Ultimate beneficial owner: how to identify and verify a UBO

An ultimate beneficial owner, usually shortened to UBO, is the natural person who ultimately owns or controls a legal entity. The word “ultimately” matters: the relevant person may sit behind several companies, partnerships, trusts or other arrangements rather than appearing as a direct shareholder.

UBO identification is often treated as a database lookup. In reality, it is a structured investigation. The analyst must resolve the correct legal entity, collect direct ownership, follow every relevant corporate shareholder, calculate indirect interests, test non-ownership forms of control and decide whether the available evidence is sufficient for the applicable policy.

That last phrase—the applicable policy—prevents a common error. There is no single global UBO threshold or definition that can be applied blindly to every company. Jurisdictions and use cases can define beneficial ownership differently. A rule may refer to ownership percentage, voting rights, the power to appoint management, control through other means, or a combination of tests. Some regimes contain special treatment for listed companies, state-owned entities, trusts or regulated institutions.

A useful UBO result therefore contains more than a person’s name. It should show:

• the verified subject company and its jurisdiction;
• each ownership or control relationship used in the conclusion;
• direct and calculated indirect interests where supported;
• the rule, threshold and policy version applied;
• source names, dates and retrieval timestamps;
• any missing branch, unavailable source or unresolved conflict;
• the date or event that should trigger a new review.

This makes UBO identification one component of a wider KYB verification process. KYB establishes which business is being assessed and whether it exists. UBO work then asks who stands behind it. Screening, risk assessment and monitoring use those resolved subjects as inputs.

Why identifying an ultimate beneficial owner is difficult

The simple case is easy: one person owns the company directly, the official record is current and the ownership percentage is stated. Enterprise cases are rarely designed to remain simple. A company can be owned by several legal entities, each registered in a different jurisdiction and each using different filing rules. One branch may end with an individual; another may end with a listed parent; a third may lead to an unavailable or restricted record.

Five problems create most of the operational difficulty.

Entity ambiguity

A correct ownership calculation attached to the wrong company is still wrong. Similar names, translated names, old addresses and group branding can cause an analyst to select a parent, branch or sister company instead of the legal entity in the transaction. Start with registration number and jurisdiction whenever possible, then preserve the exact match logic.

Ownership depth and branching

Every corporate shareholder creates another path to investigate. The number of records can expand quickly when a company has several owners and those owners are themselves held by multiple entities. A graph view helps the reviewer see the structure, but it does not replace the calculation or the source record behind each edge.

Different disclosure models

Registries vary in which relationships they disclose, how often records are updated, whether historical filings are available and whether a declared beneficial owner is verified. The absence of a published owner is not proof that no owner exists. It is a statement about evidence coverage.

Control without a simple shareholding percentage

Voting agreements, appointment rights, partnership arrangements or other mechanisms can create control that is not visible from ordinary shares alone. A workflow that searches only for a percentage threshold can miss the person who actually directs the entity.

Rules that change independently of commercial risk

Regulatory filing requirements can change while an enterprise’s need to understand counterparties remains. On 11 August 2026, FinCEN made permanent broad exemptions from federal beneficial ownership information reporting for US-created companies and US persons. That alters who must report under that regime; it does not provide an ownership answer for a bank, marketplace, insurer, investor or procurement team assessing a relationship.

A seven-step method to identify a UBO

The process below is designed for repeatability. It can be used manually for investigations or encoded into an API-led workflow, but the evidence requirements should remain the same.

1. Resolve the subject company

Confirm legal name, registration number, jurisdiction, entity type, registered address and current status against an authoritative source. Record trading names separately. If several entities remain plausible, pause the ownership work until the business is disambiguated.

2. Set the applicable ownership and control rules

Before calculating anything, define the purpose of the review and the rule set that applies. Store the threshold, relevant control tests, exemptions, look-through requirements and treatment of listed or state-owned entities. This prevents analysts from applying a familiar percentage to the wrong case.

3. Collect every direct ownership relationship

Capture all available direct shareholders, their entity types, stated interests, share classes, voting rights and source dates. Do not follow only the largest shareholder. Several smaller indirect paths may converge on the same person and cross the relevant threshold when combined.

4. Resolve each corporate shareholder

For every legal-entity owner that remains relevant, repeat the identity and ownership steps. Match it by identifier and jurisdiction, not name alone. Preserve a stable ID for every node so the same company appearing through multiple branches can be recognised instead of counted twice.

5. Calculate indirect interests along each path

Multiply the ownership percentages along one path to calculate the indirect interest contributed by that path. Where the same person reaches the subject through more than one independent path, aggregate the supported interests according to the applicable method. Keep calculation inputs visible so a reviewer can reproduce the result.

6. Test control independently

Review voting rights, appointment powers, partnership control, declared persons with significant control and other evidence relevant to the policy. A person may qualify through control even when a calculated ownership interest falls below a percentage test.

7. Record the conclusion and its limits

State which people meet which tests, the ownership paths used, the sources consulted and any unresolved branch. If nobody can be identified from the available evidence, report that as an incomplete or escalated outcome—not as proof that the company has no UBO.

How to calculate indirect beneficial ownership

Indirect ownership is calculated through the chain rather than copied from the nearest shareholder record. For a single path, multiply the interest at each layer. If Person A owns 60% of Holding Company B, and Holding Company B owns 80% of Operating Company C, Person A’s indirect interest in C through that path is 48%.

The calculation is:

This example is deliberately simple. Real structures introduce additional issues:

• the same person may hold a direct interest and one or more indirect interests;
• different share classes may carry different voting or economic rights;
• circular ownership can cause double counting if entities are not deduplicated;
• percentages may be missing, expressed as ranges or reported for different dates;
• a corporate owner may be controlled by a person without that person holding the largest economic interest.

Do not force an exact percentage when the evidence does not support one. A transparent range or an “unable to calculate” status is more defensible than false precision. Preserve the original source values and the normalisation applied before any multiplication.

Ownership percentage is not the same as control

A percentage is attractive because it is easy to compare with a threshold. Control is harder because it can arise from rights, agreements and practical influence. A robust UBO review treats ownership and control as separate tests that can lead to the same conclusion.

Document which test each person satisfies. Do not label every person with a large shareholding as a UBO without applying the policy, and do not ignore a controlling person merely because their economic interest is small or unavailable.

Declared beneficial-owner or person-with-significant-control records can accelerate the review, but they should be treated as sourced evidence rather than an unquestionable answer. Check the declaration date, source definition, verification status and consistency with the underlying ownership graph.

Build an auditable UBO evidence record

The result should let a second reviewer reproduce the conclusion without starting from zero. For every entity and relationship, capture the original source value, a normalised value where needed, the effective or filing date, the retrieval time and the transformation applied.

The evidence package should distinguish four states:

Verified

The relationship or fact is supported by an accepted source and passes the applicable checks.

Declared

The value is reported by the company, applicant or registry but has not been independently reconciled.

Calculated

The value is derived from visible inputs, with the formula and paths preserved.

Unresolved

A material branch, percentage, identity or control question remains incomplete or conflicting.

Never translate “not available” into zero ownership. Never translate a failed source request into “no change.” Keep source failure, field absence and negative findings as different states. This is essential when results are consumed by automated workflows.

Source hierarchy should follow the purpose of the review. Official registry records and filings often provide the strongest foundation for legal identity and reported relationships. Company-supplied documents can fill gaps or show what was represented. Curated datasets can connect records across jurisdictions. Each has value, but the final case should make provenance visible rather than flattening all inputs into one unexplained answer.

CompanyDelta’s sources and evidence model preserves source context alongside company facts. Teams can use the API to retrieve structured records and maintain the original relationship path inside their own case-management environment.

A UBO conclusion expires when the ownership facts change

Beneficial ownership is not a one-time attribute. Shares can transfer, a holding company can be inserted, voting rights can change, a director can gain new appointment powers or a company can move within a group. The approval remains defensible only while the underlying facts remain materially consistent.

Build monitoring around the verified legal entities in the ownership graph, not around names alone. Relevant events include:

• shareholder additions, removals and percentage changes;
• new persons with significant control or beneficial-owner declarations;
• changes to parent companies and group relationships;
• new share classes, voting arrangements or governance filings;
• status changes affecting an entity in the ownership chain;
• new filings that resolve or contradict an earlier evidence gap.

Not every event requires a full investigation. Recalculate only the affected paths first, then determine whether the change crosses a policy threshold, introduces a new controlling person or invalidates an exemption. Route the event to the right reviewer with the previous value, new value, source and affected UBO conclusion.

Use webhooks to deliver relevant changes into the review queue. The broader supplier monitoring framework explains how to prioritise events, assign ownership and prevent low-value alerts from overwhelming analysts.

Implementation checklist for ownership teams

1. Define the policy before the data workflow

   Document the rule set, ownership thresholds, control tests, exemptions, stop conditions and evidence requirements for each use case and jurisdiction.

2. Use stable entity identifiers

   Link every node to registration number and jurisdiction wherever possible. Keep source identifiers and internal IDs so duplicate companies can be reconciled across branches.

3. Preserve relationship-level provenance

   Store the source and date for every edge, not only each entity. Ownership graphs become unauditable when relationships lose their evidence.

4. Separate declared, verified and calculated fields

   Make it clear whether a percentage came from a filing, an applicant document or multiplication across a path.

5. Expose incomplete coverage

   Show restricted sources, missing percentages and unresolved corporate shareholders. Do not let automation convert uncertainty into a clean result.

6. Monitor the whole relevant graph

   Watch the subject and material ownership nodes. Recalculate affected paths when relationships change and preserve the previous conclusion for comparison.

For investigations that require cross-jurisdiction tracing, review the Ownership & Control Agent, the practical guide to finding company ownership and CompanyDelta’s current coverage.