Why Your CRM Shows 1 Record for a Company That Runs 87 Plants

Your CRM shows one record for Greif. Greif runs 118 US facilities across 30 states. Your CRM shows one record for Berry Global. Berry operates roughly 120 US manufacturing sites. Your CRM shows one record for Sonoco. Sonoco runs approximately 300 operations worldwide, more than 65% of revenue US-based.

If you are an industrial sales rep or a sales ops leader trying to understand why your account coverage looks complete and your pipeline still misses, this is the structural problem. Every major B2B database — ZoomInfo, Apollo, Cognism, D&B Hoovers — is built on a company-record architecture. One company equals one row. The row is enriched with a headquarters address, an executive contact list, and a rolled-up revenue estimate. The 86 plants that company operates in 25 other states are not individually indexed. They are invisible.

This post is for industrial sales reps who feel the gap when they pitch Fortune 500 accounts, and for sales ops leaders who are trying to figure out why territory assignments look balanced on paper and unbalanced in the field. It explains why the one-record-per-company data model exists, which real companies break it most violently, what the sales motion actually loses, and what facility-first indexing looks like instead.

Why the one-record model exists

The dominant B2B databases were not built for industrial sales. They were built for software sales.

In a SaaS sales motion, one record per company is correct. If you are selling a seat-based product to Shopify, the buyer is the Shopify procurement team in Ottawa. You do not care whether Shopify has a regional office in Austin — nobody at the Austin office is approving a software contract for the parent company. One record, one buying committee, one deal.

That architecture was then sold into the industrial segment without modification. The problem: industrial buying happens at the plant. A maintenance director in Massillon approves the MRO vendor that services that specific plant's equipment. A plant manager in Bloomington approves the cleaning contractor that works in that specific plant. Corporate procurement sets master agreements, but the spend is authorized, released, and renewed at the facility — often by someone whose name never appears in any standard B2B database.

The result is a data model with a foundational mismatch. Corporate registrations — the raw source material most databases crawl — track legal entities. A holding company that operates 80 plants may be registered as a single Delaware LLC. Its operating subsidiaries might be filed under four state business registrations, none of which carry the parent brand name. The gap between "legal entity" and "operating facility" is where standard enrichment tools fall apart.

The real companies that break the model hardest

The one-record problem is worst at the large diversified industrial conglomerates — the rollups, the acquirers, and the vertically integrated manufacturers. A few named examples make it tangible:

Greif (Delaware, OH). A packaging and industrial-container manufacturer. The company reports operations at more than 250 facilities across 37 countries. In the US alone, the public-facing footprint runs to more than 100 plants across at least 30 states. When Greif acquired Caraustar Industries in 2019, the deal folded in roughly 40 additional US operating sites, many of which continued to operate under the Caraustar name on the plant door. A typical CRM enrichment on "Greif" surfaces the Delaware, Ohio headquarters. It surfaces exactly zero of the Caraustar-branded plants.

Berry Global (Evansville, IN). A plastics-packaging giant that has made more than 40 acquisitions over three decades, including the $2.45 billion AVINTIV deal in 2015, the $765 million AEP Industries deal in 2016, and the $6.5 billion RPC Group deal in 2019. The company operates roughly 250 plants globally, around 120 of them in the United States. Every one of those acquisitions brought in a fleet of plants that had their own brand names, their own plant managers, their own local purchasing contacts. Those plants do not appear in a standard enrichment pull on "Berry Global."

3M (Saint Paul, MN). The diversified manufacturer operates more than 70 US production facilities — Maplewood, Hutchinson, Brookings, Columbia, Decatur, Cordova, Guin, Cottage Grove, Cynthiana — a footprint that stretches across nearly every region of the country. A CRM record for 3M gives the rep one Minnesota headquarters and the CEO's name.

Sonoco Products (Hartsville, SC). An industrial packaging company with roughly 300 operations worldwide across more than 35 countries. US operations generate the majority of revenue, spread across more than 100 plants and converting facilities. The Hartsville HQ is in a CRM. The 100+ plants are not.

Illinois Tool Works (Glenview, IL). A decentralized industrial conglomerate that operates through 85 divisions across more than 500 facilities globally. ITW has historically acquired between one and three companies every year, most at the $100M–$500M range, each one adding a handful of plants that typically continue operating under the acquired name. The ITW CRM record shows one Glenview address. The 250+ US plants operate under dozens of division names.

Parker Hannifin (Cleveland, OH). Closed the $4.3 billion Clarcor acquisition in February 2017, which folded in Baldwin Filters (Kearney, NE), Racor (Holly Springs, MS), AAF International (Louisville, KY), and dozens of other filtration-specific operating sites. Five years later, a CRM enrichment on "Parker Hannifin" will still frequently surface the Cleveland HQ and miss the Kearney, Holly Springs, and Louisville plants — because those plants were never legally merged into a Parker-branded corporate entity.

Magna International (Aurora, ON). A Tier 1 automotive supplier with more than 340 manufacturing operations worldwide, dozens of them in the US Midwest and South, many operating under acquired brand names that never got corporate-rebranded. Magna's US plant footprint is the single most painful example of HQ-centric data architecture in automotive supplier sales.

The pattern: every time one of these parents acquires another company, the acquired plants continue to operate under their old names on the plant door. Corporate integration absorbs the P&L, but the operating brand at each site stays put — because the relationships with local suppliers, the UL certifications on the equipment, and the state regulatory filings are all tied to the original entity name.

This is why HQ-centric enrichment tools fall further behind every year. The acquisitions keep happening. The data models do not know how to follow them.

What the one-record model costs the sales motion

When the CRM holds one record and the real account is 50 or 100 facilities, the cost shows up in five specific places.

1. You cannot pitch what you cannot see. A rep's outbound motion is gated by what is in front of them in the CRM. If Greif appears as one Delaware record, the rep pitches Delaware. The 117 other Greif sites — each with a maintenance director, a plant manager, and a local purchasing process — get zero outreach, because the rep has no reason to believe they exist as distinct buying units.

2. Account coverage metrics lie. Sales ops leaders build coverage dashboards: "We have active conversations at 80% of our target accounts." The metric treats Greif as one account. If you have one conversation with one corporate procurement person at Greif HQ, you show as "covered." You are covered at one-hundred-eighteenth of the real account. Every competitor running the same coverage metric has the same blind spot.

3. Territory assignments become unworkable. A sales ops leader splitting a national team across 8 reps assigns "Greif" to one rep — typically the rep whose geography matches the Delaware headquarters. That rep is now on the hook for 30-state coverage, which means plants in Oregon, Texas, Georgia, and Maine are either left uncovered or are called by a rep who has no physical proximity to the plant. Meanwhile the rep covering Ohio has Greif HQ plus 12 other Greif operating facilities in their physical territory — a completely different workload from the rep covering Nevada, whose assigned "Greif" amounts to zero physical plants.

4. Forecasts treat account churn wrong. Churn at the plant level happens invisibly. A facility closes. A division is divested. A purchasing contract moves from one vendor to another at one specific site. If the CRM holds one Greif record, none of those events register — you only see the impact three quarters later when the aggregated revenue from Greif declines and nobody can pinpoint why.

5. Expansion revenue stays hidden. The single most underused revenue source in industrial sales is cross-plant expansion at an account where you already hold one or two plants. You are the preferred vendor at Greif Massillon. You do not know Greif also operates plants in Longview, Meridian, Dallas, Winston-Salem, Florence, or Atlanta. The expansion motion that should be automatic — "we are doing X at your Ohio facility, can we scope the same at your Texas and Georgia facilities?" — is not initiated because the rep does not know those facilities exist as part of the same parent.

Gartner reports that poor data hygiene is one of the leading reasons forecasts miss by more than 10%, and that companies that improve CRM data hygiene can increase forecast accuracy metrics by up to 30%. The mechanism is simple: if the CRM is holding one record where the account really has 100 plants, no amount of pipeline-coverage discipline can compensate. You are forecasting against a synthetic version of the account that does not match the buying surface in the field.

What facility-first indexing looks like

The fix is not to bolt "branches" onto a company record. That has been the standard workaround, and it does not work — because a branch record in a HQ-centric schema is still treated as a child of the company, not as a first-class record with its own industry classification, its own employee count, its own contacts, and its own deal history.

The fix is to index physical facilities as the atomic unit of the database, and let parent-company relationships be a relational layer on top.

That inversion changes what the rep sees on every query. Searching "Greif" on a facility-first database returns 118 records, each one a US plant with:

A verified physical address — not a corporate registration
An employee count specific to that plant (not the parent total)
An industry classification for that specific site — a steel-drum plant, a fibre-drum operation, a reconditioning facility, a tube-and-core plant all get classified independently based on what that plant actually produces
A product list specific to that facility
Decision-makers tied to that site — plant manager, operations director, maintenance manager, local purchasing — not a corporate VP list pinned to Delaware

When the rep pitches, they pitch the plant. When the sales ops leader assigns territories, they assign facilities — which lets them balance rep workloads by actual account density, not by how many corporate names happen to appear in one geography.

Facilities Finder is structured this way. Our AI ingests billions of public signals — satellite imagery, map providers, company websites, EPA filings, permit records, trade publications — and extracts what actually matters at the plant level: products, capabilities, employees, certifications. The output is 35,000+ AI-generated industry classifications and 7 million+ products, all indexed per facility. The parent-company rollup is a relational layer on top of the facility records, which means one search on "Greif" returns all 118 US facilities across 30 states — not just the Delaware HQ. No other B2B database does this at scale.

What this means for account reviews

If you run quarterly account reviews for your top 50 industrial accounts, rerun them with facility-level data as the input. The surface area of each account will change dramatically.

An account that looked like "one conversation with one VP at Greif HQ" becomes "one conversation with one VP, and 117 plants where we have not made contact — of those, 23 are in our Southeast rep's territory, 18 are in our Midwest rep's territory, and 14 are in our Southwest rep's territory." The account moves from "covered" to "underpenetrated" — which is the correct state.

The parent-rollup pattern generalizes across industrials: Tyson Foods (over 200 production and processing sites), JBS (several dozen US beef, pork, and poultry plants), Cargill (dozens of US processing, crushing, and feed facilities), Nucor (approximately 300 operating facilities across 44 states), Steel Dynamics (roughly 80 US steel, fabrication, and recycling operations), Koch Industries (a sprawling network across chemicals, paper, refining, and ranching), Emerson Electric (multiple US manufacturing sites across automation, climate, and tools businesses), Dover (hundreds of US manufacturing sites across its operating companies), and Tyson's, Cargill's, and Smithfield's combined presence in food processing alone accounts for hundreds of plants that are invisible in HQ-centric enrichment.

FAQ

Why doesn't ZoomInfo or Apollo just add all the plants?

The underlying data acquisition model is built around company registrations, executive contact scraping, and web crawls of "About" pages. That pipeline does not produce facility-level records. To add plants as first-class records, the entire schema would have to be rebuilt — every query, every filter, every integration would shift. It is a foundational re-architecture, not a feature addition. Incumbents tend to not rebuild foundations.

Can't I just use D&B's family tree?

D&B's DUNS hierarchy is the closest structured parent/child dataset that most sales teams have access to without a specialized tool. The problem is that DUNS numbers are assigned to legal entities, not to operating plants. A single LLC may run 20 plants under one DUNS. Branch records in the D&B hierarchy represent addresses that were on file at some point — not verified current operating status. For an acquisitive parent like Berry Global or ITW, the family tree is always months to years behind the real footprint.

How is this different from pulling a list of subsidiaries from an Exhibit 21?

Exhibit 21 in a 10-K filing lists significant subsidiaries by name and state of incorporation. That tells you which legal entities exist. It does not tell you how many plants each subsidiary operates, where those plants are, or who runs them. For private companies — the majority of US manufacturing by count — there is no Exhibit 21 at all.

What about the plant manager? LinkedIn finds me those, right?

Sometimes. LinkedIn is effective when a specific plant manager has a complete, recently updated profile and a job title that includes the plant city. It is not effective when the plant manager is a 25-year veteran whose profile says "Plant Manager at Greif" with no city, or when the manager does not use LinkedIn at all — which is still common in older industrial operations. A facility-first database indexes plant leadership from multiple sources beyond LinkedIn, including public filings and industry publications.

Start with your highest-value multi-plant account

If you are selling to industrial companies, the cost of HQ-only data is measurable: accounts pitched at one-hundredth of their real footprint, plant managers you never found, forecasts that miss because the buying surface in the field is 100 plants wide when the CRM only holds one record.

Facilities Finder indexes every US industrial facility as its own first-class record — with employee count, AI-generated industry classification, and contacts keyed to that specific site. The parent-company rollup connects every plant, distribution center, and reconditioning facility back to the parent ID, so one search on "Greif" returns all 118 US facilities across 30 states, each with its own plant manager, operations director, maintenance manager, and purchasing contacts. Every plant, branch, and warehouse — not just HQ.

Facilities Finder covers 600,000+ US industrial facilities and 25 million+ decision-maker contacts across all 50 states.

Look up your next target account's full footprint — start the parent-company search.