TAM/SAM/SOM

Back when market size was quoted as a single number, the number produced no decisions. "The collaborative document market here is worth $2.3 billion" sounds reassuring, but it leaves out what anyone should do next quarter. That $2.3 billion includes systems deployed to thousands of seats at large corporations, and it includes customers abroad that the product as built cannot reach at all.

In investor meetings the same number caused trouble from the other direction. If a deck said $2.3 billion, the question that came back was always the same one: how much of that can you sell today, and how much will you actually take in three years. Without an answer in the room, the big number read less as evidence and more as proof of inflation.

So the practice settled on describing one market in three layers. The outermost is everyone who needs the product. The middle is the part the current product and current sales channels can reach. The innermost is the share you can realistically take within a few years. The numbers shrink as you move inward and become useful as they shrink. The outer number shows whether there is room for a company this size to grow; only the inner number turns into a target for the quarter.

You peel inward, one layer at a time. Run it the other way around and you end up setting the inner number first, then decorating the outer layers to match it.

1. Decide who is being counted. The unit is usually a group of companies rather than individuals, and the conditions already written in the ideal customer profile (ICP) become the counting rule. Without those conditions, the edge of the count moves every time someone redoes it.
2. Size the outermost layer. This is the figure with no constraints applied, the total addressable market (TAM). There are two ways to get it: take a total from a market report, or multiply the number of target companies by what one of them pays in a year. The second is better. Written as 3,400 companies times $3,800 a year, a multiplication lets you point at which assumption was wrong later.
3. Cut one layer using the product and the channels. What remains is the serviceable addressable market (SAM). This step removes what the product can't do and where you can't sell. For a browser-only product that takes domestic payment methods, customers who require an on-premise install and customers abroad drop out here. That's how $2.3 billion becomes $300 million.
4. Cut one more layer using competition and your own capacity. What's left is the serviceable obtainable market (SOM). Established competitors, the size of the sales team, and how fast you can build all bite here. You write it as taking $1.5 million of that $300 million within three years.
5. Attach the basis to each number. One line per figure for the source and the assumption: which statistic the company count came from, and whether the per-company amount is the average of current customers or a hope.
6. Recalculate every six to twelve months. As the product grows the middle layer widens, and as the competitive picture shifts the inner layer moves.

Three situations call for these numbers: a fundraising deck, a decision about whether to start a new line of business, and a choice about which market to attack first. The last two use only the inner figure. The first asks for all three.

The largest limitation is that all three numbers rest on assumptions. Change the assumption of $3,800 per company per year to $2,300 and the result drops by about 40 percent. Which means the real value of the method isn't producing a number; it's putting the assumptions where people can see them. With them written down, someone can point at the specific one they disagree with. Without them, the conversation is only about whether to believe a figure.

The outer number is easy to inflate. Take the whole collaboration tools market, or the whole enterprise software market, and the figure grows while most of what's inside it has no reason to buy this product.

Products creating a category that didn't exist have no report to draw from. Only the bottom-up multiplication is available, and counting the target companies becomes your own job. Accuracy suffers, but every assumption is exposed.

"I follow the $2.3 billion. How much of it can you sell with the product you have now?"

Asking for the middle layer, and declining to decide on the outer number alone.

"We put SOM at $1.5 million over three years. It comes from how many companies two salespeople can reach."

Presenting the innermost figure together with its basis.

"This whole calculation assumes $3,800 per company per year. If that's wrong, all of it is wrong."

Asking to review the assumption before the number.

Pagedeck wrote these three layers down for the first time while building a fundraising deck. Grace put $2.3 billion in the outer layer and, after removing what a browser-only product can't serve, put $300 million in the middle. The inner layer came from the beachhead market the team had picked: software companies under 50 people, 3,400 of them. A beachhead is deliberately one narrow square to take first, so the range is small, and taking 400 of those companies within three years comes to $1.5 million a year.

The deck didn't stay a fundraising document. Once the figure of 3,400 existed, the order of features Mia maintained changed. Anything companies under 50 people never asked for slid down, and the question of how many accounts a year it takes to reach 400 turned into a hiring plan.

The common belief is that sizing the three layers means you understand the market. A number came out, so something feels settled. But this method answers size and nothing else; it doesn't say why anyone inside that size pays. Size a market and start building on that alone and you end up in a large number where nobody feels the product is theirs. That's why deciding who comes before measuring how many.

The other frequent belief is that a bigger outer number is better. It's true that the market has to be large for a company to have room in it, but a larger figure is a figure with more to explain. A deck claiming $75 billion has to account for why this team takes one square of something that size. Claiming $2.3 billion and showing solid grounds for reaching $300 million of it usually lands better.