Executive summary
Market size estimates drive investment, hiring and valuation decisions, yet many rely on a single method. Top-down sizing starts from a large, published total and narrows it; bottom-up sizing builds from customers, usage and price. Each has blind spots. Triangulating the two — and testing assumptions with experts and buyers — produces numbers leaders can defend.
Key findings
- 01Top-down estimates are fast but inherit the definitions and errors of their source data.
- 02Bottom-up estimates are more transparent but depend on accurate unit counts and adoption assumptions.
- 03A large gap between the two methods is a signal to revisit market definition, not to average the numbers.
- 04The most useful market model makes every assumption explicit and adjustable.
What is top-down market sizing?
Top-down sizing starts with a broad market total — often from industry associations, government statistics or published reports — and applies filters to reach the relevant segment: geography, customer type, product category and so on.
It is quick and useful for early screening. Its weakness is that the starting number may use a different market definition from yours, and each filter compounds uncertainty.
What is bottom-up market sizing?
Bottom-up sizing builds the market from its units: the number of potential customers, the share likely to buy, purchase frequency and price. In B2B markets this might be the number of plants using a process multiplied by equipment per plant and replacement cycle.
It is transparent and easier to link to a sales plan, but relies on inputs that often need primary research to validate.
Why triangulation matters
When top-down and bottom-up estimates broadly agree, confidence rises. When they diverge significantly, the gap is informative: it usually reveals a definitional mismatch, an unrealistic adoption assumption or a missing segment.
Expert and buyer interviews are the tie-breaker. They test the assumptions that move the number most — penetration, pricing, replacement cycles — and add the context no database provides.
- Define the market boundary before collecting data.
- Build both estimates independently.
- Identify the three to five assumptions with the greatest impact.
- Validate those assumptions in primary research.
- Present a base case with scenarios, not a single 'true' number.
Business implications
- Ask for both methods in any market sizing you commission or review.
- Treat a single-source market number as a starting point, not an answer.
- Keep the model live: update assumptions as your own sales data accumulates.