The standard advice on seed portfolio construction is a math problem, and the math is not wrong. Returns follow a power law. Most companies return nothing. A few return everything. You cannot know in advance which, so you take more shots.
We agree with that. We want shots on goal. A seed fund holding a handful of names is not a portfolio, it is a bet.
Where we part company with the standard advice is on what sets the number. In the US the answer is capital: write as many first checks as the fund can afford, and if you want more of them, write smaller ones. In Malaysia, Thailand and the Philippines, capital is not the binding constraint. Two other things are, and both of them cap the number well before the money runs out.
The first cap is attention, and it is not a soft factor
In a deep market, a seed company sits inside a dense supply of everything it needs. Operators who have scaled the same function before and can be hired. Design partners within driving distance. Lawyers, recruiters and agencies who have run the same playbook a hundred times. Above all, a thick layer of investors above you who will do the work at the next round.
A seed investor in that market can be almost purely capital and the company still works, because everything it needs is available somewhere else. Help is a differentiator at the margin. It shows up in the pitch. It is rarely load-bearing.
That supply is thinner here. The first enterprise customer in a second market. A manufacturing partner in Penang who will accept a small first order. A finance function that produces numbers a Series A investor can read without a translator. In our markets those frequently arrive through the investor, because no other supplier is standing by.
That changes what attention is. It stops being a marketing line and becomes an input to the outcome.
And attention does not scale with fund size. It is a person's calendar. Each additional company takes time out of the others, and it usually takes it from the ones that need it most, because the companies that go quiet are rarely the ones doing well.
So the honest version of the constraint is this. We would happily hold more names if we could be genuinely useful to more names. The cap is capacity. It is not a belief that owning fewer things is virtuous.
The second cap is the round that has to come after ours
This is the one most often missed by people applying US portfolio math to this region.
In a deep capital market the layers above seed are continuous. A company that performs gets funded. The seed investor's job is largely finished at entry, which is why so much of the standard advice reduces to a single instruction: put everything into first checks and let the later market sort out the winners.
The layers here are thin and uneven. A good company in Kuala Lumpur or Manila can hit a gap between rounds for reasons that have nothing to do with its performance. The layer above it is underpopulated. The regional funds are between vintages. Growth capital that would be routine elsewhere is simply not in the market that quarter. Early capital is scarce across the board here, and it is scarce in a lumpy way.
A fund that put every dollar into entries has no answer when that happens. It watches its best company stall for want of a bridge that a deeper market would have supplied without anyone noticing.
So we reserve. Not as return optimisation, and not because we expect to price the next round. As insurance against a gap in the capital stack that we already know is there.
Then why not simply write a bigger first check
That is the obvious objection. If you are holding money back anyway, put it in at the start and be done.
Because at this stage companies die from indigestion more often than from starvation.
A company that has not yet proven how it sells does not become more likely to prove it by having more money. It becomes more likely to hire ahead of the answer, open a second market before the first one works, and set a burn rate that assumes a round it has not earned. Capital committed against an unproven motion does not buy learning. It buys more of whatever the company is currently doing, which is usually the wrong thing, and it removes the pressure that would have corrected it.
The same dollar does more work released in stages against real milestones than handed over on day one. That is not caution for its own sake. It is the difference between funding a company's next question and funding its current assumptions.
What the number actually comes out to
Put both caps into the same fund and the count falls out of them rather than being chosen.
Our first fund was $3.4M and made fifteen investments. The second is $5M, and it is built on a split we can state plainly. Twenty percent goes to management fees across the life of the fund. Fifty percent goes into first checks. Thirty percent is held back for follow-ons.
That is $2.5M of first checks and $1.5M of reserves. Across roughly twenty-five companies the first check averages about $100K, and the reserve is deep enough that the handful who earn more can receive several times their original check rather than a gesture.
Twenty-five is more shots on goal than our first fund took, deliberately. A larger fund should buy more of them.
What it is not is whatever the arithmetic alone would allow. Collapse the reserve, put the full $4M into first checks at the same $100K, and you get forty companies. Nothing in the math stops you. What stops you is that a small team cannot be materially useful to forty companies across three countries, and that a fund with forty entries and nothing held back has no answer the first time a good company hits a gap between rounds.
The other direction fails for a different reason. We could have kept the shape of the first fund, held the count near fifteen, and written first checks two or three times larger. More ownership per name, fewer names to look after. That is the version that feeds companies more than they can digest before they have proven how they sell.
The thirty percent is the part that reflects where we invest rather than how we feel about risk. In a market with a continuous capital stack it would be too high, because somebody else would fund the winners. Here it is roughly what it costs to be sure that we can.
Twenty-five is not a cautious number. For a fund this size it is a wide portfolio, and it is wider than our first one. It is simply not forty, and the reason it is not forty has nothing to do with a preference for concentration.
What to ask a manager in this region
The useful question is not how many companies a fund holds. It is what sets the number.
Ask what caps it, and listen for whether the answer is a belief or a capacity. "We hold few names because returns follow a power law" is a belief, and it is available to anyone with a slide deck. "We stop around twenty-five because that is how many founders we can be genuinely useful to, and here is what useful means" is a capacity, and you can check it.
Ask what happens if the round above one of their best companies does not appear. A manager who has not thought about that in this region has not been here long.
Ask what the reserve policy is, what triggers a follow-on, and what they do when a good company raises at a price they cannot justify. The honest version sounds like a policy, not a hope.
And ask them to walk the portfolio from memory. What each company sells, who buys it, what would kill it. It is a crude test and it works.
The takeaway
We want more shots on goal. Anyone honest about early-stage returns does. The question is what a fund gives up to get them, and here the answer is not abstract. It gives up the ability to be useful, and it gives up the money that keeps its best company alive when the market above it goes quiet. Those are expensive things to trade for a longer list of names.
The number of names is an output. Ask what produced it.