The short version
- Your prospect list is built from HQ records, but the risk you're pricing lives at the plant: a specific building, a specific headcount, and in some lines a specific EPA file.
- Firmographic tools give you a company name and a headcount guess with no building square footage and no per-site hazard signal. LinkedIn Sales Nav finds you a title, not a facility over 75,000 square feet with an open RCRA flag.
- The fix: filter manufacturing plants by building size, facility-level headcount, and EPA/RCRA enforcement history, then reach the plant manager or EHS lead directly, matched by domain to that facility instead of guessed at the corporate level.
A renewal-season list built from a firmographic database looks the same for a 40,000-square-foot job shop and an 800,000-square-foot stamping plant that happen to share a parent company name. We've written about the HQ-record problem generally; for insurance it's sharper, because size and hazard exposure are literally what you're pricing.
The market makes that gap more expensive to ignore. NCCI's 2025 State of the Line report put workers' comp at an 86.1% combined ratio for 2024, the eighth straight year under 90%. Healthy underwriting, but net written premium fell 3.2% to $41.6 billion while indemnity severity climbed 5% and medical severity climbed 6%, even as claim frequency dropped 6%. Fewer claims, each one costing more, on a shrinking premium base. The broader commercial-lines picture backs that up: the Insurance Information Institute puts industry-wide commercial-lines incurred losses at $222.5 billion in 2023, up from $154.9 billion in 2019, even as commercial P&C net premiums written grew 10.2% that year to $857.8 billion. Holding revenue flat means writing more new business, not just retaining the book, and that's a discovery problem before it's a sales problem.
What actually finds the right accounts
- Building square footage, measured from satellite imagery rather than self-reported, as the replacement-cost proxy that ranks prospects by size instead of alphabetically.
- Facility-level headcount, not the parent company's global employee count, since that's the number that actually drives a WC exposure base.
- EPA registry status, RCRA hazardous-waste generator class, and enforcement or penalty history as an environmental-liability signal firmographic tools don't carry at all: a plant with an open enforcement action is either a harder placement or a live opportunity to replace an incumbent who missed it.
- A domain-matched plant manager or EHS contact, the person who actually knows the local claims history and current broker relationship, instead of a corporate risk-management contact three layers removed from the building you're trying to insure.
Where the data actually comes from
Facility-level data like this doesn't come from a single feed, because no single feed has it. Facilities Finder tracks over 600,000 plants, warehouses, and branches across all 50 states, each geocoded and linked to its parent company, with more than 20 structured fields per site. Twenty-five million employees are tied to the specific facility they work at, not just to a company name, which is what makes a facility-level headcount and a role-searchable contact list possible in the first place.
Building that record means combining sources that don't talk to each other anywhere else: satellite imagery for building footprint and location, company websites for what a plant actually makes or processes, EPA and state filings for permits, waste-generator status, and violation history, business registries and public records for ownership and corporate lineage, and professional networks for who works where. AI agents read every one of those sources for a given address and resolve them onto one record, so a plant's square footage, its EPA flags, and its current EHS contact all live on the same profile instead of three separate lookups.
Manufacturing is the largest single category in the database at 182,000 facilities, followed by industrial services at 168,000 and warehouse/distribution at 98,000. That matters for a P&C book because it means the exposure base a broker is trying to reach isn't a long tail of edge cases. It's the core of what's out there to write.
Sizing the risk before you pick up the phone
The facility profile itself is built around the fields an underwriter actually asks for: satellite-measured square footage, reported headcount, EPA registry status, hazardous-waste generator class, and permits held, all on one page instead of pulled from four sources by hand. Facility search filters on that same data, so a broker can pull every plant in a territory over a given square footage with an active RCRA flag in one query, or describe the same thing in plain English through AI search instead of building a filter string.
Once a target list exists, the people tab shows every employee at a site, searchable by role, so "EHS manager" or "plant manager" returns the right name instead of a generic corporate contact. Verified contact info gets unlocked once and shared with the whole team, so the next renewal cycle doesn't mean re-buying the same list.
A manufacturer's exposure lives at the plant, not the headquarters. See how facility-level data compares to headquarters-only providers in our D&B Hoovers breakdown.