Portfolio Construction: How Top VCs Build Winning Funds

Portfolio construction is the architecture of venture capital. It's the set of decisions—how many companies to invest in, how much to put into each one, how much to reserve for follow-on investments—that determines whether a fund can generate the returns that LPs expect. Get it wrong, and even a fund full of great individual picks can produce mediocre results.

Most aspiring VCs focus obsessively on deal selection—picking the right companies—and completely ignore the structural decisions that are equally important. Understanding portfolio construction is what separates someone who can identify good startups from someone who can actually manage a fund.

The Basic Math: Fund Size and Check Size

Everything in portfolio construction starts with fund size. A $50 million seed fund operates completely differently from a $500 million growth fund, and both operate differently from a $2 billion multi-stage platform. The fund size dictates check sizes, which dictates the number of companies you can invest in, which dictates your entire strategy.

Here's a typical framework for a $100 million early-stage fund. You might allocate 60% of the fund ($60 million) to initial investments and reserve 40% ($40 million) for follow-on investments in your best-performing companies. If your average initial check is $2 million, that gives you 30 initial investments. Your follow-on reserve of $40 million might be deployed across your top 10-15 companies, with an average follow-on of $2.5-4 million each.

These numbers aren't arbitrary. They reflect the fundamental reality of venture capital: most investments will fail or return modest amounts, a few will do well, and one or two might return the entire fund. Your portfolio needs to be large enough to give you enough shots on goal, but concentrated enough that your winners can meaningfully move the needle.

Reserve Ratios: The Hidden Strategy

How much of the fund you reserve for follow-on investments is one of the most consequential decisions a fund manager makes—and one of the least discussed. Reserve too little, and you'll get diluted out of your best companies when they raise subsequent rounds. Reserve too much, and you're sitting on dry powder while missing new opportunities.

The industry standard for early-stage funds is roughly 40-50% reserves. But this varies by strategy. Seed funds that don't plan to follow on aggressively might reserve only 20-30%, deploying most of their capital into initial checks and accepting the dilution. Late-stage funds might reserve 50-60% because the follow-on checks are so large. Some firms, like Benchmark with their equal-partnership model, take a more concentrated approach overall.

The tricky part is that reserve decisions have to be made before you know which companies will succeed. You're committing today to have capital available for follow-ons that might not happen for three to five years. If you're too generous with initial checks early in the fund's life, you might not have enough reserves to support your winners later. This is a mistake first-time fund managers frequently make.

Concentration vs. Diversification

This is the great philosophical debate of venture capital. On one extreme, you have the Benchmark model: five partners, one fund, roughly 15-20 investments, enormous concentration. On the other extreme, you have the 500 Startups model (now 500 Global): hundreds of small checks, maximum diversification, betting that a broad portfolio will capture outliers through sheer volume.

The data is nuanced. Academic research on VC fund returns suggests that funds with 20-40 investments tend to produce the best risk-adjusted returns. Fewer than 20 and you're taking on significant concentration risk—one bad outcome can sink the fund. More than 50-60 and you're diluting your winners' impact on overall fund performance. Each fund can only have so many 100x outcomes, and if they represent too small a percentage of the portfolio, even massive individual winners don't move the aggregate return.

The concentration question also affects how partners spend their time. A fund with 20 companies can be deeply involved with each one. A fund with 100 companies is spread thin, which means less value-add per company but potentially more serendipity and a wider information network.

Follow-On Strategy: Doubling Down on Winners

Follow-on investing is where portfolio construction gets really interesting. The basic idea is simple: invest more money into the companies that are working, and let the ones that aren't die. But the execution is full of psychological traps and structural complications.

The most common mistake is what the industry calls "watering the weeds"—putting follow-on capital into struggling companies because you're emotionally invested or because the founder is persuasive. The data overwhelmingly shows that follow-on capital is best deployed into companies that are clearly working: growing fast, hitting milestones, attracting strong co-investors. Throwing good money after bad is how funds destroy returns.

The best follow-on investors are ruthlessly disciplined. They have clear criteria for when they'll invest more—specific revenue thresholds, product milestones, or market validation signals—and they stick to those criteria regardless of the personal relationship with the founder. This is harder than it sounds because venture capital is an intensely relationship-driven business.

The Math Behind Fund Returners

Here's the math that every aspiring VC needs to internalize. For a $100 million fund to return 3x to its LPs (a strong result that puts you in roughly the top quartile), it needs to generate $300 million in total returns. After management fees consume roughly $20 million over the fund's life, you have about $80 million in investable capital to work with.

In a typical 30-company portfolio, the outcome distribution might look something like this: 10 companies return zero (total losses), 10 companies return 1-2x (modest outcomes that roughly return their invested capital), 7 companies return 3-5x (solid outcomes), 2 companies return 10-20x (strong winners), and 1 company returns 50x or more (the fund maker). That single 50x return on a $2 million check is $100 million—already more than the entire fund size.

This power-law distribution is the fundamental reality of venture capital. Your portfolio needs to be constructed in a way that gives you the best chance of catching a 50x-plus outlier, while maintaining enough diversification that you don't rely on any single bet. It's a delicate balance, and it's why portfolio construction is as much art as science.

Understanding these mechanics isn't just useful if you want to be a fund manager someday. It shapes how every VC thinks about deals. When a partner says "this company can't return the fund," they're doing portfolio construction math in their head—calculating whether the potential outcome is large enough to matter given their ownership percentage and fund size. Learn to think in these terms, and you'll understand why VCs make the decisions they do.