Market sizing is the process of estimating how many customers, purchases, or units exist in a clearly defined market and what that demand could be worth. A useful estimate states the customer, need, geography, offer, time period, price basis, evidence, and assumptions. It is a range for a decision, not a claim that the business will capture the whole market.
What do TAM, SAM, and SOM mean?
Total addressable market, or TAM, describes the demand if every customer meeting the broad definition bought the relevant offer. Serviceable available market, or SAM, narrows that total to customers the current product, geography, channel, and operating model can serve. Serviceable obtainable market, or SOM, is the portion the business might realistically win in a defined period given awareness, competition, capacity, and sales execution.
These labels are useful only when the boundaries are explicit. A global industry revenue figure is not the TAM for a local product simply because both share a category name. Define the problem and buying unit first. One company may be the customer, but several employees may be paid seats; one consumer may make several purchases per year. Keep customer count, purchase frequency, units, and revenue as separate lines in the model.
How do top-down and bottom-up estimates differ?
A top-down estimate starts with a broad published total and applies filters for geography, segment, or use case. It is quick and can provide a reasonableness check, but small percentage assumptions can create a false sense of precision. A bottom-up estimate starts with observable units: eligible customers, expected purchase frequency, plausible quantity, and net price. It usually makes operating constraints easier to see.
Use both when possible. If the answers differ sharply, examine definitions instead of averaging them. Build low, base, and high cases around the assumptions that matter most. For an early business, customer interviews, transaction records, public registries, channel data, and small demand tests can improve the inputs. Evidence that people experience a problem does not by itself prove they will pay the assumed price.
How should a market-size estimate be used?
Match the model to the decision. A founder choosing whether to test one city may need the number of reachable buyers and a twelve-month obtainable case, while a team comparing segments may need relative demand and economics. Connect the estimate to acquisition capacity, delivery capacity, contribution margin, and the time needed to win customers. A large TAM cannot compensate for an unreachable audience or unworkable unit economics.
Keep a calculation sheet with source dates, formulas, exclusions, and owners. Mark observations separately from assumptions. Review the model after customer research or real sales provide better evidence, and record why important boundaries changed. A useful market size becomes more accurate as uncertainty is reduced; it should not remain a polished number that nobody can reconstruct.
A fictional bottom-up estimate for a Bengaluru training offer
A training provider identifies 1,200 Bengaluru organisations that match its employee-size, function, and delivery criteria. That count is its current serviceable customer pool, not proof of demand. For a base obtainable case, it assumes 10% buy one cohort in the next year, with 15 paid seats at a net price of ₹4,000. The model produces 120 organisations, 1,800 seats, and ₹72,00,000 in annual revenue. A low case uses fewer buyers and seats; a high case is capped by instructor capacity. The team then tests the assumptions through outreach and paid pilots rather than presenting ₹72 lakh as guaranteed revenue.