A market research report will cost up to 25,000 euros and is likely wrong in ways you cannot detect. The methodology is not made transparent enough to be challenged. For most strategy and diligence purposes, you can build a more defensible estimate yourself using the TAM/SAM/SOM framework, applied with discipline.

TAM is not the number that matters for a decision

TAM / SAM / SOM

  • TAM: total revenue pool at 100% market share across all buyers. The ceiling, not the target.
  • SAM: the portion reachable given geographic scope, channel, and product fit. This is the number that matters for underwriting revenue.
  • SOM: realistic near-term share of SAM given competition and sales capacity. This is what goes into the financial model.

Conflating TAM and SAM is the most common error in growth-stage analysis. A business selling mid-market ERP software in the Netherlands does not compete in the global ERP market. Its SAM is the Dutch companies of the right size, not locked in with a dominant incumbent, with a procurement cycle that matches the sales model. Every filter skipped inflates the number and the conviction attached to it.

Use bottom-up as the primary estimate, top-down as the check

Top-down: start from a large industry figure, apply filters. Each assumption compounds the previous one. Four filters at 10% error each produce a number off by 40% or more.

Bottom-up: start from the unit, the buyer, the transaction. How many buyers exist in the target segment? What do they spend annually on the relevant category? Multiply through. Every assumption is explicit and individually challengeable.

Build the bottom-up estimate first. Use top-down as a sanity check. If they diverge by more than 30%, find the assumption doing the most damage before presenting either number.

Use public data: it is better than you think

Company filings in adjacent sectors disclose revenue by segment, customer counts, and average contract values. These are audited numbers, not analyst projections. Trade body publications often contain unit data on licensed operators or transaction volumes. LinkedIn headcount by job title is a reasonable proxy for the number of organisations that have invested in a function. Job posting volume across a competitive set reveals where capital is being deployed before revenue follows.

A range is more useful than a point estimate

A board decision does not require a number accurate to two decimal places. "The serviceable market is between 80 and 120 million euros annually" is a better input than "97.3 million euros" when the latter carries false precision. State the assumptions explicitly and rank them by sensitivity. The weakest assumption is the one that needs to be challenged.