TAM, SAM and SOM: Market Sizing That Survives a Second Question

TAM, SAM and SOM are three nested estimates of how much revenue a market could provide. The total addressable market is the whole opportunity if every organisation with the underlying need bought something like your product. The serviceable addressable market is the part of that you could actually sell to given your product, price, geography and distribution. The serviceable obtainable market is what you could realistically win in a defined period given your capacity and your conversion rates. Each layer strips away an assumption, and the value of the exercise lies almost entirely in making those assumptions explicit rather than in the final number.

Top-down

Start from a published market figure and narrow it with percentages. It is fast and it produces a number that sounds authoritative, but every percentage in the chain is a judgement that cannot be checked, and the published starting figure was usually built for a different purpose with a different definition.

Bottom-up

Start from units. Count the organisations that match your criteria, multiply by what one of them would pay in a year, and repeat by segment. It takes longer, and every assumption is visible: the count can be challenged, the price can be checked against your own deals, and the arithmetic connects directly to how the business sells. Use top-down afterwards as a sanity check on the result, never as the source of it.

A worked bottom-up example (illustrative figures)

Suppose your criteria describe organisations in particular industries, of a particular size, in particular cities, and a structured count gives 40,000 of them. Your average annual contract value in that segment is ₹1,20,000. Total addressable revenue for the segment is 40,000 multiplied by ₹1,20,000, which is ₹480 crore. Now apply reality: you can currently support and sell to 15,000 of those, giving a serviceable figure of ₹180 crore. Finally apply capacity: your team can genuinely work 1,200 accounts in a year and converts 12% of the accounts it works, producing 144 new customers at ₹1,20,000, or roughly ₹1.7 crore of new annual contract value. That obtainable figure is the only one of the three you can check against last year, which is exactly why it is the one worth planning on.

The most useful decision it drives is comparative. Absolute market size rarely changes a plan, but the relative size of two segments does, particularly when they differ in deal value, win rate and cycle length as well as in count. A segment with half the accounts but twice the contract value and a much shorter cycle is often the better place to concentrate, and only a segment-level model will show that.

The second use is capacity planning. Knowing how many organisations match your profile tells you how many can be covered by a given number of salespeople, and therefore whether the constraint on growth is demand or headcount. Businesses regularly hire against an assumption of unlimited market and then discover their serviceable market is thinner than the plan required, which is a much more expensive lesson than an afternoon of arithmetic.

Definition creep

The commonest inflation is widening the definition until the market includes organisations with no realistic need or budget. Counting every registered business in a country, or every company in an industry regardless of size, produces a headline number that is technically arithmetic and practically meaningless. If the estimate does not state how many organisations were counted and on what criteria, it cannot be evaluated.

List price instead of realised price

Multiplying a large count by a published price rather than an actual average is the second inflation. Discounts, plan mix and the fact that smaller organisations buy smaller configurations all pull the realised figure down, often substantially. Using your own average contract value by segment is both more honest and easier to defend.

An obtainable figure pulled from the air

Stating the obtainable market as a modest-sounding share of the serviceable one is the most common shortcut of all, and it converts the one layer that could have been grounded in evidence into a guess. Deriving it from accounts worked, win rate and deal value takes ten minutes longer and produces a figure that can be compared against actual performance at the end of the period.

Ask three questions of any estimate. How many organisations are in it, and by what criteria were they counted? What price was applied, and is it a list price or a realised average? And what period does the obtainable figure cover? An estimate that cannot answer all three is a headline rather than a model.

Then look at composition rather than total. A market whose value is concentrated in a small number of large accounts demands a different go-to-market approach from one spread thinly across many small ones, even where the totals match. And treat the model as something to revise: as you win and lose deals, your real win rates and contract values replace the assumptions, and the estimate should get better every quarter rather than sitting unchanged in a deck.

Each layer answers a different question and carries a different level of reliability.

The fourth row is worth including deliberately. Pipeline coverage is not part of the traditional trio, and it is the number that tells you whether any of the other three is being acted on. A large obtainable market with no pipeline in it is a model, not a plan.

Sizing improves as soon as it is fed by real conversion data rather than assumptions. Start with what a CRM is, then look at lead management software, CRM for small business, AI CRM for scoring by fit, sales automation for covering more accounts per person, and sales use cases.