It is a pattern that repeats. A family office invests in a new reporting platform, or begins a risk system implementation, and within six months the project is struggling. Not because the technology is wrong, but because the data feeding it is inconsistent, unclean, or ungoverned. Two systems hold different valuations for the same position. Three people maintain three versions of the same entity record. No one is certain which is correct.

Data governance is the discipline that prevents this. It is not primarily a technology problem. It is an ownership problem — and solving it requires decisions, not just tools.

What data governance actually means

In a family office context, data governance means establishing clear answers to three questions for every material data type the office works with:

  • Who owns this data — meaning who is accountable for its accuracy and completeness?
  • Where is the master record — meaning which system or source is authoritative when there is a conflict?
  • How does it flow — meaning what is the documented path from source to consumption, and where are the transformation and reconciliation points?

Most family offices can answer these questions for some of their data. Very few can answer them for all of it. The gaps are where the problems live.

The four data domains that matter most

1. Security master and reference data

Reference data — the attributes that define what a security is, how it is priced, how it is classified — is foundational to everything else. If the same instrument is classified differently in your portfolio system and your risk system, your risk figures are wrong. If pricing sources conflict and there is no hierarchy to resolve them, your valuations are unreliable. A governed security master establishes a single source of truth for instrument attributes, a defined pricing hierarchy, and a clear process for adding new instruments and resolving conflicts.

2. Entity and counterparty data

Family offices typically hold assets through multiple legal entities, trust structures, and holding vehicles. The relationships between those entities — and their mapping to counterparties, custodians, and investment vehicles — need to be maintained consistently across every system that uses them. When they are not, consolidated reporting becomes unreliable, and any attempt to produce a consolidated balance sheet or risk view becomes a manual reconciliation exercise.

3. Position and transaction data

Position data is the most operationally critical data in a family office. It needs to be timely, complete, and reconciled. The questions that matter here are: how frequently is position data reconciled against custodian records? Who is responsible for break resolution? What is the tolerance for unreconciled positions before escalation? The absence of documented answers to these questions does not mean the work is not happening — it usually means it is happening informally, which means it depends on individuals rather than process.

4. Performance and valuation data

Performance data is downstream of position data, pricing data, and transaction data. Errors in any of those flows propagate into performance figures. A governance framework for performance data needs to establish the calculation methodology, the reconciliation process, and — critically — the sign-off process before figures are shared with the Principal or investment committee. Performance figures that are shared without a documented review process create reputational risk if they are later found to be incorrect.

Where to start

The most useful starting point for a family office without a formal data governance framework is a data inventory: a structured record of what data the office holds, where it lives, who owns it, and what systems consume it. This exercise typically takes two to three weeks and produces a clear picture of where the gaps and risks are. It also surfaces the informal dependencies — the spreadsheet that feeds the board report, the manual extract that reconciles the custodian feed — that are invisible until they break.

From the inventory, prioritisation becomes straightforward. The data domains that feed the most critical outputs — investment committee reporting, risk monitoring, client reporting — are addressed first. The framework is built incrementally, starting with the highest-risk areas rather than attempting to govern everything at once.

The goal is not a perfect data governance framework on day one. It is a documented, owned, and operational framework that improves over time — and that does not fail silently when something goes wrong.