TAM, SAM and SOM: A Bottom-Up Market-Sizing Guide
TAM, SAM and SOM describe different boundaries around a market opportunity. Total addressable market is the broad demand for a defined offer. Serviceable available market is the portion your offer and operating scope can serve. Serviceable obtainable market is a plausible near-term share, constrained by acquisition and delivery capacity.
Definitions vary between presentations, so state your own scope and units explicitly. A transparent estimate with uncertain assumptions is more useful than a large number that cannot be reconstructed.
Define what is being counted
Specify the product category, customer type, geography, reference period and unit. Are you counting organizations, locations, individual users, transactions or annual revenue? Avoid moving between those units without explaining the conversion.
For a subscription sold to organizations, a starting formula is eligible buying organizations multiplied by annual revenue per organization. For a transaction service, estimate relevant transactions and the revenue earned per transaction. Gross transaction value is not the same as the startup’s revenue.
Decide whether the estimate concerns the current offer or an imagined future category. Keep speculative expansion separate so it does not inflate the opportunity available to today’s product.
Build the estimate from the bottom up
Find a documented starting population. For United States employer businesses, Census County Business Patterns can provide counts by geography and industry. Its unit is an establishment, which may not correspond to a separate buyer. Nonemployer Statistics covers businesses without paid employees and should be considered separately when relevant.
For other countries, use the national statistical office or sector authority. Record the dataset, reference year, coverage, exclusions and any transformation you make. A directory can help construct a prospect list, but check duplicates and whether listed organizations still operate.
Filter the population using criteria that affect your ability to sell and deliver: workflow, required technology, language, location and customer type. Distinguish documented exclusions from assumptions based on limited research.
Distinguish market size from a sales forecast
A market estimate describes an opportunity under stated boundaries. A sales forecast describes what your business expects to acquire and earn over time. A large TAM does not make the forecast more credible, and a modest near-term forecast does not necessarily mean the broader market is small.
Write the intended use of each number. TAM can explain the category’s scale. SAM can explain the fit between the current offer and a subset of customers. SOM should connect ambition to access, buying cycles and delivery capacity within a stated horizon. Keep a speculative future product outside the current SAM unless its requirements and evidence are clear.
For a multi-sided business, define whose spending is counted. A marketplace’s gross transaction value, seller revenue and platform revenue are different quantities. If you use a transaction fee, show the fee assumption and avoid adding several measures of the same transaction together.
Work through a labelled example
Illustrative example, using invented planning assumptions: a startup sells a reporting service at an assumed $1,200 per organization annually. It estimates 20,000 organizations in a broad category, of which 4,000 fit its initial language, workflow and support scope.
| Layer | Calculation | Illustrative annual revenue opportunity |
|---|---|---|
| TAM | 20,000 organizations × $1,200 | $24 million |
| SAM | 4,000 suitable organizations × $1,200 | $4.8 million |
| Capacity-constrained SOM | 60 supportable customers × $1,200 | $72,000 |
The SOM assumes the team can acquire and serve 60 paying customers in the stated annual period. It is not justified simply by taking an arbitrary percentage of SAM. It needs a route to those customers and a delivery plan. It also assumes a full year of revenue from each customer; staggered acquisition would produce less revenue within the first calendar year.
Constrain SOM using operations
Estimate how many qualified prospects the team can reach, what evidence supports conversion assumptions, how long buying takes and how many customers can be onboarded. Use ranges when the channel has not been tested.
Check support and retention as well as acquisition. If the team can sign customers faster than it can deliver, the sales forecast is not an obtainable market estimate. If the need is seasonal, timing may dominate the annual opportunity.
Test sensitivity. In the fictional example, 30 customers at $900 annually produce $27,000 of full-year revenue, while 80 at $1,500 produce $120,000. Neither scenario is a prediction. The range reveals which assumptions need evidence before a major commitment.
Build SOM from a reachable funnel
Start with the number of suitable prospects a specific channel can reach in the period. Then model the stages required for purchase and implementation. Use actual evidence where available and clearly labelled ranges where it is not. A funnel is a scenario, not a guarantee that prospects progress independently or at a stable rate.
Additional illustrative assumptions: the reporting startup can contact 600 suitable organizations over a year. It assumes 120 enter a serious discussion, 60 accept an offer and 40 can be onboarded with current capacity. Its operationally constrained acquisition case is therefore 40 customers, not 60, despite the higher sales assumption.
At $1,200 of full-year revenue per customer, 40 customers represent a $48,000 annualized revenue run rate once all are active. If customers join evenly through the year, the first year’s recognized or collected revenue will depend on the contract, billing and accounting treatment; it should not be reported automatically as $48,000 of first-year cash. Use a dated forecast for that question.
The earlier 60-customer SOM scenario might remain a longer-term or higher-capacity case, but the team should label the difference. Showing both without an explanation would mix assumptions and make the estimate difficult to assess.
Check filters and double-counting
Do not apply multiple filters as though they were independent unless the evidence supports it. Company size, technology use and workflow can be related. Multiplying separate guessed percentages may create apparent precision while compounding uncertainty.
Where feasible, inspect a sample of actual organizations against the complete eligibility criteria. Record uncertain classifications and the source date. A directory can contain dormant organizations, duplicate branches or companies whose services have changed. A population dataset can be valid while its classification is too broad for your offer.
Use the ideal customer profile to make the inclusion rule explicit. If one organization has several relevant teams, decide whether it can purchase one subscription or several. Apply that commercial rule consistently rather than switching units to increase the total.
Test sensitivity one assumption at a time
Identify which variables drive the result: eligible population, annual spend, purchase frequency, channel access or delivery capacity. Change them individually before combining an upside or downside case. This reveals what evidence would most improve the estimate.
For example, doubling SAM has little effect on the next year’s obtainable customers if onboarding remains limited to 40. Improving onboarding may be more consequential than finding a larger industry total. If pricing is uncertain, a specific paid offer can be more informative than another market report.
Keep price and scope linked. A higher assumed price may require more support or a different buyer, so it cannot always be changed in isolation. Explain the operating implications of each scenario and revisit the model when an experiment changes the underlying offer.
Preserve a reproducible worksheet
Store the source population, filters, units, reference year, currency, revenue convention and formulas together. Mark invented planning inputs visibly, including in exported versions. The market-sizing calculator helps with arithmetic, but its output is only as defensible as the assumptions entered.
The result should let another person reconstruct the estimate and identify its weakest input. That is more useful than a polished chart with a large unsupported total.
Present the limits honestly
Show the calculation, source dates, uncertain filters and pricing assumption. Explain why an organization qualifies and how double-counting was avoided. Market size does not demonstrate customer urgency, viable margins or access to buyers.
Use market research to investigate the population and idea validation to test behavior. The market-sizing calculator can organize arithmetic after the assumptions are explicit. Your next step is to replace the most consequential invented assumption with a documented source or a bounded customer experiment.