Portfolio Construction for New VCs: How to build a fund that works when most startups miss.

You do not need to predict the future to build a strong venture portfolio. You need a plan that fits how venture actually behaves.

Outcomes swing wide. Timelines stretch. Plenty of good teams do not reach the next round on schedule. A small number of companies drive most of the results. That is not a reason to get cynical. It is a reason to get systematic.

Portfolio construction is the part you control. It turns uncertainty into repeatable choices: how many companies you back, how much you own, and how you keep capital ready when the winners emerge.

A quick reality anchor

Seed graduation has slowed hard. Carta shows that 30.6% of companies that raised a seed round in Q1 2018 made it to Series A within two years. Only 15.4% of Q1 2022 seed startups did so in the same timeframe.  

Seed is also getting “stuck” longer. Carta data reported by Axios shows 46% of all seed deals in Q1 2025 were bridge rounds, up from 39% in full year 2024 and 31% in 2022. In the same dataset, Series A deal count fell 79% between Q1 2022 and Q1 2025.  

This does not mean you should stop investing at seed. It means you should size a seed portfolio for the market you are in, not the market you wish you had.

Why portfolios beat “picking” in venture

Most new investors pour energy into selection. You should. But even great selection does not change the shape of venture returns.

Returns concentrate. Andreessen Horowitz analyzed VC fund performance data and found that about 6% of investments generated about 60% of total returns in their dataset.  

A separate Correlation Ventures dataset, shared by Seth Levine, found 65% of financings fail to return 1x, while only 4% return 10x or more (and 10% return 5x or more).  

So the job is not to avoid misses. The job is to build a portfolio that survives misses and still participates meaningfully in the few outcomes that matter. That is the whole game.

Here is what that distribution can look like in plain English:

If you make 20 investments, a realistic base case looks like 13 that return less than 1x, 5 that land between 1x and 5x, 1 that returns 5x to 10x, and 1 that returns 10x plus.  

That last category is rare. It is also where fund outcomes get made.

The Coverage, Ownership, Continuation framework

You can run this framework in a spreadsheet, in IC, and in Monday morning planning.

1) Coverage: make enough investments for the distribution to show up

If you only make a handful of investments, you turn venture into a coin flip. You give yourself too few chances for the few big outcomes to appear.

Coverage does not mean writing tiny checks into everything. Coverage means you set a portfolio size that matches your fund size, your target check size, and your ability to support companies after the first check.

How to apply it this year:

Underwrite coverage using today’s seed to A reality. Two year graduation has fallen from 30.6% to 15.4% across the cohorts Carta highlights.  

Treat “time” as a risk factor, not just “quality.” If seed companies sit longer and bridge more often, you need enough shots for the outliers to show up anyway. Seed bridges hit 46% of deals in Q1 2025.  

Pace deployment so you can keep standards consistent across the fund, not just early in the cycle.

A simple gut check: if one write off would make you abandon your strategy, your portfolio is too concentrated.

2) Ownership: diversification helps you survive, ownership helps you win

Diversification keeps you in the game. Ownership determines whether winning outcomes move the fund.

This is where many early portfolios drift. You write smaller checks to get into more deals. You “get access.” Then a winner shows up and your stake is too small to matter.

How to apply it this year:

Set a minimum ownership target at entry that you can hit repeatedly.

Size checks to reach that target, then let the portfolio count flex around it.

Treat follow on rights and pro rata as part of portfolio design, not an afterthought.

This matters because the return curve does not reward “some exposure.” It rewards meaningful participation in the small slice of outcomes that carry the total.  

3) Continuation: reserves keep you in the game when the breakout arrives

This is the quiet portfolio killer: you back a great company, it starts working, then you cannot support it because you already spent the reserves.

That outcome is avoidable. It matters more now because timelines stretch and bridge rounds show up more often. Again, seed bridges reached 46% in Q1 2025, and Series A deal count fell 79% from Q1 2022 to Q1 2025.  

How to apply it this year:

Decide your reserve policy before you deploy.

Protect reserves from “just one more new deal.”

Use simple follow on rules so decisions stay consistent under pressure.

A clean way to decide follow ons: write down three milestones that earn more capital. Keep them measurable. Keep them boring. Boring beats improvisation.

The portfolio mistake most new investors make

They build a portfolio that assumes smooth progress. They assume seed leads to Series A on a tidy timeline. They assume most companies that “look good” will raise. They assume reserves will be there later.

But the market does not care about your assumptions.

A better approach: build a portfolio that still works when progress slows, pricing becomes uneven, and only a small slice of companies drives outcomes.

A practical portfolio model you can use

You do not need a complex simulation. Start with a simple map.

1. Portfolio size: choose an initial number of positions that fits your fund and your check size.
2. Ownership target: define what meaningful ownership looks like for your fund economics, then size checks accordingly.
3. Reserve plan: decide how much capital you hold for follow ons and how you release it.
4. Timeline reality: model the portfolio with today’s seed to A conversion rate, not a best case.  

This model will not predict outcomes. It will prevent avoidable failure modes.

Conclusion: design for resilience and upside

Venture has brutal math. Most investments do not return capital, and a tiny minority drives most of the returns.  

So design your fund for resilience: expect many bets to falter, place enough bets for the distribution to show up, and hold enough reserves to stay in the game when timelines stretch.

Then design for upside: maintain ownership that can matter in the outliers, and support the breakouts when they earn it.

You cannot remove uncertainty from venture. You can stop letting it surprise you.

‍

‍