FOUNDATION SERIES FOR FAMILY OFFICES · No. 2 of 5
Funds, Direct, or Both?
How family offices really get paid in venture — and why the instinct to go direct is right in spirit and wrong in structure.
Pascal Bouvier, MiddleGame Ventures · July 2026
This the second article directed at family offices. The first article can be found here.
Families do not usually arrive at venture through a fund. They arrive through a company. Someone built wealth by owning things directly — operating them, sitting on the board, backing people they knew — and that instinct, to own the asset rather than rent a manager, is the very thing that built the fortune in the first place. So when a family office turns to venture, the reflex is natural: why pay someone two and twenty to do what we could do ourselves? It is a fair question, and the honest answer is not that direct investing is a mistake. It is that venture punishes the direct investor in ways that stay invisible until they have already happened. The question worth asking is not whether to be in venture — you should be — but how to get the exposure without walking into the traps the asset class sets for the untrained.
Why venture belongs in the portfolio at all
Start with the case for the asset class, because it has changed in the family office’s lifetime. The public market is shrinking: the number of listed U.S. companies has fallen from roughly 8,000 in the late 1990s to about 3,700 today, while companies stay private far longer and raise enormous sums without ever ringing the bell.[1] The practical consequence is stark. A great deal of the value that once accrued to public shareholders — the compounding between a company’s early growth and its maturity — now accrues in private hands, before any listing. A portfolio built only of public equities and bonds is, by construction, absent from a growing share of where value is actually created. Venture is the entry point to that private compounding.
Families have noticed. Direct deals now make up more than 40% of the typical family-office private-equity sleeve, and venture and growth allocations keep climbing.[2] But the appetite is uneven, and European principals in particular should sit with that. North American families are well ahead on venture: roughly 60% of all family-office venture capital flows to the United States, with Europe and Asia splitting much of the rest.[3] Outside North America — across Europe, the Middle East and Asia — family-office alternative books are still dominated by real estate and traditional, plain-vanilla private equity, with venture and growth a thin slice on top. That is a gap, not a virtue, and it needs to close. The families that built their fortunes by backing the new before it was obvious are, for the most part, the ones least exposed today to where the new is being built. And noticing an opportunity and capturing it well are different problems — the gap between them is where this piece lives. Appetite is not a strategy.
Why families reach for direct — the case, stated fairly
The pull toward direct investing is not naïve, and I will not pretend it is. Going direct offers control, transparency into a single asset, no second layer of fees, clean alignment with a founder you have chosen, and the chance to put a family’s own operating expertise to work. For the rare family with a genuine sourcing edge in a sector it knows cold, the staff to diligence and support companies, and the discipline to build a real portfolio rather than collect a few trophies, pure-direct can work. The trouble is that this describes very few family offices. Most have one or two of those three ingredients. Almost none have all three — and venture is unforgiving of the missing one.
The hidden costs of going pure-direct
Six costs sit beneath the surface of a direct-only program. None of them shows up on a term sheet; all of them show up in the returns.
Adverse selection. This is the deepest and least understood. The deals that reach a family office are not a random sample of the market, still less the top of it. The best companies are competed for by full-time investors with proprietary access, brand and speed. By the time an opportunity arrives through a friend, a banker or an inbound email, the right question is why it is still available. You are not seeing the market; you are seeing the slice that happened to find its way to you — and that slice is adversely selected against you.
The judgment is a craft, not a checklist. Suppose the best deals did reach you. Picking among them is a skill, and a narrow one. A venture investor spends a career learning to answer a short list of deceptively simple questions: is this exceptional talent — genuinely curious, resilient, truth-seeking founders who can recruit and execute; is the market large enough and structured to be disrupted — big enough to carry a company toward nine figures of revenue within five to seven years, and fragmented or poorly served enough that incumbents can actually be displaced; and is there real momentum behind the technology or the shift the company is riding, rather than a slide claiming there is. Those questions take years of repetition to answer well, and the humbling part is that even people who answer them for a living are wrong a great deal of the time. A family office cannot manufacture that judgment in a season, however sharp it is at everything else.
The power law is extreme, and it punishes the few-bet portfolio. Venture returns do not merely skew; they follow one of the most extreme distributions in finance. A tiny number of winners pay for everything, while most investments return less than the capital put in.[4] Capturing the rare winner is a numbers game: the evidence on portfolio construction suggests you need on the order of twenty to thirty investments to reach the top performers with any reliability, and many funds deploy across fifty to eighty.[5] A family making five or ten direct bets is not running a venture portfolio; it is placing a concentrated wager against a distribution engineered to punish concentration — and the likeliest result is that it ends up owning several of the many losers and none of the few winners. That is not bad luck. It is arithmetic.
No reserves, no follow-on. A direct investor backs a company early and then, round after round, faces a choice: keep writing larger cheques to defend ownership, or watch the stake dilute through exactly the rounds in which the winners compound. Disciplined funds hold back a third or more of their capital precisely to follow their winners. The typical direct investor reserves nothing, and is quietly diluted out of the best outcome they ever had.
Time and operational load. Sourcing, diligence, legal negotiation, board seats, portfolio support, follow-on decisions, reporting — venture done well is a full-time operating business, not a sideline run between other commitments. Most family offices are not staffed for it, and the price of doing it badly is not a visible fee. It is the winner you never saw, or saw and mishandled.
The survivorship illusion. Families remember the one direct deal that worked and quietly forget the diffuse losses around it, which leads them to overestimate their own direct track record. Without portfolio-level bookkeeping, the flattering story beats the arithmetic — and a family that believes its own story is the one most likely to over-allocate to direct at precisely the wrong size.
Two different crafts: allocation and experimentation
Step back from the costs and the deeper point comes into view. A family office and a venture firm are good at two different things. The family office is, at its core, an allocator: its craft is deciding how capital should be spread across asset classes, managers and generations, and judging those choices over long horizons. A venture firm does something else entirely — it runs a continuous series of small experiments, most of which fail, in order to find the few that change everything. Selection at the level of a single company, conviction ahead of consensus, the willingness to be wrong often in pursuit of being spectacularly right occasionally: that is a distinct muscle, built over years of repetition. Neither craft is superior. But they are not the same, and the error families make is assuming that excellence at allocation transfers to excellence at experimentation. It does not — any more than a great pianist is, for that reason, a great composer.
The barbell: funds for access, direct for edge
The resolution is not to choose between the two. It is to use each instrument for what it does well. A fund buys you three things a family cannot easily manufacture: diversified, professionally constructed exposure to the asset class; access to a manager’s proprietary deal flow — the companies that will never reach your inbox; and the reserves and discipline to keep backing the winners. A direct sleeve, sized deliberately and kept smaller than instinct suggests, is where a family deploys the edge it genuinely has — the sector it understands, the founder it can uniquely help — without staking its entire venture allocation on its own blind spots.
Co-investment is the bridge between the two, and it is the single most valuable right a family office can hold. It delivers the direct exposure families want — a specific company, often with little or no fee or carry — but on a deal that a manager has sourced and diligenced after seeing the whole market. That is direct investing with the adverse-selection problem largely removed. It is also, not coincidentally, usually granted to the LPs who bring strategic value or a meaningful commitment, which is to say: the families who show up as serious fund investors first.
Think of it as a core and a satellite. The fund commitments are the core: they give you the asset class, built properly. The direct and co-investment capital is the satellite: it gives you the handful of companies you have real conviction in. A family that does only direct is over-exposed to what it cannot see. A family that does only funds forgoes the edge it actually has. The barbell uses both, on purpose.
Which is why the choice of partner matters more than almost any other decision a family makes in venture. The right firm hands you the thing you actually want — exposure to the winners, and the insight that comes from sitting inside the market — while absorbing the losses into a portfolio built to expect them. The write-offs still happen; they simply happen inside a structure designed to survive them, rather than landing one at a time on your own balance sheet as a string of disappointing single bets. You get the returns and the learning; you are spared the heartache. Choosing that partner well is hard enough to deserve its own treatment, which is where this series eventually goes.
What this means in practice
None of this is an argument against your instincts. It is an argument for structuring them. The family office is right to want to own venture, and right to want some of it directly. What it needs is the discipline to let funds do the work of building a portfolio and reaching the deals it will never see on its own, while reserving direct and co-investment capital for the places it holds a real advantage. Decide that structure first, and one question remains — the one that catches even sophisticated families: how much to commit, and at what pace. That is the subject of the next piece.
[1]The number of U.S. listed companies fell from roughly 8,000 in 1996 to about 3,700, even as companies stay private far longer (Center for Research in Security Prices data). See CNN Business, “America has lost half its public companies since the 1990s.”
[2]UBS Global Family Office Report 2025: direct deals now account for over 40% of the typical family-office private-equity sleeve, a sharp rise over a decade. FINTRX (Q1 2026) finds first-generation family offices increasingly favouring direct, private equity and venture over hedge funds and funds-of-funds. See UBS Global Family Office Report 2025.
[3]Roughly 60% of family-office venture capital flows to U.S. companies, with Europe and Asia splitting much of the remainder; outside North America, family-office alternative allocations remain weighted toward real estate and traditional private equity rather than venture and growth. See VC Lab, “Family Office Investment in Venture Capital”, and the UBS Global Family Office Report 2025.
[4]Correlation Ventures’ analysis of more than 21,000 U.S. financings (2004–2013), as widely reported: about 65% returned less than the capital invested, and only ~4% returned more than 10x. See The VC Factory, “The Super Power Law.”
[5]AngelList data and academic work on portfolio construction: reliably reaching the top performers takes on the order of 20–30 investments, and many funds deploy across 50–80; the majority of AngelList funds with at least 20 investments show a TVPI above 1x. See AngelList, “The Basics of VC Portfolio Construction.”
