Market sizing is the arithmetic that turns a defined market into a number, and it settles the question market definition deliberately leaves open.
A usable definition names who the buyers are and then asks for a count of them, with a stated method behind it. Market sizing is that method. The vocabulary of total, serviceable and obtainable markets came out of venture capital and became standard in investor decks through the 2000s, and product teams inherited the habit of quoting the largest of the three figures along with it.
A business case fails at the arithmetic far more often than it fails at the idea. The idea survives the first review, and then a finance director asks where the eleven thousand sites came from, and behind the number there is only a report somebody skimmed. So the arithmetic is the product manager's problem long before the pitch.
The sections below define the three sizes and give the two directions a calculation can run. One calculation is then worked through with real figures. Finally they set out the four ways a market size stops being believable.
The three sizes and what each one assumes
Each step inwards removes buyers for a reason somebody can state out loud, which is what separates a market size from a guess. A business case that shows only the outer rectangle has skipped both arguments.
Total addressable market is the annual revenue available if every buyer who has the problem bought a solution and this product supplied all of them. It is a ceiling and a fiction. It is useful only as the starting point for the two cuts that follow.
Serviceable addressable market removes the buyers this product could not serve even with unlimited demand. Those are buyers in countries it does not sell into, buyers whose regulator it has no approval from, buyers on a platform it does not support and buyers in segments it was never built for. Each exclusion is an argument, and writing the arguments down is most of the value of the exercise.
Serviceable obtainable market removes the buyers a competitor will keep. It is a share of the serviceable market, won by a stated date, given the competitors already in place and the organisation's own route to the customer.
Top down and bottom up
A top down calculation starts from a published figure for an industry and cuts it down with a series of percentages. It is quick, it is easy to check against other people's numbers and it inherits whatever definition the original analyst was working to.
A bottom up calculation starts from a single unit and multiplies upwards. The method is to count the buyers, multiply by how much each one buys and multiply again by the price. It is slower, it forces every assumption into the open and it is the only one of the two that a team can defend line by line.
The rule is to run both and reconcile them. If the two land within a factor of two of each other, the shape of the market is probably right. If they differ by a factor of ten, one of them rests on an assumption nobody has examined. Finding out which one is the most valuable hour in the whole exercise.
One bottom up calculation, worked through
The worked example is a product that schedules refrigeration maintenance for grocery chains in the United Kingdom. Every figure below is an assumption, and they are set out in a column so that each one can be argued with on its own.
| Step | Assumption | Running figure |
|---|---|---|
| Sites in scope | Grocery stores above 280 square metres | 11,000 |
| Segment | Share belonging to chains of ten stores or more | 6,820 |
| Qualified | Share with refrigeration under a maintenance contract | 5,456 |
| Price | Annual licence per site | £1,400 |
| Serviceable addressable market | 5,456 sites at £1,400 | £7.6 million |
| Obtainable in three years | 12 per cent of qualified sites | £917,000 |
A serviceable market of £7.6 million is a solid business for a team of eight and no business at all for a company carrying a hundred million pound plan. The arithmetic does not say which situation the organisation is in. It says where the ceiling sits, and the plan has to fit underneath it.
The four ways a market size stops being believable
Quoting the industry figure as the total addressable market. An analyst draws a boundary around a category of supplier. A market definition draws one around a problem and a buyer. Treating the supplier boundary as though it were the buyer boundary hands the most load bearing number in the business case to somebody who was answering a different question.
Running the whole thing as desk research. Published figures give the count of buyers. They cannot give willingness to pay. That number is the multiplier the answer is most sensitive to, and only conversations with buyers produce it. Steve Blank built customer development around this point in The Four Steps to the Epiphany in 2005. His instruction to get out of the building and test the assumptions against real buyers governs the market size as much as it governs the product.
Never reconciling the two directions. A team that runs one calculation has a number. A team that runs both and explains the gap has a model, and the model survives contact with a finance director.
Sizing on seats in a market that has stopped buying seats. If pricing is moving from a charge per user towards a charge per outcome or per unit of consumption, a seat count sizes a market that is being taken apart. The pricing metric has to be settled before the multiplication means anything.
A market large enough to be worth entering is almost always too large to win whole. The next decision cuts it into groups that behave differently and chooses one of them.