What a family office actually costs
Four structures, the running cost in basis points, and the asset level below which the arithmetic simply does not work.
A family office is not a status object, it is a small company with a payroll, and the payroll does not care how much you like the idea. Industry practice puts running costs somewhere between 30 and 120 basis points of assets, personnel accounts for two thirds of that, and below roughly 250 million the arithmetic stops working no matter how it is arranged.
What the thing actually is
Strip the mystique and a family office is an entity that manages one family's money and affairs: investments, tax, structures, property, sometimes the operating businesses, sometimes the philanthropy, and in the fuller versions the education of the next generation. It exists because past a certain scale those things stop being manageable as a set of relationships with outside advisors and start needing somebody whose job it is.
There are four shapes it takes, and the differences are practical rather than definitional.
| Single | Multi | |
|---|---|---|
| Who it serves | One family, exclusively | Several unrelated families, sharing the infrastructure |
| Control | Complete. You hire, you set the mandate, you decide | Shared. The firm sets its own strategy and you are a client of it |
| Cost shape | Fixed. A payroll that exists whether or not much happens | Variable. A fee, usually on assets, that scales with you |
| Where it makes sense | Large enough that the fixed cost is a small percentage | Almost everywhere below that, and plenty of places above it |
| The real difference | You own the capability and the loyalty of the people | You rent both, which is cheaper and occasionally less aligned |
Two more variants matter and are less discussed. An embedded family office sits inside an operating business the family already owns, using its finance function and its back office to handle family matters, which is how a very large share of them actually begin. A virtual one is a small coordinator, sometimes one person, who holds the relationships and buys everything else in from outside firms.
The cost, in numbers
The figures below come from industry practice rather than academic research, so treat them as the shape of the thing rather than as measurements. They are consistent enough across sources to be useful.
Running costs typically fall between 30 and 120 basis points of assets under management, with the wide spread explained almost entirely by scale. A smaller office might run on a budget of one to one and a half million a year at 50 to 60 basis points. Averages across the sector land near 40 basis points and roughly 3.2 million a year, rising to something like 6.6 million for offices above a billion.
Personnel is 60 to 70 percent of that and is the only line that really matters. For an office in the 300 million to one billion range, a chief investment officer's base salary sits in the region of half a million dollars, with total compensation typically 40 to 60 percent above base once bonus and co-investment are counted. That is one person.
The comparison that decides it
A family with $150m running a deliberately lean office at $1.2m a year is paying 0.80% of assets in overhead, before any investment costs at all. An outside advisory firm covering the same family might charge 0.30% to 0.50% all in. The office is costing 30 to 50 basis points more, every year, forever.
That premium is not automatically wrong. It is the price of control, and control is worth something. But it should be a decision somebody made with the number in front of them, rather than a structure that grew because it seemed like what families at this level do.
Thirty basis points a year on a hundred and fifty million is four hundred and fifty thousand. Ask what that buys and the answer had better not be a nicer letterhead.
The threshold, and why it moved
The number usually quoted as a minimum is 100 million in investable assets. Practitioners increasingly say the real figure is 250 million, on the basis that this is where a proper office can be run for under one percent of assets, and that the point where it genuinely makes sense is somewhere between 250 and 500 million depending on jurisdiction and how much the office is asked to do.
The mechanism behind every one of those numbers is that the cost is a payroll and the payroll is fixed. Hold the lean budget above steady and watch what the denominator does to it:
| Assets under management | A $1.2m budget, as a share of them |
|---|---|
| $50m | 2.40% |
| $100m | 1.20% |
| $150m | 0.80% |
| $250m | 0.48% |
| $500m | 0.24% |
| $1bn | 0.12% |
Nothing in that column is about skill, strategy or market conditions. It is one number divided by another, and it is why the threshold exists at all. The same office, staffed identically and performing identically, is a 2.4 percent drag at fifty million and a rounding error at a billion.
One honest correction to that table, because it flatters the top rows. A billion-dollar office does not run on 1.2 million; it hires more people, which is exactly why sector averages climb toward 6.6 million at that size. So the real curve is flatter than this one. But it still falls, because headcount grows more slowly than assets, and that gap is the entire economic case for the structure.
The useful line to take from it is the crossing point. At a genuinely lean 1.2 million, the one percent test the industry uses is met at about 120 million. Anything below that and you are paying more than one percent before a single investment decision has been made, which is a hurdle the portfolio then has to clear just to break even against the simpler alternative.
The threshold has drifted upward for unglamorous reasons. Compensation for good investment people has risen faster than most things. Compliance and reporting obligations have expanded. And the outside alternative got considerably better, because the advisory market that a family office competes against is more capable and cheaper than it was twenty years ago.
For scale, the UBS survey of 307 family offices across more than thirty markets in early 2026 found an average family net worth of 2.7 billion dollars and an average of 1.3 billion managed per office. That is the population most published material describes, and it is a long way above where most families asking the question actually sit.
What the premium genuinely buys
Four things, and knowing which of them you actually want matters, because you can buy them separately.
Undivided attention
Nobody in the office has another client. That matters most when something goes wrong at an awkward time, and it matters much less during the years when nothing does.
The ability to do direct deals
The main reason most of these get built now. Buying whole companies takes people who can assess and then oversee them, and no advisory relationship provides that. This is the capability that actually justifies the structure.
Coordination across everything
Investments, tax, structures, property and the operating business considered as one position rather than by five firms who never speak. The gains here are quietly large and almost never quantified.
Privacy and control of information
Fewer people know the whole picture. Worth a great deal to some families and almost nothing to others, and worth being honest about which one you are.
Direct deals are the interesting entry, because that is the only capability an outside firm genuinely cannot supply. An advisor can allocate your money, run your tax and hold your structures. What they cannot do is sit on the board of a company you bought, work out why its margins slipped, and decide whether to replace the managing director. If you intend to own operating businesses, you need that capability in-house, and that is a hiring decision rather than an assets-under-management decision.
The cheaper structure most families should start with
If the direct-ownership answer is yes but the assets are well short of 250 million, the sensible path is not a scaled-down single family office. It is the embedded version.
Run the family's affairs out of the operating company that made the money in the first place. The finance function already exists, the controller already produces accounts, and the marginal cost of having them handle family structures alongside is a fraction of a standalone office. Add one genuinely good generalist rather than a full team, and buy investment management, tax and legal from outside on a fee basis.
Two rules make this work rather than turn into a mess. Keep the family's affairs and the company's books cleanly separate, in the accounting and in the governance, because blurring them creates tax and partner problems that are expensive to unwind later. And write down what the arrangement is, so that when the operating company is eventually sold the family capability does not leave with it. What that written governance should contain, and how few families have any, is worth its own treatment.
The test I would apply
Five questions, in order, and the first two are disqualifying.
- Are you going to own and oversee operating businesses directly, or only hold financial assets
- Does the total annual cost, honestly estimated with real salaries, come to under one percent of what it manages
- Is there a specific person you would hire, by name or by profile, and would they be fully occupied
- Would you still want this structure if the family's assets fell by a third
- Is anybody in the next generation actually going to be involved, or is this being built for an audience of one
If the answer to the first is only financial assets, an outside firm will do it better and cheaper, and the honest reason to build an office anyway is preference rather than economics. That is a legitimate reason and it should be named as such rather than dressed up in a cost argument that does not hold.
The direction of travel across the sector is toward direct ownership, which is why this question is coming up more, and why the answer is increasingly yes for families who would have said no ten years ago. The case for holding companies rather than fund positions is in private equity is becoming a family business, and the discipline that has to sit on top of it, once you are the one deciding where the money goes, is in where the cash goes after a good year. If you are weighing a first direct acquisition, what I screen for is the same list a family office committee should be running.
The short version
- Running costs sit roughly between 30 and 120 basis points of assets, averaging near 40, with personnel at 60 to 70 percent of the total.
- A lean office on 150 million at 1.2 million a year is 0.80 percent of assets in overhead alone, against 0.30 to 0.50 percent for an outside firm covering the same ground.
- The commonly cited minimum is 100 million, but practitioners increasingly put the working threshold at 250 million and the real inflection between 250 and 500.
- The whole threshold is one number divided by another. A lean 1.2 million budget is 2.40 percent of 50 million and 0.12 percent of a billion, and it meets the one percent test at about 120 million.
- The one capability you cannot rent is operating oversight of companies you own. That, not asset level, is the actual trigger.
- Below the threshold, the answer is an embedded office inside the operating company, not a miniature version of a full one.
Questions I get on this
How much money do you need for a family office?
What does it cost to run a single family office?
Single family office or multi family office?
Survey scale, average net worth and assets per office from the UBS Global Family Office Report 2026 (307 family offices across more than 30 markets, surveyed 22 January to 30 March 2026). Cost ratios, staffing splits and threshold figures reflect published industry practice rather than peer-reviewed research and vary considerably by jurisdiction and scope.