TAM, SAM and SOM: A Market Sizing Guide for Startups

Total addressable market, or TAM, represents the annual revenue opportunity if the defined eligible market were fully served. Serviceable available market, or SAM, narrows that scope to customers the offering can address given constraints such as geography, product capability and customer requirements. Serviceable obtainable market, or SOM, adds a realistic acquisition and delivery horizon. State your definitions because usage of the labels varies.

Choose a paying unit before counting: a household, company, location, user seat or transaction. Keep it consistent with the price. Multiplying company counts by a per-location price can undercount multi-site buyers; multiplying locations by a company-wide contract can overstate opportunity. The Census Bureau glossary distinguishes establishments from enterprises. An establishment is a location, while an enterprise can contain several establishments.

For each input, record the source, reference year, geography, classification, unit and known exclusions. Then mark what still requires direct customer research. A dataset can tell you how many establishments exist without establishing that each has your problem, can use your product or will pay the proposed price. A market report for broad industry spending may also include products your company will never sell.

The SBA’s market-research guidance combines existing data with direct research into demand, alternatives and pricing. Use those two forms of evidence together. Start with a defensible population, then investigate buying conditions through interviews and commercial tests. Keep an assumption register beside the spreadsheet so an attractive total does not obscure an untested price or compatibility requirement.

Consider an illustrative scheduling product priced per dental-office location. The Census Bureau’s 2023 County Business Patterns table reports 135,665 US employer establishments in Offices of dentists, NAICS 621210. It reports 48,076 establishments with five to nine employees. These are dated employer-establishment counts, not a count of dentists or a current list of buyers.

Now assume, purely for this example, a $1,200 annual price per location and a product initially designed for the five-to-nine-employee segment. The broad population multiplied by that hypothetical price produces a $162,798,000 annual opportunity proxy. The size-filtered population produces $57,691,200. Neither figure demonstrates actual demand. The smaller figure still needs filters for compatibility, buying authority and product requirements before being treated as a validated SAM.

For the same hypothetical product, suppose the team plans to reach 150 retained paying locations by the end of twenty-four months. Explain how that number could be achieved: qualified accounts reached, evaluations started, conversion timing, losses and onboarding capacity. It is a plan to test, not a market statistic. A percentage of the national market does not supply this operating explanation.

Check the capacity constraint separately. If each new location requires several days of specialist implementation, a small team may be unable to onboard every account implied by the sales forecast. Model the limiting step and consider what changes would increase capacity. Also distinguish the $180,000 annual run rate at the endpoint from cumulative revenue during the two-year period. Customers arrive at different times, so multiplying the endpoint run rate by two would misstate the revenue path.

Vary one material assumption at a time: price, eligible locations, time to conversion, customer losses or implementation capacity. Then test combinations that could occur together. A lower price might increase interest but also reduce the resources available for onboarding. Keep the relationships explicit rather than changing every input independently in the optimistic direction.

Use a top-down industry estimate as a reasonableness check, with matching geography and product scope. Investigate large discrepancies before averaging two incompatible numbers. For the dental example, verify whether central purchasing means several locations share one buying decision and whether the product’s pricing still applies per location. Update the dataset when a suitable newer release is available, while preserving the reference year in the model.

Present the result as a dated range with a short explanation of the key assumptions. Link the obtainable customer base to the go-to-market plan and use only the necessary summary in the pitch deck. The spreadsheet should answer follow-up questions without relying on a large, unsupported headline.

Explore NextUnicorn’s published Rating framework