Home Use cases Family Office Direct-Investment Pipeline
Use case

A direct-investment pipeline built the way families actually buy

Referral networks show you what intermediaries choose to circulate.

A mapped universe shows you what exists: every independent, long-established company matching the family's criteria, researched from public websites with evidence per row — and no outreach, no market signal, until you decide otherwise.

700+
industry categories
15
signals per company
0
targets contacted by us

The structural disadvantage patient capital accepts — and shouldn't

Families hold two genuine advantages — decades-long horizons and the freedom to weight fit over financials. Standard deal channels neutralize both.

Auctions neutralize patience

Banked deals arrive pre-shopped, priced by competition, and timed by someone else's calendar. The independent, long-tenured businesses families want most rarely appear in a process at all.

Referrals carry hidden selection

Intermediaries circulate what they have mandates on and what they can fee. The result is a pipeline drawn by other people's incentives — authentic but arbitrary.

Lean teams hit a research ceiling

One principal and an analyst cannot hand-research four hundred candidates across three verticals. Patient-capital advantage converts into a staffing problem.

Databases miss the criteria families use

"A business we understand, with a real trade behind it." "Certifications its industry respects." These markers live in website sentences, not database columns — and until recently nothing could read forty thousand sites carefully.

The gap this closes: not more deal flow, but a complete, quiet census of the companies matching the family's own criteria — built once, refreshed on your rhythm, and touching no one.

Why the family's criteria defeat ordinary tools

Fit lives in language, not fields

Longevity, independence, craft pride, named principals — the markers families screen for are sentences on websites, not columns in a schema. Reading them requires reading, which is what an LLM pipeline does at census scale.

Discretion is a hard requirement

Families do not want their interest circulating. Our screen reads public pages only, contacts no one, and creates no signal of intent. The pipeline exists on your desk and nowhere else.

Lean teams need pre-done research

The deliverable moves diligence's starting line: by the time a company reaches the principal, its history, certifications, footprint, and ownership language are already quoted and sourced. Judgment starts at the shortlist, not the search bar.

A census, not a search

The method's discipline is what makes the output trustworthy enough to put in front of a principal. Profile-based databases index the companies they previously found — we start from the entire active web.

1

Full-web base

100M+ classified domains, 24.7M business sites in 700+ categories. LLM analysis runs on the family's exact criteria written in plain language — not keyword approximations.

2

Two-pass triage

Pass 1 sorts every domain: operating company vs. directory, independent vs. chain. In one specimen run, 25,000 domains yielded 17,300 live companies — the rest was noise that would have consumed a quarter.

3

Deep extraction

15 signals per survivor — services, customers, founding language, certifications, leadership, hiring — each captured as a verbatim quote with source URL. Unverifiable claims say "unverified", never inferred.

70%Mandate Fit
20%Outreach Suitability
10%Transition Context (capped)

Group-owned companies score zero outright. We do not profile individuals, estimate ages, or predict anyone's plans — a line we hold even when asked nicely. See our standards.

The signals a principal actually reads

Five of the fifteen signals do most of the work in family-office configurations.

Operating history & continued independence

Stated founding years, tenure language, and explicit independence claims are extracted verbatim. A business that has stayed independent for forty years has already made, many times, the choice the family is hoping it will make once more.

Founder-led / family-led association

Roughly half of confirmed fits carry explicit founder or family evidence — “Family-operated. Personally invested.” — and the other half are honestly labeled unverified, which is itself information. We find the companies that already tell the story; we never manufacture it.

Visible leadership bench depth

A site naming a single principal describes one kind of conversation; one naming an owner/CEO, a president, and a VP of operations describes another — continuity a patient buyer can build on. The signal informs how you approach, not whether someone “must” transact.

Compliance & regulated-market readiness

ASME stamps, ISO/IEC 17025, NADCAP, UL 508A — respected credentials do double duty as quality assurance for families without sector operating partners on staff. We capture the exact claim text so the family's advisors can verify rather than research.

Recurring-offering indicators

Maintenance agreements, service contracts, scheduled inspection programs — companies describe these economics openly, and the language is extracted per company. In service-led industrials, the split between contract-backed and project work is often visible from the services pages alone.

What one vertical looks like, enumerated

Specimen numbers from boiler, burner and steam-system services — a trade many industrial families know first-hand. One full US screen yielded 172 eligible independent companies, each entry evidenced.

Target B-01 — Maryland
Confirmed fit
  • ASME and National Board credentials
  • History page: "the founder and entrepreneur"
  • Team page: owner/CEO, president, EVP, VP of operations & compliance
Hidden fit B-01
Keyword-missed
  • Homepage lacks obvious category keywords
  • History page: "The family of entrepreneurs began with the founder"
  • Services match criteria precisely — a fifth of fits share this profile
172
boiler & steam
545
equipment repair
254
calibration & testing
513
material handling

Referral flow and the mapped universe, side by side

Not a replacement — a floor under it. The comparison families find clarifying:

Referral & intermediary flowMapped universe
SelectionOther people's mandates and incentivesThe family's written criteria, applied to everything
CompletenessUnknowable — you see what circulatesA counted census with a stated boundary
Competitive postureOften pre-shopped; price set by processUn-shopped; no signal exists until you create one
TimingThe seller's process clockThe family's clock — patience becomes usable again
EvidenceA teaser, then a data roomQuoted site evidence per company before any contact
Cost of waitingDeals expireThe universe refreshes; waiting is a strategy, not a loss

Where this stops — stated plainly

We would rather lose an engagement than overclaim. The boundaries are explicit.

No intent signals

Cannot tell you whether any owner wishes to sell. No signal on the public web supports that claim.

No financial estimates

Websites do not carry financials. A guessed number in an otherwise evidenced file would be the worst sentence in it.

No individual profiling

No age hunting, no retirement speculation, no reading of anyone's personal circumstances.

No gap-filling

Thin web presence yields flagged rows, not dressed-up ones. In the specimen format, two of every twenty entries are exactly such flags.

What we refuse to sell: no “ready to sell” flags, no revenue or EBITDA guesses, no owner-age profiling, no distress detection — and no engagements in consumer-captive verticals. Read our standards; serious buyers tell us this page is why they trusted the rest.

How offices run it in practice

The engagement begins with a criteria conversation — the family's actual language about what it wants to own, translated into screenable markers. The census is delivered as a ranked file with evidence inline.

Active pursuit — quarterly

Refresh via delta monitoring so ownership changes and new entrants arrive as structured diffs rather than surprises.

Patient build — semi-annual

Refresh every six months and let the pipeline mature alongside relationships. Waiting becomes a strategy, not a loss.

Advisor-assisted

Hand the advisor the unbranded file directly. The advisor-facing variant works equally well when the mandate sits on their desk.

Beyond sourcing

Memo-ready citations

Quoted evidence drops directly into the internal write-up — founding history, certifications, footprint, leadership — built on citations, not paraphrase.

Governance answer

A counted universe with a dated boundary answers "are we seeing enough?" in one line — the standing question every direct program faces.

Battle-tested format

300+ organisations have run engagements on this machinery, from single-thesis buyers to large European corporates. The format has been shaped by readers who audit before they trust.

Entry economics

Proof project from €4,900 · Full universe from €9,900 · Annual monitoring from €18,000/thesis. Custom ICP re-runs included.

See the specimen first

Questions families ask us

Completely, by construction. The screen reads publicly published websites — the same pages any visitor sees — and no company is contacted, emailed, or called at any stage. No list is shared with intermediaries, no interest is registered anywhere, and each family's universe is confidential to that family. Until you choose to reach out, the only place this pipeline exists is in your own files.

Partly, and we are precise about which part. Values-adjacent markers that companies publish — craft language, community tenure, named people, respected certifications, association memberships — screen well, because they are text on pages. Genuinely interior qualities like culture and integrity do not screen from websites, and we will say so rather than pretend. The file gets you to the right doors with real context; what you find behind them is the family's judgment.

Because referral flow cannot tell you what it is not showing you. The census provides the denominator — in one specimen vertical, 172 independent companies, counted — so the family knows whether its referrals represent the market's best or merely its most circulated. Offices typically keep every referral channel open and use the universe to rank inbound against the full field, which quietly improves their negotiating posture on both.

The family owns it outright — the universe, the scoring configuration, and every evidence row are engagement deliverables, unbranded by default. Advisors, operating partners, and co-investors can work from it freely on your behalf. We take no success fees, no introductions, and no position in any transaction; the engagement is research, priced as research, per the terms on our pricing page.

It is accurate as of screening, and it ages the way markets do — quietly. Ownership changes are the fastest-moving risk: in our runs, roughly one in ten keyword-perfect candidates was already group-owned, and that number only grows in consolidating trades. Quarterly or semi-annual re-screens catch acquisitions, leadership changes, and new entrants as structured deltas. For a standing pipeline we recommend the monitoring tier; for a one-time purchase decision, a point-in-time census with a dated boundary is usually enough.

See the universe before anyone knows you are looking

Request the specimen report — every company scored and ranked with full signal transcripts and every exclusion documented. Same-day, no signal to anyone.

Request the specimen report