It is Sunday night. The pitch is Tuesday. The founder, building software for independent crane rental operators in the Gulf states, opens a fresh tab and types the category into Statista. Two unrelated reports surface, both about construction equipment broadly, both behind a paywall, neither useful. She tries IBISWorld, which has a report on “Crane, Hoist, and Monorail Manufacturing” but nothing on the rental side. Crunchbase shows three companies tagged “construction software,” none of them competitors. She closes the laptop. The deck still has a slide titled “Market Size” with three empty boxes.
This is the wall. It is the most predictable failure mode in a Series Seed deck, and it has a method behind it that works whether the category is crane rental software, mobile dental clinics, freeze-dried pet food, or industrial sensors for grain elevators. The method does not require a paid subscription. It requires reading public records carefully and triangulating.
Section 1: Why the data is missing
The research industry sells reports on categories that already have buyers for those reports. A category becomes worth covering when a critical mass of investment bankers, corporate strategists, and consultants will pay $3,000 to $5,000 a copy to read it. That economic gate is the reason Statista, IBISWorld, Grand View Research, and Gartner cluster their coverage around mature, broad, consumer-visible categories.
Four conditions cause a market to fall through the cracks. The first is novelty. If the category did not exist in the previous NAICS revision cycle, no government statistician has assigned it a code, and no analyst house has assigned it a tracker. The second is fragmentation. A market made up of 4,000 small private operators with no dominant player generates no analyst interest because there is no acquirer paying for the report. The third is regional concentration. A service that exists in twelve metros but not nationally rarely justifies a national report. The fourth is the B2B-only filter. Consumer markets get covered because consumer brands buy the reports. A market made of dentists buying scheduling software, or warehouses buying lift truck telematics, gets covered when one of the buyers gets large enough to acquire a competitor and trigger a banker pitch.
None of these conditions mean the market is small. They mean the market is invisible to the tools that report on markets. The data exists. It is distributed across federal statistical agencies, trade associations, regulatory filings, and the LinkedIn pages of every operator. The job is to assemble it.
Section 2: The top-down move
Top-down sizing through adjacent-market triangulation works as follows. Identify a broader category that contains your target market as a subset. Find the public data series that measures the broader category. Then estimate, with a defensible ratio, what fraction of the broader category your target represents.
Take the crane rental software example. Statista does not cover the buyer. The Census Bureau does. The U.S. Census Bureau publishes County Business Patterns annually, with establishment counts, employment, and payroll broken down by 6-digit NAICS code, by state and by county. NAICS 532412 covers “Construction, Mining, and Forestry Machinery and Equipment Rental and Leasing.” That is the universe of buyers, plus some adjacent rental categories that include crane rental. The CBP tables let the founder pull the count of establishments in that NAICS code in Texas, Louisiana, Mississippi, Alabama, and Florida.
That gives a population. To convert population into market size, layer in an estimate of how many of those establishments are crane-focused (industry trade associations like the Specialized Carriers and Rigging Association publish member counts that establish a ratio), and apply a software spend per establishment derived from comparable verticals. Equipment rental ERP and dispatch software typically runs $200 to $800 per user per month, depending on tier. The founder needs a defensible mid-point and a citation. The result is a top-down number anchored to a real government data series, not a guess.
The same pattern works for almost any niche B2B category. Find the broader NAICS code. Pull the establishment count from County Business Patterns or the BLS Quarterly Census of Employment and Wages. Estimate the share of that broader category your target represents. Apply a per-establishment spend. Cite each step.
Section 3: The bottom-up move
Top-down is half the answer. A founder who only presents a top-down number is presenting a slide that an investor will discount on instinct. The bottom-up move counts the buyers directly and multiplies by the price.
The mechanics are simple. Start with a verifiable list of every plausible buyer. For consumer-facing categories, this might be the count of licensed practitioners in a state, available through state professional licensing boards. For B2B categories, it is the count of firms in the relevant NAICS code from County Business Patterns, cross-checked against trade association membership rosters. For categories where the buyer is an enterprise, LinkedIn’s company search produces a count of employers above a given employee threshold in a given industry. The SBA’s table of small business size standards provides the revenue and employee thresholds that the federal government uses to define “small” within each NAICS code, which gives a defensible bucket for segmentation.
Once the buyer count is established, the price per buyer comes from one of three places. Comparable public company filings disclose ARPU in 10-K and S-1 filings, where it is often broken out by segment. Comparable private company pricing pages disclose list prices. Founder interviews, ten of them, asked directly what the buyer pays for the closest substitute today, produce a defensible ARPU range.
Take an illustrative example. A founder is building scheduling software for mobile dog grooming businesses in the U.S. The Census Bureau records roughly 30,000 establishments under NAICS 812910 (“Pet Care, Except Veterinary, Services”) in recent County Business Patterns releases. Of those, the founder estimates from trade association data that 15% are mobile, not storefront, giving 4,500 buyers. The closest pricing comparable is a horizontal scheduling tool at $60 per month per location. Multiplied out, that is roughly $3.2 million in fully penetrated annual revenue. This is the illustrative bottom-up TAM. It is small, which is itself the answer. The founder either expands the buyer definition, widens geography, or reconsiders the category. The number did its job.
Section 4: The cross-check
The cross-check is the move most founders skip and most experienced investors silently grade on. The top-down and bottom-up numbers must reconcile within an order of magnitude. If the top-down says $400 million and the bottom-up says $4 million, one of them is wrong, and the founder needs to know which one before the pitch, not during the Q&A.
When they diverge, the failure mode is almost always one of three. First, the top-down adjacent market is too broad. If the founder used “all construction software” as the parent category and the actual target is “dispatch software for crane rental operators in the Gulf states,” the top-down will inflate by 50x or more. Tighten the adjacent category until the ratio is defensible. Second, the bottom-up buyer count is too narrow. If the founder counted only crane rental firms with their own websites listed in Google, the count missed the long tail of firms that exist in NAICS data but have no public web presence. Pull the count from the Bureau of Economic Analysis industry data and from Census, not from Google. Third, the ARPU is borrowed from the wrong comparable. A horizontal scheduling tool priced for groomers is not the same as a vertical dispatch system priced for crane operators, where downtime costs $5,000 a day and willingness to pay is correspondingly higher.
When the two numbers reconcile, the founder has a defensible TAM. When they diverge, the divergence itself becomes the next research task. Either outcome is useful. The mistake is presenting one number without having computed the other.
Section 5: How to present this to an investor
The slide does not say “the data was not available.” The slide presents a number, the method, and the sources, in that order. Investors discount slides that sound apologetic. They engage with slides that show work.
A useful format is three lines on the slide and a paragraph in the appendix. Line one is the TAM, with the method abbreviated (“$84M annual U.S. spend, bottom-up, 4,500 establishments x $1,560 annual ARPU”). Line two is the SAM, with the geographic or segment filter (“$22M, Gulf states only, large operators, year one focus”). Line three is the SOM, the realistic capture in three to five years given the go-to-market plan. The appendix paragraph names the data sources by URL and acknowledges the limits of the estimate. Investors who care about rigor read the appendix. Investors who do not care about rigor are not the ones to optimize for.
The defensive register is the trap. A founder who says “Statista does not cover this category” is conceding that Statista is the right source and the absence is a problem. A founder who says “we sized this from Census NAICS 812910 establishment counts cross-checked against trade association data and primary interviews with 12 operators” has reframed the absence of Statista coverage as a research finding, not a research failure.
Section 6: When to commission a market research report instead
The method above takes a sharp founder roughly 20 to 40 hours to execute well. It involves downloading CSV files from Census and BLS, filtering NAICS codes, reconciling state-level and national-level data, building a working spreadsheet, and writing methodology notes that survive an investor’s scrutiny. For founders who have the time and the appetite for the work, doing it themselves is the right call. The exercise produces a deeper understanding of the buyer that pays off in every subsequent pitch.
For founders who do not have 40 hours in the window before the raise, or who have done a draft and want it stress-tested before the pitch, an industry-specific market research report is the faster path. Pondera’s Market Research Industry Report is built for this case. It executes the same top-down, bottom-up, and cross-check methodology, with the data pulled from the same public sources, written up in a format that drops into a Series Seed deck and survives diligence. The deliverable is a 25 to 35 page report with an executive summary, methodology section, sourced TAM/SAM/SOM, competitive landscape, and an investor-facing appendix. Turnaround is four business days.
The decision is mostly about time and confidence. If the founder is the right person to build the model and has the time, building it is better. If the founder is past the napkin stage, has a pitch in 10 days, and needs the model to survive a partner-level review, commissioning it is the better trade.