What a family office directory can and cannot tell you

A private equity fund has to tell you things. It raises from institutions, so it files, it reports, and it publishes a team page because that is how it wins mandates. A single-family office has no such obligation and usually no such incentive. It exists to manage one family’s capital, it is not raising, and in a good number of cases it would prefer you did not know it exists at all.

That asymmetry is the whole problem with family office data, and it is worth being precise about where the line falls before you spend money on any list, ours included.

What is realistically knowable

That the office exists, and its legal form. Company registries are the floor. Hong Kong’s Companies Registry and Singapore’s ACRA both publish incorporation records, and a family office structured as a local entity appears in them. This is reliable but thin. It establishes a name and a jurisdiction and very little else.

Whether it is single-family or multi-family. A multi-family office serves several unrelated families and therefore has to market itself. It has a website, a services page, and often a regulatory licence. A single-family office may have none of these. The distinction is usually determinable and it matters more than almost anything else on the record, because it predicts whether the office will respond to an unsolicited approach at all.

Whether it invests directly. Some family offices allocate to funds and managers. Others buy companies and assets themselves. Offices that invest directly tend to say so, because they want deal flow. If you are looking for a buyer rather than an LP, this is the field that matters.

A contact route, sometimes. A general office email, a switchboard number, a website contact form. Roughly three-quarters of the offices in our directory have one of these. The rest have a working website and named principals, and nothing more.

Declared appetite. Where an office publishes what it invests in, whether private equity, venture, real estate or operating businesses, that is a statement of intent from the office itself and can be taken at face value as such.

What is not knowable, and is routinely faked

Assets under management. This is the single most over-claimed field in family office data. Single-family offices almost never publish AUM. There is no filing that compels them to. When a directory shows you an AUM figure for hundreds of single-family offices, it is either derived from a wealth estimate of the underlying family, itself an estimate and usually of illiquid holdings, or it is a band invented to make the row look complete. We hold a disclosed AUM figure for a very small number of offices and we show a blank for the rest, because a blank is honest and a guess is not.

Who makes the decisions. You can often learn who runs the office. You can rarely learn who has to approve a transaction, which family members hold sway, or whether the person listed as CIO is empowered or ceremonial. No dataset will give you this. It comes from the conversation.

Whether they are currently active. A record can tell you an office existed and had a stated appetite on the date it was checked. It cannot tell you the family sold the operating business last quarter and stopped deploying. Any list, however well built, is a snapshot.

Personal contact details. We do not hold direct mobile numbers or private email addresses for family principals. That information is not public, and a vendor offering it has either scraped it or bought it from someone who did.

The registry trap

There is a specific failure mode worth naming, because it is how most large family office lists get large.

Both Hong Kong and Singapore make it easy to enumerate companies whose name contains “family office”, or whose registered industry code corresponds to family office activity. Run that query and you get thousands of rows in an afternoon. Almost none of them are usable: they are holding vehicles, dormant shells, and administrative entities that happen to carry the words in their legal name. There is no website behind them, no contact, and often no operating office.

We hold several thousand such rows internally because they are a useful starting point for research. None of them are in the product. A row only becomes something we will sell once it has the office’s own website, a direct contact route, or named principals behind it. The majority of registry-sourced rows never do.

This is why a smaller list can be worth more than a larger one, and why the first question to ask any vendor is not “how many records” but “how many of them have something I can act on”.

How to read a record

Treat every field as one of three kinds:

1. Structural. Name, jurisdiction, legal form, single- or multi-family. Stable, verifiable, and safe to rely on.
2. Declared. Stated investment focus, described strategy, named principals. Accurate as a statement of what the office says about itself, on the date it was checked.
3. Absent. Blank fields. Read these as “not published”, never as “none”. An office with no recorded private equity focus may well do private equity. It has simply never said so anywhere we could find.

The mistake that wastes the most time is treating a declared field as structural, or an absent field as a negative. A directory is a starting list of who to approach and how to reach them. The qualification happens after that, in the conversation, the way it always has.

Sources

1. Companies Registry — Government of the Hong Kong Special Administrative Region; accessed September 4, 2026
2. Accounting and Corporate Regulatory Authority — Government of Singapore; accessed September 4, 2026