What are TAM, SAM and SOM?
TAM is the total market for the problem, SAM is the portion your model can actually serve, and SOM is the share you could realistically win in a few years. The purpose is not the headline number — it is the reasoning that gets you from a large abstraction to a small defensible one, and a bottom-up calculation is worth more than any industry report.
The three figures
TAM — total addressable market. Everyone with the problem, if you could serve all of them with no constraints. Deliberately theoretical.
SAM — serviceable addressable market. The portion your product, geography, language, regulation and business model can actually reach today. TAM minus reality.
SOM — serviceable obtainable market. What you could plausibly capture in three to five years given competitors, your sales capacity and your distribution. This is the only figure with any operational meaning.
Bottom-up beats top-down every time
Top-down goes: the global market is 50 billion, we will take 1%, that is 500 million. The 1% is invented, and everyone reading it knows so. It is the single fastest way to lose credibility in a strategy document.
Bottom-up goes: there are roughly 180,000 companies in this segment in our launch markets, our realistic price is 400 per month, giving a SAM of about 860 million annually; our sales capacity supports reaching perhaps 3% within three years, so SOM is around 26 million.
Every number in the second version is challengeable, which is precisely why it persuades. It also produces something useful — if the arithmetic only works at a price nobody will pay, or at a customer count exceeding the entire industry, you have learned something about the business rather than about the spreadsheet.
What sizing is actually for
Three genuine uses:
Sanity-checking ambition against the model. If reaching your target requires 40% share of a market with three entrenched incumbents, the plan needs changing now.
Choosing between segments. Relative sizing across two candidate markets is far more reliable than absolute sizing of either, because the same systematic errors apply to both.
Finding the wedge. The valuable output is usually not the total but the discovery of a specific subsegment that is underserved, reachable and large enough to start in.
Zerodha did not enter a large broking market and take a slice of it. It priced for a segment the incumbents were structurally unable to serve profitably, and grew from there. The sizing that mattered was of the wedge, not of the industry.
Where the number misleads
Redefining the market to inflate it. Every failing product can reach a huge TAM by describing itself more broadly. Communication, productivity and wellness are not markets; they are categories of aspiration.
Assuming the market is available. A large market with entrenched network effects is not addressable just because it is large. Foursquare's unbundling into Swarm faced a market that existed but had already resolved around habits the split disrupted.
Confusing a market with a moment. MoviePass had a real and large market of people who liked cheap cinema tickets. Nothing about the size of that market made the unit economics survivable.
If your strategy depends on the TAM being large, it is not a strategy. If it depends on a specific reachable wedge inside it, it might be.
Seen in practice
Case studies where this shows up as a real decision, not a definition.
Related questions
What is the difference between top-down and bottom-up market sizing?
Top-down starts from a published industry figure and applies percentages to it, which is fast and almost impossible to defend because the percentages are invented. Bottom-up starts from the number of potential customers times a realistic price, which is slower and can be argued with — which is the point.
Does a huge TAM make a business more attractive?
Not on its own. A huge TAM with no wedge into it is worse than a modest market you can dominate, because the large number attracts competitors and the lack of a wedge means you meet all of them at once. Investors read the reasoning, not the total.
How do I size a market that does not exist yet?
Size the behaviour it would replace. Nintendo could not have sized non-gamers buying a console from console data, but the population of households not currently gaming was countable — and that framing was the strategy, not just the arithmetic.
More on strategy
Last reviewed 2026-09-08