Ask for a market sizing report on a niche B2B software category and you will usually get a number from a research firm that covers "vertical software" as one line item inside a much larger report. The number is not wrong exactly. It is just measuring something else.
For a business selling scheduling software to regional dental practices or compliance software to credit unions, the market that matters is not a slice of a billion-dollar category estimate. It is a specific, countable, finite population of businesses that could plausibly buy this specific product. That population can be counted directly, and counting it directly produces a far more useful number than any top-down report will.
Total addressable market is a headcount problem, not a research problem
Start with the actual count of the target buyer unit: how many independent dental practices, how many regional credit unions, how many crane operators exist in the relevant geography. Government business registries, trade association membership counts, and licensing databases usually get you within a reasonable range. Multiply that count by a realistic average contract value, not the top-tier customer's contract value, and the resulting number is both defensible and far more specific than anything a syndicated report will hand you.
Serviceable addressable market means applying real constraints
Not every business in the total count can actually buy the product. Some are too small to afford it, some are locked into a competitor with high switching costs, some operate in a geography or regulatory environment the product does not support. This is where most sizing exercises get lazy and apply a flat percentage haircut instead of naming the actual constraints. Naming them is what makes the number defensible under diligence pressure instead of collapsing the first time someone asks how you got there.
Serviceable obtainable market is where honesty actually pays off
This is the number that should scare people a little, because it is bounded by realistic penetration rates in a finite market, and finite markets have real ceilings. A company that already owns 35 percent of a addressable population of 1,200 businesses does not have unlimited runway left, and pretending otherwise in a model just moves the disappointment from the sizing slide to the year-three actuals. The honest version of this number is also the more useful one: it tells you what the growth story actually depends on, whether that is price increases, adjacent product expansion, or genuine geographic expansion into a new but comparably finite population.
What this means for underwriting
A bottom-up count takes longer to build than pulling a number from a report, and it is worth the extra hours every time in a niche category. The output is not just a bigger or smaller number. It is a number you can defend line by line in front of an investment committee, which is the entire point of doing the sizing work in the first place.