Compressor houses sell uptime, not keywords — and their websites read like it. We screened an entire 367,478-domain industrial category to isolate every independent US operator, with the evidence attached.
Definitions decide universes. Before any scoring, the thesis fixes which operating models belong in yours — and which nearby lookalikes do not.
Companies whose visible revenue centers on compressed air or vacuum systems: sales and installation, field service and repair, air treatment, and rentals. Most run a distributor-plus-service model around one or more compressor brands. Pure equipment brokers and OEM-owned branch networks are excluded by design.
Typical subsegments we classify separately, so your ICP can include or drop each one:
A compressor house in Missouri describes itself by the brands it services and the plants it keeps running. The phrase "compressed air services" may never appear on its homepage.
These are actual entries from the published compressed air run — the deliverable carries the company names and source URLs; the public specimen does not.
"founded in 1983 by the founder and the co-founder… over 40 employees"
"an industrial service provider, founded in 1978… degreed mechanical engineers on staff"
"The founder purchased the company with a second-generation owner in 2003"
Excluded: already part of a group/consolidator — documented before it reached your list
The method is the same for every vertical; the tuning is not. Here is how it ran for compressed air.
From 100M+ classified domains, the industrial category slice: 367,478 domains worldwide.
A fast screen over 25,000 US-focused domains confirmed roughly 17,300 live operating companies.
Each survivor read against the thesis: 15 signals, verbatim quotes, source URLs captured.
Mandate Fit 70%, Outreach Suitability 20%, Transition Context 10% — zeroed on group ownership.
Fifteen signals run on every company. In this vertical, these six do the sorting work.
Named brand authorizations tell you the product line, the territory logic, and the switching costs. We capture the exact claim — "authorized distributor since 1994" — not a guessed affiliation.
PM agreements, leak surveys and air audits are the recurring-revenue spine of a good compressor house. Their language is unmistakable once a model actually reads the service pages.
An owned rental fleet signals capital discipline and customer stickiness. Fleet pages disclose sizes and horsepower ranges that never reach any structured database field.
Round-the-clock response commitments separate plant-critical service operations from box movers. We log the commitment and the coverage area it claims.
Roughly half of confirmed industrial fits carry explicit founder or family language. We quote it — "founded in 1983 by the founder" — and never infer beyond what the site states.
About one in ten keyword-perfect candidates is already owned by a consolidator. Acquisition-program language zeroes the score and moves the company to the documented-exclusion list.
A classification you cannot audit is an opinion. Each row in the deliverable pairs every extracted signal with a verbatim snippet and the page it came from — inclusion or exclusion alike.
That standard is published in full in our standards, and it applies to this vertical exactly as to the other 24.
Structured fields plus quoted proof — nothing summarized beyond recovery.
Leading company databases index the companies they found. In a 276-company universe, what they missed is the deal you never saw.
| Dimension | Profile-based databases | Full-web screening |
|---|---|---|
| Starting population | Companies previously profiled | Every active domain in the category — 367,478 |
| The plain-spoken website | Missed without category keywords | Read and classified on what it actually does |
| OEM branch vs independent | Not distinguished | Ownership language checked, exclusions documented |
| Rental & PM revenue signals | No structured field exists | Quoted from service and fleet pages |
| Custom ICP re-screens | New filters over the same rows | Re-run of the full universe, included at every tier |
| Audit trail | Vendor's word | Verbatim snippet + source URL per claim |
Trust in the positives requires candor about the limits. Three that matter in this vertical.
Websites do not disclose revenue, and no page signal predicts a decision to sell. We refuse to estimate either — vendors who do are guessing with your credibility.
A five-page site from 2011 yields thin evidence. Such companies land in the insufficient-evidence bucket — flagged honestly rather than scored on imagination.
Universes drift as companies sell, close and launch. Annual monitoring re-screens your universe and ships deltas; a one-time run is dated the day it lands.
The deliverable is built to drop into an existing origination stack, whether you run a search fund process or a platform add-on program.
Pricing is public: proof project from €4,900, full universe with deep shortlist from €9,900, annual monitoring from €18,000 per thesis — details on the pricing page.
"A fifth or more of confirmed fits never use the category's obvious keywords. In compressed air, the best independents talk about uptime, plants and response times — which is precisely why lists built on keywords are short."
It is the count of independent, thesis-eligible operators surviving evidence review in our documented run — after removing OEM branches, group-owned distributors and out-of-scope models from roughly 17,300 live companies.
The broader population is larger; the eligible pool is not. That distinction is the deliverable.
Yes — it is a first-class classification, not a keyword guess. Sites are read for how revenue visibly arrives: equipment sales pages, service-agreement language, rebuild capabilities, rental fleets.
Your ICP can then include hybrids, exclude pure resellers, or weight service mix however your thesis requires.
Authorization language differs from ownership language, and we quote both. "Authorized distributor for" keeps a company eligible; "a division of" or acquisition announcements zero the score.
Each exclusion ships with its disqualifying quote, so you can audit every removal.
The public specimen shows 8 top fits, 5 keyword-missed fits, 5 documented exclusions and 2 insufficient-evidence rows with anonymized labels and verbatim quotes.
Client deliverables add company names, domains, source URLs and the full 15-signal extraction per row.
Yes. Adjacent verticals like filtration and boiler and steam services come from the same source category, so a combined universe is one engagement with per-segment tuning.
Custom ICP re-runs on any subset are included.
Each run is dated, and the specimen's run date is printed on it. For an active mandate we re-verify the eligible set at delivery.
Annual monitoring keeps the universe current with scheduled re-screens and delta reports — new entrants, ownership changes, gone-dark domains.
The specimen is the argument: real classifications, real quotes, documented refusals. If the evidence standard holds up to your scrutiny, the full 276 are one engagement away.
Open the specimen report