Market size startup: the framework, the common mistakes, and the evidence that separates a defensible answer from a confident one.
The TAM slide is the most abused slide in any deck. The pattern is always the same: a Gartner or IDC headline, a number with a B after it, and an implicit argument that capturing a small percentage of that number produces a large business. The investor has seen this slide five hundred times this year. It impresses no one. The version that survives diligence is bottoms-up, not top-down.
The problem is not the number. The problem is that a top-down market size number is not an argument. It is a citation. It tells you what a research firm estimated the total spend in a category to be. It does not tell you whether your business can reach the buyer, at what price, with what conversion rate, and at what cost. Those are the questions that determine whether the business is real.
The argument that actually matters is a bottom-up reconstruction. Start with the specific buyer. How many of them exist? In which geographies can you reach them in year one? What is the maximum price they will pay, not the price you want to charge? What fraction of the addressable segment will realistically convert in 24 months? Multiply those together and you have a number that is defensible because you built it from assumptions you can test.
A $4 billion TAM that produces a $40 million bottoms-up addressable segment is not a problem. It is accurate information. The question is whether a $40 million segment, at the margin structure your model requires, produces a fundable business. For many products, it does. For some it does not. The answer to that question is more useful than a defensible headline number.
The other issue with top-down market sizing is that it anchors the pitch to a category rather than to a problem. When a founder says "the HR software market is $35 billion," they are describing a taxonomy. When they say "there are 280,000 companies in the US with between 50 and 200 employees, all of which use at least one scheduling tool, and none of the current tools solve the compliance reporting problem that costs these companies an average of $4,200 per year in manual labor," they are describing a specific, priced gap in a segment they can reach.
One of these is a slide. The other is a business.
At Verdikt, market sizing is one of the first places we look for kill criteria. Not because a small market automatically kills an idea, but because the reasoning founders use to size the market tells you a lot about whether they have actually talked to buyers. A founder who can tell you the number of potential customers, the price point they tested against, and the two assumptions most likely to be wrong is a founder who has done the work. The TAM slide is often the first place that work is absent.
The memo worth writing does not lead with a market size. It leads with a buyer, a problem, and a price. The market size follows from those three things, not the other way around.
A worked bottom-up TAM, end to end
Take a vertical SaaS for outpatient mental health practices in the United States. A bottom-up TAM starts with three numbers and one assumption. Number one is the count of practices: per BLS occupational employment data, there are roughly 75,000 mental health practitioners across solo and small-group practices, of which an estimated 35,000 operate as billable practices large enough to need scheduling and billing software. Number two is what they pay today: most practices use a stack like SimplePractice or TherapyNotes at $50 to $90 per practitioner per month, suggesting a benchmark ACV band of $1,500 to $3,500 per practice annually. Number three is the assumption: you target the 35,000 billable practices at a $2,400 ACV midpoint. Bottom-up TAM lands at $84M. SAM, defined as practices in your two beachhead states where you have referral access, lands at $7M. SOM, defined as the practices you can close in year one with a founder-led motion of three demos a week, lands at $360K. Those three numbers tell the whole story. None of them require a Gartner report.
The argument is more useful than the size. If you can defend the 35,000 count from the BLS source, defend the $2,400 ACV midpoint from comparable subscription benchmarks, and defend the SOM math from your demo capacity, an investor knows you have done the work. If you can only show a $12B "behavioral health software" TAM from a press release, the investor knows you have not.
Where founders confuse market size with market access
A 100,000-buyer TAM is meaningless if 99,000 of those buyers are unreachable. Market access is the third leg of the argument and it is where most decks fall apart. Reachability has three dimensions: can you find the buyer (data exists), can you talk to them (a channel exists), and can you afford to acquire them (the CAC math works at your price point). First Round Review’s archive is full of GTM postmortems where the market was real and the access was not. The classic example is selling to small business owners where the buyer is reachable but the CAC to convert is two to three times the average lifetime value. Big TAM, no business.
Market access also shifts the math on SOM. A founder-led motion realistically books 8 to 15 first meetings per week per founder; a tightly targeted outbound machine with a sales rep adds another 25 to 40. That capacity ceiling is what produces a credible year-one revenue number, not the assumption of "1 percent market penetration." The 1 percent assumption is the single most common signal that the model was built bottom of the deck rather than top of the conviction.
What investors check, in order
The order matters. First, they look for evidence that you have talked to buyers in the TAM. The single most common partner question after a market-size slide is "How many of those have you actually spoken to?" Second, they check the ACV assumption against comparable wedges. If you are pricing 40 percent above the comparable wedge, you need pricing-research evidence; if you are pricing 40 percent below, you need a CAC argument that explains the discount. Third, they back-solve from SOM to first-year ARR. If your SOM does not produce a first-year ARR that justifies the round, the round is wrong, not the TAM.
This is the structure Verdikt’s methodology uses for every sizing memo. Bottom-up from primary sources, ACV cross-checked against three comparable wedges, SOM modeled across three GTM scenarios with named penetration ceilings, and growth rates triangulated from at least two independent sources rather than averaged. Top-down Gartner multipliers are explicitly excluded from the source library. The reason is not religion. It is that top-down numbers cannot be defended in a room where someone is trying to find the hole.