Root cause: three systems of record, three different questions
The fragmentation is a consequence of system design rather than poor data entry. Each system in a typical private-markets stack was built to answer a different question.
- The fund-administration register identifies the legal entity that signs the subscription and holds the capital account.
- Know-your-customer and anti-money-laundering onboarding identifies the natural person associated with a given vehicle, as required by regulation.
- The customer relationship management system records a contact for investor-relations purposes.
Business impact on the capital-raising function
On the servicing side, duplicate investor records are largely an operational inconvenience. On the capital side, they can affect placed capital and reporting accuracy in four principal ways.
- First, re-up and outreach prioritisation is distorted. Senior outreach ahead of a fundraise is typically ranked by committed capital, and a principal recorded as several mid-sized investors may not reach the priority threshold, reducing the likelihood of re-commitment from the firm’s larger relationships.
- Second, co-investment offers are sized to a single visible vehicle rather than to the investor’s full capacity, which can lead to under-allocation to significant backers.
- Third, single-investor concentration is measured inaccurately, which affects risk limits, Limited Partner Advisory Committee reporting and side-letter compliance.
- Fourth, the firm incurs duplicate know-your-customer and servicing costs for the same individual, with a corresponding increase in the risk of inconsistent records.
The materiality of these effects varies by firm. Managers whose investors each commit through a single vehicle may see limited benefit from further resolution, and existing strong-key matching may be sufficient. The impact increases with the proportion of the investor base that commits through multiple vehicles, which tends to be concentrated among the firm’s larger relationships.
Why joining on the beneficial-owner key does not resolve investors
A common proposed remedy is to join investor records on the beneficial-owner identifier captured during onboarding, on the assumption that a single such key exists for each principal. In most cases it does not.
The same principal is typically onboarded under a different strong document for each vehicle: for example, a passport for the trust, a national identity document or driving licence for the co-investment entity, and a Social Security or tax number for the direct holding. Each document is a valid and verified identifier, but the identifiers are not equivalent. A join across these fields therefore matches no records, and each vehicle remains isolated. The strong key reliably consolidates only those vehicles that were onboarded under the same document, which tends to be the less material case; the multi-vehicle principals that carry the most value are precisely those recorded under different documents.
Resolution approach: name and date of birth, with LEI for institutions
Effective resolution uses three methods within a single multi-signal framework, selected according to the record type.
For individual principals, the primary bridge is a match on name combined with an exact date of birth. Name alone is insufficient, as it can incorrectly merge relatives who share a name; the exact date-of-birth condition materially reduces that risk while allowing the differently-documented vehicles to be linked. Where two vehicles were onboarded under the same document, the shared strong key continues to consolidate them directly. For institutional limited partners, such as pension funds, endowments and funds-of-funds, there is no natural person behind the entity, and resolution is performed on the Legal Entity Identifier (LEI) at the entity level. Used together, these methods address the same-document, different-document and institutional cases respectively.
Governance: precision testing and use-based permissions
In a private-markets context, an incorrect merge carries greater consequence than in consumer marketing. Merging two investors in error can expose one investor’s capital account, holdings and commitment history to another, which may constitute a personal-data breach under GDPR and, in the United States, an issue under Reg S-P, as well as a potential information-barrier breach where the investors are separated by such a barrier.
Two controls are therefore recommended. First, each probabilistic rule, including the name-and-date-of-birth bridge, should be precision-tested against a held-out set of records with known outcomes before it is used, and rules that do not meet the required precision should not be deployed. Second, a use-based permission model should govern what each resolved link is allowed to do. High-confidence links equivalent to a strong key may be used for reportable purposes such as concentration and committee reporting; lower-confidence probabilistic links may support an internal consolidated view but should be excluded from reportable or externally disclosed outputs until confirmed by a stronger signal. Under this model, coverage can be extended for analytical purposes without a corresponding increase in disclosure risk.
Building the business case: a single firm-wide measure
The opportunity is best quantified as a single firm-wide measure rather than as separate metrics per source system. The recommended measure is the proportion of the investor base, weighted by assets under management, that resolves to a single underlying person or entity once the bridges are applied.
The unresolved remainder should then be segmented by cause: records with no shared document, institutional entities without a captured LEI, and genuinely distinct investors. The resolvable but currently fragmented portion can be translated into financial terms the firm already tracks, principally the assets under management that are mis-ranked at re-up and the duplicated know-your-customer and servicing costs associated with repeated individuals. Reporting a single firm-wide figure, rather than a resolved rate per source, keeps the assessment focused on the business outcome rather than on the relative performance of individual systems.
Scope: batch resolution (L2), not real-time (L3)
The capability described here is batch identity resolution, corresponding to Level 2 in the staged model of identity resolution. It runs on a scheduled basis against the systems the firm already operates, including the fund-administration register, the know-your-customer store and the customer relationship management system, and does not require streaming infrastructure or platform replacement.
Resolving an investor for the next fundraise, co-investment offer or concentration report is inherently a periodic task and does not require real-time processing. Extending resolution to real time corresponds to Level 3, which is a larger and separate undertaking and is generally warranted only after the batch capability is established. Applying real-time processing to an unresolved base tends to accelerate decisions made on an incorrect identity. Scoping the initiative to Level 2 limits implementation risk.
Conclusion and recommended actions
On the capital side of private equity, the firm’s most significant relationships are frequently the least visible, because value is concentrated among principals who commit through multiple vehicles. The commonly proposed beneficial-owner join resolves only the same-document cases and leaves the more material different-document cases unaddressed. Effective resolution combines a name-and-date-of-birth bridge for individuals, the strong key for same-document vehicles and the Legal Entity Identifier for institutions, with each probabilistic rule precision-tested and governed by use-based permissions, and with the overall opportunity quantified as a single firm-wide measure delivered in batch.
As an initial step, firms can assess the scale of the issue directly: for the largest relationships by committed capital, determine how many distinct records each principal spans across the register, the know-your-customer store and the customer relationship management system. Where the largest relationships resolve to a single record, the issue is limited. Where they span multiple vehicles that current ranking does not consolidate, the assessment identifies both the affected relationships and the approximate scale of the opportunity.
Next step
Firms evaluating this capability can use the identity resolution playbook, a fixed-scope batch engagement that resolves principals across the existing register, know-your-customer store and customer relationship management system, with precision and permission controls applied. A short scoping discussion can establish the likely size of the gap and its cost ahead of any commitment.
Identity Resolution for Private Equity — FAQ
Why does joining on the beneficial-owner key not resolve LPs across their vehicles?
The same principal is typically onboarded under a different strong document for each vehicle, such as a passport for a trust, a driving licence for an LLC and a tax number for a direct holding. These identifiers are each valid but are not equivalent, so a join across them matches no records. The beneficial-owner key reliably consolidates only vehicles onboarded under the same document.
What method links one investor across a trust, an SPV and a feeder fund?
For individuals, the primary method is a match on name combined with an exact date of birth, which links vehicles that strong keys cannot while reducing the risk of merging relatives who share a name. Vehicles onboarded under the same document continue to be consolidated by that shared key. Institutional limited partners are resolved on the Legal Entity Identifier at the entity level.
Does investor identity resolution create compliance risk under GDPR or Reg S-P?
It can if a probabilistic match is treated as certain, because an incorrect merge may expose one investor's account to another and, where an information barrier applies, breach it. The recommended controls are to precision-test each rule against known outcomes before use and to apply a use-based permission model that limits what lower-confidence links may drive. Applied in this way, the approach supports resolution without a corresponding increase in disclosure risk.
How does batch identity resolution differ from a real-time customer data platform project?
Batch identity resolution corresponds to Level 2 and runs on a scheduled basis against the register, know-your-customer store and customer relationship management system already in use, without streaming infrastructure or platform replacement. Resolving investors for fundraising and reporting is a periodic task that does not require real-time processing. Real-time resolution corresponds to Level 3 and is generally warranted only after the batch capability is established.
How can a firm size the LP identity gap before committing to a project?
The recommended measure is a single firm-wide figure: the proportion of assets under management that resolves to a single underlying person or entity once the bridges are applied, with the unresolved remainder segmented by cause. The resolvable portion can be expressed in financial terms already tracked by the firm, principally mis-ranked assets at re-up and duplicated know-your-customer and servicing costs. A single firm-wide figure is more useful for this purpose than a resolved rate per source system.
