you’ve probably defined the category too broadly. Investors will discount it on principle.
  • If SOM is < $50M over 3-5 years, the opportunity may be too small to attract growth capital — but may still be a great business if you’re bootstrapped.
PRODUCT: [one-line product description]
ICP: [from /library/strategy/icp/]
TIME HORIZON: [years for SOM — usually 3-5]
TAM (top-down):
Source: [industry analyst report / public market report]
Category total: $______
Year: [YYYY]
TAM (bottom-up):
Addressable accounts (global, all segments): N = ______
Average ACV at maturity: $______
TAM = N × ACV: $______
SAM:
Geo constraint: [e.g. US + Canada only at launch]
Segment constraint: [e.g. 5-50 person SaaS teams]
Addressable accounts in reach: N = ______
Average ACV at our pricing: $______
SAM = N × ACV: $______
SOM (years 1-3):
Realistic market share in years 1-3: ______%
or
Accounts winnable per quarter × quarters: ______
SOM: $______
Sanity checks:
Bottom-up TAM / Top-down TAM ratio: ______ (should be 0.1-1.0×)
SAM / TAM ratio: ______ (typically 1-10%)
SOM / SAM ratio: ______ (typically <5% for new entrants)
  • TAM growth rate — how fast is the category expanding? Stagnant categories make scaling harder, even at high market share.
  • SAM penetration — your current revenue ÷ SAM. Tracks “are we getting bigger in our actual reachable market?”
  • SOM accuracy — actual revenue at year N vs. your year-N SOM forecast. Calibration data for the next round of sizing.
  • ACV trend — average contract value over time. If ACV is climbing, your bottom-up SAM should be re-estimated upward (and your ICP may be shifting up-market).
  • Adjacent-segment readiness — accounts in adjacent SAM segments that have approached you organically. Signal of where to expand SAM next.

Top-down:

  • Project / workflow management software category (analyst report): ~$8B globally, growing ~12%/year.
  • Reachable geo (US + EU + Canada + AUS): ~60%.
  • Target segment (5-50 person SaaS teams): ~12% of category buyers.
  • Top-down SAM: $8B × 60% × 12% = ~$580M.

Bottom-up:

  • ~120,000 SaaS companies globally with 5-50 employees (estimated from Crunchbase + LinkedIn).
  • In reachable geos: ~75,000 accounts.
  • Average ACV at $29/seat × ~7 seats × 12 mo = ~$2,500.
  • Bottom-up SAM: 75,000 × $2,500 = ~$190M.

Bottom-up is ~30% of top-down — within the 10-100% sanity-check range. The bottom-up number is the one used internally; the top-down is the headline for investor narrative.

SOM (years 1-3):

  • Realistic year-3 share of bottom-up SAM: 2% (a stretch but plausible with PLG and an existing community presence).
  • SOM: $190M × 2% = ~$3.8M ARR by end of year 3.

This anchored the year-3 plan: ~1,500 paid accounts at average $2,500 ACV. The number was used to sanity-check the Workbook → Targets baseline — and revealed that the initial $1.2M year-1 target was about right (roughly 30% of SOM by year 1).

B2C sizing uses people-not-accounts at the bottom-up level. Same framework, different units:

  • TAM (global home-fitness mobile app users, all categories): ~250M users.
  • SAM (US + EU + UK, casual home exercisers 25-50): ~25M people.
  • ACV at $5.99/mo with 25% annual retention: ~$18/user/year blended.
  • Bottom-up SAM: 25M × $18 = ~$450M.
  • SOM at 0.5% share in 3 years: ~$2.25M ARR.

The lesson: SOM at year 3 is a very small fraction of SAM (0.5%). This is normal for a new B2C entrant in a crowded category. Investors who push for 5% SOM in year 3 are asking for fantasy numbers; founders who promise it are setting themselves up.

  • Confusing TAM with the opportunity. TAM is the upper bound of fantasy. SOM is the actual near-term opportunity. Investors care about both, but board operators should run on SOM.
  • Top-down without a bottom-up cross-check. Top-down numbers are easy to manufacture. A bottom-up estimate that lands within an order of magnitude is what makes the number defensible.
  • Defining the category too broadly. “We’re in the $X-trillion business software market” is a tell. Define narrowly enough that the SAM number is small enough to be embarrassing — that’s usually the right zoom.
  • Static sizing. Markets and ICPs change. Re-size annually; expect SAM to grow as you broaden ICP, and shrink as you tighten it.
  • Forgetting “do-nothing” reduces SAM. Not every account in your reachable segment will ever pay for anything in your category. A practical SAM excludes the perpetual do-nothing population — typically 30-50% of the gross count.
  • Mixing time horizons. TAM is timeless. SAM is “at maturity.” SOM is “in years 1-3.” Don’t compute them under different time assumptions.