ALLOCATOR SERIES FOR FUNDS OF FUNDS · No. 2 of 5
Portfolio Construction for Early-Stage Fintech
This is the second installment of our series dedicated to fund of funds. You can find the first article here.
Ownership, reserves and follow-on logic: the engine an allocator is actually buying when it commits to an early-stage fund.
Pascal Bouvier, MiddleGame Ventures · August 2026
When a fund of funds underwrites an early-stage manager, it is not buying the manager’s stories. It is buying an engine: how many positions, entered at what ownership, defended with what reserves, followed on under what rules. Two funds with identical logos on the portfolio page and different construction will produce different outcomes — reliably, and by a wide margin. Deal anecdotes are the paint job. This piece is about the engine.
Start with the arithmetic nobody escapes. Venture returns follow a power law: in most funds, one or two outcomes carry the entire vehicle, and the median investment rounds to zero. The consequence is brutally clarifying. A fund that owns 10% of a company at exit needs a €1 billion outcome to return €100 million of committed capital once — before fees. Halve the ownership and you double the outcome required. Every construction decision — entry price, cheque size, reserves, pacing — is downstream of that equation, which is why an allocator should insist on seeing the equation before the anecdotes.
Ownership is the product
Most institutional-grade seed funds target roughly 7–15% ownership at entry, and the discipline shows up not in the target but in the delivered median. Ownership per euro invested is the cleanest efficiency measure an emerging fund can show: what a manager paid, and what stake that bought. Europe currently sharpens this materially. Median Series A pre-money valuations run near $28 million in Europe against $48 million in the United States — a 30–50% structural discount that persists at every stage — so the same cheque simply buys more company. For a fintech specialist, the discount compounds with knowledge: knowing which regulated niches are mispriced is ownership bought with expertise rather than capital.
Reserves are a strategy, not a savings account
Industry norms put follow-on reserves anywhere between 20% and 40% of an early-stage fund, with some managers running ratios of two reserved euros for every euro initially deployed. The number matters less than the logic behind it. Reserves exist for exactly one purpose: to buy more of the positions where the evidence has improved — to defend and extend ownership in the companies external markets are validating. The characteristic failure mode is the opposite: reserves quietly become survival capital, bridging the middle of the book because the manager cannot bear to recognize a loss. A manager with no reserve policy is leaving its best outcomes on the table; a manager whose reserves flow to its weakest names has misunderstood its own business. The follow-on ledger — which companies received a second cheque, and what evidence triggered it — is, as we argued in the first piece of this series, the most candid document a young fund possesses.
How many positions
Portfolio breadth is the third leg. Too concentrated, and power-law variance can sink a well-selected fund on bad luck alone; too diffuse, and neither capital nor partner attention is meaningful in any single company. At seed, portfolios of roughly 20–30 core positions have become the institutional consensus range, sized so that graduation rates — the share of companies that reach a competitive Series A and beyond — translate into a handful of genuine fund-return candidates. At A, portfolios of roughly 15 to 20 core positions have also become the norm. Recycling policy belongs in this conversation too: a manager who recycles early proceeds into new and follow-on positions is buying its LPs more invested capital per committed euro, and should be able to say precisely how much.
What construction looks like in a specialist fintech book
Construction is not generic; it should embody the thesis. Ours is European fintech from seed through Series B, with first cheques between €500K and €5 million, and the construction logic follows from how regulated businesses actually create value. Fintech companies pass discrete, underwritable inflection points that generalist SaaS does not have: licenses granted, regulatory perimeters entered, bank partnerships signed, payment volumes crossing thresholds where unit economics invert. Reserve depth and follow-on triggers can be calibrated to those milestones rather than to sentiment. This is also why fintech construction rewards the specialist: judging whether a company will clear a regulatory milestone is ground truth, not pattern-matching — the kind of judgment that, as we argued last time, persists across funds.
The questions
For an allocator, five questions surface most when analyzing the venture “engine”:
- What is your ownership target at entry — and what is your delivered median across the current portfolio?
- What is your reserve ratio, and what written rule governs a follow-on cheque?
- Walk me through your last three follow-ons: what evidence triggered each?
- What is your recycling policy, and what invested-to-committed ratio does it produce?
- What would break your model — and what did you change in it after 2021?
A manager who answers these fluently has an engine. A manager who answers with portfolio company stories has a scrapbook. Both can be charming over dinner; only one is underwritable.
This is No. 2 in a five-part series for allocators. Next: emerging versus established managers — and where the European edge sits.
Sources
Kauffman Fellows Journal, “Venture Fund Portfolio Construction” — ownership and reserve norms (kauffmanfellows.org).
Industry portfolio-construction guides on seed ownership targets (7–15%) and reserve ratios (20–40%; 2:1), incl. The VC Factory and Signature Block.
Value Add VC, “European VC Valuation Discount 2026” — median Series A pre-money ≈$28M Europe vs ≈$48M US (valueaddvc.com).
MiddleGame Ventures — stated stage focus and first-cheque range (middlegamevc.com).