TL;DR
- Choose Dun & Bradstreet when the account-planning problem is connecting commercial sites, subsidiaries and ultimate parents through a common business identifier. Choose Moody's Orbis when ownership detail, financial context and cross-border legal entities drive the decision.
- GLEIF is the strongest free starting point for LEI-registered entities and accounting-consolidation relationships. It does not describe every company or establish who controls a purchase.
- OpenCorporates helps resolve registry identity and check provenance. Treat it as an evidence layer rather than assume every record arrives with a complete commercial family tree.
- Keep legal ownership, operating structure and sales territory as separate relationships. Collapsing them into one CRM parent field creates avoidable account conflicts.
- Buy accepted, current relationships at the entity level you sell to. A large company count or a visually impressive tree is insufficient.
Three trees that should remain separate
An enterprise can own a German subsidiary, publish all its products under one brand and buy a particular service through a regional operating company. Each statement describes a different structure. Corporate hierarchy data becomes useful when it preserves those distinctions.
The legal ownership tree identifies entities and their relationships. The operating tree describes divisions, branches, plants and business units. The commercial tree represents who buys, which contract applies, where a sales team has rights and which owner coordinates the relationship. External data can inform all three; it should not silently dictate the third.
Consider a hypothetical industrial group with a US parent, a UK distribution subsidiary and a German manufacturing subsidiary. The group might negotiate security software centrally while buying warehouse services locally. Rolling both opportunities to the ultimate parent improves exposure reporting. Assigning both to the same seller without reference to the product and contract can be wrong.
Create separate fields or relationship objects for legal parent, ultimate group, local selling entity and commercial account owner. Attach the relationship type, source and observation date. A parent value without its meaning is difficult to reconcile later.
Our broader company-data provider guide covers general account records. This guide focuses on which relationships to buy and accept, rather than re-ranking firmographic databases.
Four evidence sources compared
| Source | Strongest buying reason | Identifier and relationship basis | Main limitation | Commercial basis |
|---|---|---|---|---|
| Dun & Bradstreet | Build commercial account families and link sites | D-U-N-S and corporate linkage | Legal linkage still needs a separate buying-authority policy | Scope-specific data/API agreement |
| Moody's Orbis | Investigate ownership and financial context across jurisdictions | Company reference data and ownership structures | Filing availability and relationship dates vary | Scope-specific licence |
| GLEIF | Obtain an open legal-identity/consolidation baseline | LEI and reported accounting parents | LEI population and reporting exceptions limit completeness | Public LEI data accessible free |
| OpenCorporates | Resolve registry entities and inspect their sources | Jurisdiction plus company number, sourced records | Registry identity alone is not a full buying-group map | Commercial API or bulk licence |
These are documented buying differences checked on 6 October 2026. They are not a four-provider accuracy benchmark. A provider's global coverage claim cannot settle coverage for your particular subsidiaries and countries.
Dun & Bradstreet: commercial linkage
D&B is a sensible first shortlist when a revenue team needs to move between a known business, its domestic group and its global family. Its ownership documentation distinguishes parents, subsidiaries and branches; its family-tree report connects entities through D-U-N-S identifiers.
That makes the data useful for deduplicating apparent account coverage. Ten local records belonging to one group may represent ten operating opportunities, one strategic relationship, or both. Retaining local identifiers while adding a group key supports either view without deleting the local accounts.
The purchase should specify which linkage product, entity population and delivery method are included. A report a researcher can inspect is different from an API feed the CRM can refresh. Price initial matching, relationship retrieval, exports, ongoing updates and permitted storage together. No universal per-account price is asserted here.
Choose D&B when commercial entity linkage and consistent business identifiers are the core requirement. Do not choose it solely to infer buying authority. A correct group relationship still does not prove that the head office buys your category or that a local entity is covered by the parent agreement.
Moody's Orbis: ownership depth
Orbis combines company reference data, financial context and ownership structures. That makes it a stronger candidate when the question includes intermediate owners, cross-border entities or whether a relationship is based on ownership rather than a shared brand.
For enterprise prospecting, the practical benefit is a more defensible account thesis. A seller can distinguish an operating subsidiary from a holding company and identify a relevant group relationship before investing in a complex pursuit. Financial context can also help determine which entity deserves further research. It should not become a claim about the prospect's budget.
Scope the licence around the countries, company types, relationship fields and historical data you need. An interface subscription, an API and a bulk dataset are different purchases. Request terms for CRM storage, refresh, analyst access and any client-facing use instead of assuming a research seat permits redistribution.
Choose Orbis when ownership research is valuable enough to justify specialist data and interpretation. Avoid buying its breadth for a narrow territory problem already resolved by confirmed CRM entities. More links are useful only when their definitions and dates improve the decision.
GLEIF: an open consolidation baseline
GLEIF Level 2 data records direct and ultimate accounting-consolidating parents for entities in the LEI system. Public access includes search, downloadable files and an API. This is a credible free starting point for enterprises and financial-sector account sets with substantial LEI coverage.
The relationship has a specific meaning: accounting consolidation. It is not a comprehensive map of operational reporting lines, procurement teams or all shareholdings. Preserve reporting exceptions rather than replacing a missing parent with the closest similarly named company. “No reported parent relationship” and “independent company” are different conclusions.
A useful workflow begins with a known legal identity, matches the LEI and stores the relationship status alongside the parent. Review the registration and update context before using it for current planning. The open licence removes a data-subscription line; engineering and exception handling still cost money.
Start with GLEIF when an open baseline answers a substantial part of the target market. Supplement it for uncovered entities and operational buying levels. Do not reject a real subsidiary merely because it lacks an LEI or a published parent link.
OpenCorporates: registry identity and provenance
OpenCorporates' API provides legal-entity data with sources. Jurisdiction and company number are particularly useful when a trading name, shared website or abbreviation makes a name-only match unsafe. Its role here is checking the identity behind a proposed relationship.
The public pricing page lists Essentials at £225 monthly or £2,250 annually for up to 500 monthly API calls, Starter at £660 monthly or £6,600 annually for 2,500, and Basic at £1,200 monthly or £12,000 annually for 5,000. Those are call allowances, not the same number of fully researched hierarchies. A search followed by a detail request consumes a different workload from one lookup against a known identifier.
There is a consequential licence issue. The September 2026 self-service terms distinguish internal use and external distribution, and address records marked restricted_for_marketing. A commercial subscription does not erase the source jurisdiction's restrictions. Have the permitted use confirmed for a prospecting deployment.
Choose OpenCorporates for sourced registry identity and investigations. Use another relationship source where a complete parent chain is required and not available in the relevant records. A reliable company match is valuable even when it deliberately leaves ownership unresolved.
A thousand-account acceptance exercise
Freeze 1,000 account inputs and 200 proposed parent links. Include shared domains, recent acquisitions, similarly named firms, branches, renamed entities and subsidiaries in the countries you sell to. This is a proposed exercise, not a reported run.
Each candidate link should have three acceptance decisions:
| Decision | Accept when | Do not accept when |
|---|---|---|
| Entity identity | Returned identifier describes the intended legal or operating entity | Only the name or group domain matches |
| Relationship | Parent identifier, relationship type and dated support are consistent | Link is inferred from branding or stale ownership |
| Commercial usefulness | Link supports the intended roll-up without changing the local buyer incorrectly | Correct legal ownership is used to force an unsupported territory assignment |
Suppose the review labels 160 links supported, 25 ambiguous and 15 wrong. Acceptance is 160/200, or 80%, in this hypothetical sample. Keep ambiguous separate from wrong. Neither should trigger automatic reparenting.
Then mark which supported links help a real action: group exposure, coordinated outreach, local ownership or contract coverage. If only 120 do, the commercial output is 120 useful links. The broader 160 remains useful research, but it is not the denominator for a purchase justified by actionable account planning.
Use the ABM account-scoring template to record the selling entity and qualification policy before adding hierarchy-based rules. For resolving a person to the correct CRM account, continue with lead-to-account matching. That is a downstream decision, not a substitute for defining the tree.
Budget the useful relationships
Let the allocated monthly licence be £600, review 200 links for three minutes each, and value analyst time at £50 per hour. These are assumptions, not provider prices. Review labour is 10 hours × £50 = £500. If integration and maintenance add £150, the total is £1,250.
At 160 supported links, cost is £7.81 per supported link. At 120 commercially useful links, it is £10.42 per useful link. The same invoice can look materially different depending on the denominator.
If review time rises to six minutes, labour becomes £1,000 and total cost £1,750, or £14.58 for each of 120 useful links. That sensitivity makes the provenance and ambiguity queue a buying criterion, not a technical afterthought.
For public API tiers, count every required call and respect daily caps. Do not assume a 1,000-account project fits a 500-call allowance merely because some companies return no useful relationship. Also separate initial linkage from refresh: a dataset that must be repurchased or heavily reviewed every month has different economics from a maintained relationship feed.
Maintain the tree without rewriting territories
Store the incoming relationship as a proposed revision. A confirmed acquisition can update the legal group while leaving the active opportunity owner unchanged until the commercial policy is reviewed. Preserve the earlier relationship and effective date so historic pipeline reporting remains intelligible.
Route disagreements to the person who owns account identity, with the exact entities and supporting source. A domain change should prompt a check rather than an automatic group merge. Refresh high-value and recently changed accounts more often than dormant records, subject to the licence and workload.
Our recommendation is to start with a source appropriate to the required tree, then prove acceptance on difficult entities. Combine open identity evidence with paid linkage only where the incremental coverage or interpretation is useful. The next step for turning those relationships into a pursuit is account-based sales research.
FAQ
Is the ultimate parent always the correct sales account?
No. It is useful for group reporting and coordination, but the buyer may be a subsidiary, branch or operating unit. Keep the group key separate from the account that owns the opportunity. Let the commercial policy determine territory and contract coverage.
Can GLEIF replace a commercial hierarchy provider?
It can replace paid work where the required entities and consolidation relationships are covered. It will not establish complete operating structures for every target. Compare the accepted coverage of your actual account list before buying a supplement; a free baseline can narrow the paid scope substantially.
Should subsidiaries sharing a website be merged?
Usually not merely because of the shared website. Confirm legal identity, local operations and active CRM relationships first. A common domain is a useful candidate signal but can describe an entire group. Preserve separate account IDs when the entities or buying decisions differ.
How should recent acquisitions be handled?
Retain the announcement or filing, relationship type and effective date. Separate announced, completed and reflected-in-data states. A transaction can change legal ownership before systems, brands or procurement are combined. Existing opportunities should not be reassigned on a news headline alone.
What should an enterprise data agreement include?
Specify entity levels, jurisdictions, delivery, update rights, storage, exports, external use and source restrictions. Price matching and relationship retrieval together with refresh. The relevant unit is the useful, permitted relationship, not an undifferentiated company count.
