Each vertical below gets the same treatment: the full 100M+ domain universe, cut to the category, read company by company against your thesis. Ten already carry a published specimen report you can inspect.
Each page explains how full-web screening behaves in that specific market: which website signals carry weight, which keyword traps exist, and how we separate independent operators from group-owned branches. Where a specimen run exists, the page quotes its anonymized evidence.
Ten subverticals from one 367,478-domain category — every one carries a published specimen.
A specimen is not a teaser list. It shows our confident fits, the fits keyword tools miss, and the lookalikes we refused — with the exact site quote behind each call.
Roughly one in ten keyword-perfect candidates turns out to be group-owned on inspection. The specimen shows you those exclusions rather than hiding them.
Explore the live specimen| ID | Class | Founder | Evidence |
|---|---|---|---|
| Target H-01 | fit | family | "family-owned and operated" — About page |
| Hidden H-02 | keyword-missed | founder | founder story, no category keyword on the homepage |
| Excluded H-01 | excluded | — | group-ownership language — already consolidated |
Commercial-side trades with inspection cadences and contract language keyword tools cannot parse.
Asset-light books, niche lanes and value-added services — legible only on the operator's own site.
A screening vendor is only as credible as what it declines to infer. Every vertical page on this site carries the same constraints, documented in our standards.
People businesses whose websites disclose licensure, accreditation and client mix.
Recurring-revenue models where the website is the product's own best disclosure.
The method never changes. What changes per vertical is which website signals carry weight — ASME stamps here, carrier networks there, accreditation scopes elsewhere.
From 100M+ domains to the vertical's full population — 367,478 in our industrial reference run.
Fast classification confirms live operating companies — 25,000 became ~17,300.
Every surviving site is read against the thesis, signal by signal, with quotes captured.
Mandate Fit 70%, Outreach Suitability 20%, Transition Context 10% — zeroed on group ownership.
We decline engagements in consumer-captive care verticals — veterinary, dental, medical practices, home health and similar. For dual-sided trades we screen the B2B and commercial side only. The reasoning is in our standards.
Almost certainly. These 25 pages document verticals we know well, but the underlying universe spans 700+ industry categories across 100M+ domains. A new vertical means writing a thesis prompt and selecting the category slice — not building new infrastructure.
Send a one-paragraph thesis and we will tell you what the population looks like.
It marks the ten verticals where we have already published an anonymized sample run: 8 top fits, 5 keyword-missed fits, 5 documented exclusions and 2 insufficient-evidence cases, each with verbatim site quotes.
Those pages link their specimen directly, and the master version lives at the specimen report page.
From a real US-focused screening run over one 367,478-domain industrial category. After triage confirmed roughly 17,300 live operating companies, deep extraction left the counts shown — for example 702 eligible independent machining shops and 93 surface finishers.
They are measured outputs of a documented run, not market-size estimates.
The opposite. In a 93-company pool, sampling is fatal: every missed company is a measurable share of the market. Exhaustive coverage matters most where populations are small and fragmented, because each additional confirmed independent changes your process.
That is why filtration and surface finishing carry specimens despite being the smallest pools.
Yes. A thesis often spans adjacent verticals — compressed air plus filtration plus boiler service is a common industrial-services combination. We screen them as one universe with per-vertical signal tuning, and custom ICP re-runs on any subset are included at every pricing tier.
Signals and counts reference our most recent documented runs, and annual monitoring clients receive refreshed universes with delta reports on a schedule. When a new specimen run changes a number materially, the page is updated with it.
The counts you see are dated artifacts of real work — we prefer that to rounded marketing figures.
Tell us the thesis in one paragraph. We reply with the population shape, the signals we would tune, and a specimen so you can judge the evidence standard yourself.
Start with one email