Last post we said the structured database of a family’s financial life doesn’t exist as a product. It exists as labor.
This post is about who’s doing that labor. Because this is the industry’s dirty secret, and we think it’s time the industry said it out loud.
The cast of characters
Let’s talk about how this actually works, today, for most folks. When a UHNW family or a family office “has their data together,” what that actually means, behind the curtain, is some combination of these people:
Bucket one — the in-house family office analyst or accountant. Usually 28 to 45 years old. Often a CPA. Likely your wife’s sister or your fraternity brother’s kid that is going to have the most absurd resume at 22 you’ve ever seen. Comp in the $90K–$180K range depending on the city, per the latest Botoff Consulting family-office compensation data referenced by Fortune. Their job description says “operations.” Their actual job is pulling statements, reconciling positions, chasing K-1s, and rebuilding a spreadsheet every quarter that nobody else fully understands.
Bucket two — the external CPA firm. Billing somewhere between $300 and $700 an hour, depending on the partner and the market. They get the data the in-house analyst couldn’t get, run the tax work, and quietly maintain the second version of the family’s books. Often the only version that ties to the return.
Bucket three — the wealth owner themselves. We have watched principals with nine- and ten-figure balance sheets log into a GP portal at 11pm to download a capital-call notice because nobody else has the credentials. This is not rare. This is most weeks.
Bucket four — the outsourced fund administrator or family-office service provider. SEI, Eton, BBH, the big trust companies, the regional players. Excellent at what they do. Still, fundamentally, labor runs this industry at its core — people are in buildings doing work by hand for firms.
How much labor are we actually talking about
Deloitte’s 2024 family-office research found that North American family offices spend roughly 27% of their day-to-day time on administration and compliance — more than any other region (Deloitte). Personnel is 60–70% of family-office operating cost, according to the operational benchmarks Aleta has published (Aleta). A full-service single-family office can cost over $1M a year to run, per Fortune’s reporting on the J.P. Morgan Global Family Office Report.
Most of that cost is not portfolio managers. Most of that cost is the structured-database labor we described above.
And the family-office population is growing. Deloitte counts 8,030 single-family offices globally today, projected to reach 10,720 by 2030 — a 75% increase from 2019 (Deloitte). Every one of them needs this labor. There aren’t enough qualified humans, at any price, to scale linearly with that.
This is not to mention that you go ask 20 people what the threshold for a proper family office is, net worth wise, and you’ll get at least 15 different answers. The $10–$20–$30 millionaires want the bells and whistles of a family office — they can’t compete, at all, with the labor costs. There’s no way. (They can’t compete with the software costs either, but that’s another long, drawn-out blog.)
The K-1 problem, in one paragraph
A K-1 is a partnership tax document. For a UHNW family invested in, say, 40 private vehicles, K-1s arrive between February and October, in PDF, by email, by portal, by mailed paper, sometimes amended, almost never standardized. They have to be opened, classified, extracted, tied to the right entity in the family’s structure, and reconciled against capital-call and distribution history. Each one is a small, weird, idiosyncratic data problem. There are not enough hours in the calendar year for one analyst to do this well for one large family, never mind a multi-family office’s worth of relationships.
Canoe Intelligence, one of the better alternative-data shops, processed more than 150,000 documents on behalf of just one platform partner’s clients. That gives you a sense of the volume. And that’s one slice — funds — of the broader weird-asset problem, which also includes operating-company K-1s, real-estate LLCs, carry interests, intercompany loans, family loans, and the various trust structures that hold all of it.
Why all of this work combined has resisted productization
There’s a question we get a lot: if this is so painful, why hasn’t software eaten it already?
Three reasons.
One — the work is heterogeneous. No two GP statements look alike. No two family structures look alike. A vendor that wants to fully automate this either has to constrain the input (and lose the long tail of weird) or get really good at handling exceptions (which is expensive and not very VC-friendly until very recently).
Two — the buyers are private. Family offices and UHNW households don’t issue RFPs. They ask a friend. The category does not get the kind of public pressure that forced, say, payroll into the cloud.
Three — the buyers can afford the status quo. A family with $500M in assets paying $400K a year for in-house accounting is paying 8 basis points for the back office. That is not the budget line that keeps a principal up at night. The pain is real but it is not financial — it is operational, error-prone, and key-person-risky.
Why this is changing now
Two things have changed.
First — alternatives allocations among HNW and UHNW households have moved from a side dish to a main course. Bain projects private-market assets reaching $60–$65 trillion globally by 2032 (Bain & Company). UBS’s Global Family Office Report shows family offices allocating 42% of portfolios to alternatives (PR Newswire). Bank of America’s 2024 study found 93% of younger HNW investors plan to increase their alts allocation further (Bank of America). The volume of weird data per household is increasing faster than the supply of analysts to handle it.
Again, we’d add — much lower net worths view themselves as mini-family offices now. Even the $10 millionaires are heavily leaning into alts. They are even demanding them of their PWMs and RIAs.
Second — the actual cost of skilled financial-operations labor is going up. McKinsey’s wealth-management research projects a shortfall of 90,000–110,000 advisors by 2034, and the squeeze on adjacent operations talent is comparable (McKinsey). “Just hire another analyst” stopped being an answer about two years ago.
Not to mention every person you hire into your family’s closest, most intimate walls is another risk. Not even just of security — sure, that’s true — but of reputation and legacy and so many other elements.
The takeaway
The structured database of a family’s financial life exists today. It is built, maintained, and kept honest by humans. The cost runs into the hundreds of thousands of dollars per family per year, and the supply of those humans is not keeping up with the complexity of the assets they’re being asked to track.
That’s not a bug in the industry. That has been the industry. The question for the next decade is whether the database stops being labor and starts being product.
Spoiler — it does. We’ll get there in post six.