Episodic list-buying finds the companies everyone already knows. A standing radar starts from the entire active web and holds the complete add-on universe against your thesis — with evidence for every inclusion and every exclusion.
The platform thesis is written. The integration playbook works. What runs out is the list — and the list runs out because it was never complete to begin with.
A typical mandate starts with a few hundred names from banked deals, conferences, and database exports. Eighteen months in the pipeline feels “picked over” — but what is picked over is the visible slice, not the market.
Leading company databases start from companies someone already profiled. Small, founder-run operators in niche subverticals — exactly the add-on sweet spot — are the ones most likely to be missing.
A one-off long list is a photograph. Companies get acquired, pivot their service mix, or professionalize between your purchase and your outreach — and the list quietly rots on the shared drive.
If your universe came from the same tool your competitors subscribe to, your proprietary pipeline is nothing of the kind. Differentiated sourcing needs a differentiated denominator.
We separate cheap breadth from expensive depth. The first pass can afford to look at everything; the second pass earns its cost only on plausible candidates.
We start from 100M+ classified domains — 24.7M of them business and finance sites across 700+ industry categories. One industrial category alone held 367,478 domains globally.
A fast screen removes parked domains, directories, and non-operators. In our specimen run, a 25,000-domain US-focused triage left roughly 17,300 live operating companies worth a closer look.
Survivors get full-site LLM analysis against your written thesis: 15 signals, verbatim evidence snippets, source URLs. Nothing is inferred that the company did not publish itself.
Every company carries three scores — Mandate Fit 70%, Outreach Suitability 20%, Transition Context 10% — zeroed out on group ownership. The result is a ranked, defensible universe.
We do not screen against a category label. We screen against your exact add-on criteria, written down: subsector boundaries, service mix, customer types, geographic footprint, certifications that matter to your integration model.
That document becomes the instrument every domain is scored against — and when your thesis shifts after the next platform acquisition, we re-run the same universe against the new ICP. Custom re-runs are included in every engagement.
Every filter maps to a website-visible signal. Nothing is guessed.
These are not marketing round numbers. They are the working scale of the classification platform your add-on universe is cut from — the same data 300+ enterprise organisations already run on.
Every company on the radar is scored on fifteen website-visible signals. These six do the heaviest lifting for buy-and-build work.
We read for group membership, “a division of” language, and companies running their own acquisition programs. Roughly one in ten keyword-perfect candidates was already group-owned in our specimen work; the radar removes them before they waste a partner’s afternoon.
Service contracts, maintenance agreements, scheduled programs, consumables — the published language of repeat revenue. For a platform underwriting integration cost, an add-on with visible contract language is a different asset than a pure project shop, even when both carry the same category keyword.
Explicit certifications only — ISO, AS9100, ITAR, ISO/IEC 17025, ASME — captured as exact claim text with the source page. If your platform sells into aerospace or regulated utilities, an add-on’s certification page is a gating fact, and we quote it rather than summarize it.
HQ, branch locations, and stated service area, with owned locations distinguished from partner mentions. Density strategies live or die on this signal: a two-branch operator adjacent to your existing coverage is worth more than a larger company three states away.
Customer industries evidenced by case studies and named end markets — not guessed from keywords. A controls integrator serving food-and-beverage plants and one serving automotive lines read identically in a directory export; on the radar they sit in different buckets, because their own project pages say so.
Visible non-founder functions — finance, operations, HR, marketing roles named on the site. A thin bench can mean integration lift or a clean tuck-in; either way you want to know before the first call, from what the company itself publishes.
The initial map is the start of the engagement, not the end. The value compounds in the delta runs, where a static list would already be losing accuracy month by month.
Full two-pass screen of your platform’s space: universe cut, triage, deep extraction, scoring. You receive the complete eligible population — top fits, keyword-missed fits, documented exclusions, and insufficient-evidence cases, in the same 8/5/5/2 specimen format we publish.
Your deal team reacts to the first ranked slice; we tighten the ICP against those reactions. Disagreements are resolved against evidence snippets, not adjectives — if a company was scored wrong, the instrument gets fixed and the universe is re-scored.
The same universe is re-screened on a cadence. New entrants appear, acquired companies drop to excluded, dormant sites get flagged on activity trajectory. Your pipeline meeting starts from what changed, not from a stale export.
New platform, adjacent subvertical, changed size band — the written ICP is edited and re-run against the standing universe at no extra data cost. This is where owning the denominator beats re-buying lists every time the mandate moves.
All three approaches produce names. They differ on completeness, currency, and whether you can defend the list in an IC memo.
| Dimension | Standing add-on radar | Episodic list purchase | Profile-based database |
|---|---|---|---|
| Starting population | Entire active web, 100M+ domains | Whatever the vendor had on file | Companies already profiled |
| Keyword-missed fits | Found via classification, not keywords | Systematically absent | Dependent on tagging quality |
| Ownership screening | Group-owned flagged with quoted evidence | Rarely checked | Often stale after the deal closes |
| Currency | Quarterly deltas on the same universe | Decays from purchase day | Refreshed, but on their schedule |
| Thesis pivots | ICP re-runs included | Buy a new list | Re-filter the same partial pool |
| Evidence per name | Verbatim snippets + source URLs | Firmographic fields only | Aggregated profile data |
Anonymized specimens from our published industrial run. On the public site names are withheld; client deliverables carry full identities, snippets, and source URLs. The full specimen report is free.
Scored 92.7 with a five-location Indiana footprint, UL508 compliance and visible recurring offerings. Twelve of twelve evidence snippets verified against site text.
An equipment-repair operator that matches every category keyword — and is quoted on its own site as a subsidiary of a larger group. It ships in the exclusions file with the disqualifying sentence attached.
This is the fate of roughly one in ten keyword-perfect candidates, and the reason a radar documents its “no” decisions: your team should never rediscover an exclusion the hard way.
A sourcing instrument you trust is one whose limits are written down. These are ours — the same boundaries we publish for every engagement, stated before you buy rather than discovered after.
The radar ships as structured data — one row per company, one column per signal, evidence linked at cell level. It imports into any CRM your sourcing team uses, with exclusions and insufficient-evidence cases as separate, auditable files.
Deal leads take the ranked slice into outreach sequencing. IC memos cite the universe count and screening criteria directly, because the denominator is now a documented fact rather than a banker’s adjective. Quarterly deltas arrive as changesets, not fresh spreadsheets to reconcile.
Pricing is public: proof project from €4,900, full universe from €9,900, annual monitoring from €18,000 per thesis. Details here.
A database is a pool of profiles someone already collected; you filter what is there. The radar starts from the entire active web — 100M+ classified domains — screened against your written thesis, so your criteria define the population, not prior coverage.
The practical difference shows up at the edges: small, founder-run, keyword-atypical operators that never got profiled, which in our specimen work amounted to a fifth or more of confirmed fits.
Three changesets: new entrants that now meet your ICP, companies that left the eligible universe (acquired, pivoted, gone dormant), and score movements with the evidence that caused them.
Each change names its trigger — a new certification page, group-ownership language, a dead news feed — so your team reviews facts, not a re-shuffled spreadsheet. The underlying universe stays constant, which is what makes the deltas meaningful.
That is the expected case, and it is why custom ICP re-runs are included rather than sold as change orders.
You edit the written instrument — new subvertical, tighter service-mix rule, different geography — and we re-score the standing universe against it. Because the denominator is already built, a re-run is fast and does not reset your pipeline history.
Only from what is published: “a division of,” “acquired by,” group navigation, investor pages, or the parent’s own site.
Each exclusion carries the quoted sentence and its source URL, and companies with no such evidence are not assumed independent — independence language is itself one of the fifteen signals. Roughly one in ten keyword-perfect candidates in our specimen category failed this screen.
No, and we would not trust anyone who says otherwise. Willingness to transact is not a website-visible fact, and we do not manufacture it.
What we can show is transition context from published evidence — founder-associated, long-established, independently positioned businesses with an identifiable decision-maker and limited visible leadership bench. That framing is deliberate and it is the boundary of the claim.
Public pricing: a proof project from €4,900 to validate the approach on your space, the full universe with deep shortlist from €9,900, and annual monitoring from €18,000 per thesis with the quarterly cadence described above.
Against two or three episodic list purchases a year — each partial, each decaying — the standing universe is usually the cheaper instrument per qualified name reached.
Send us the thesis. We will show you the eligible universe it actually implies — complete, evidenced, and kept current — starting with a proof project on one subvertical.