How to do it
Filter by what the product actually does
If your tool only works with one accounting system, SAM is the buyers on that system. Not the ones you could support later — the ones you could onboard this quarter. Future capability belongs in a future SAM.
Filter by where you can legally and practically sell
Regulation, language, payment rails, data residency, support hours. A product that cannot take payment in a country is not addressing that country, however many buyers it has.
Filter by who you can reach at all
This is the filter founders skip. A segment you have no route to — no list, no channel, no community, no search intent — is not serviceable no matter how well the product fits. Reachability is part of the market, not part of the marketing plan.
Name each filter and its source
Three named filters that each remove a countable group are defensible. "About 20% of TAM" is not a calculation; it is a wish with a decimal point.
Worked example
Filters that actually narrow a market, and what each one costs to remove
The last row is the one founders leave out, and it is usually the most expensive to remove. A segment you cannot find is not addressable, and no amount of product work changes that — it is a distribution problem wearing a product costume.
Reachability is part of the market, not part of the marketing plan
A founder can usually tell you why their product fits a segment and almost never how they would find the first hundred of them. Those are the same question. If there is no list, no community, no search intent and no partner who already has them, the segment is not serviceable this year regardless of fit.
The practical test: name the channel, then name how many of the right buyers it puts in front of you per month. If you cannot answer the second half, leave the segment out of SAM and put it in a note about where the market could extend to.
SAM is the number your roadmap argues about
TAM does not move when you build something. SOM moves when distribution works. SAM is the only one of the three that responds directly to product decisions, which makes it the right number to attach to a roadmap debate.
Put it to work: for each candidate integration or locale, write what it adds to SAM and what it costs to build. A quarter of engineering that adds 400 reachable buyers is a different decision from one that adds 4,000, and most teams argue about those two options without ever writing either number down.
What goes wrong
- Multiplying TAM by a round percentage instead of applying real filters
- Filtering only by geography and ignoring whether you can reach anyone
- Counting segments your product could serve after a rebuild
- Making SAM so close to TAM that the distinction stops meaning anything
Is your number plausible?
Each check catches a different way the arithmetic can be right and the answer still wrong.
- If SAM is more than about 70% of TAM, your filters are not doing any work — look again at reachability
- If SAM is under 5% of TAM, check whether you are filtering on things the product could support with a week of work rather than genuine limits
- Name the single filter that removes the most buyers. That is your biggest roadmap or distribution decision, and it should feel like one
Related answers
How is SAM different from a target segment?
A target segment is who you are going after first. SAM is everyone you could serve today if they all arrived at once. The segment sits inside SAM, and confusing the two makes SAM too small to be useful as a ceiling.
Should SAM include buyers a competitor already has?
Yes. SAM is what your product can serve, not what is unclaimed. Existing competitor customers are usually the easiest to sell to, because they have already proved they will pay for the category.
What if my filters remove almost everything?
Then the honest reading is that version one serves a narrow slice, which is fine — but say so, and size the slice properly rather than quoting TAM. A small SAM with a plan to widen it is a strategy; a large TAM with no serviceable middle is a hope.