Untangling Luck from Skill
Why Indexing at Seed Doesn’t Work in 2024
Follow the discussion on x.com
Venture capital has a habit of relying on outdated practices and data. I’m seeing a lot of commentary on what fund sizes are most effective on Twitter, but the truth is that the fund strategies of individual firms are so nuanced that when you zoom too far out on the data, it loses a lot of signal.
More recently, I’ve seen a lot of investors and LPs moving further upstream and frankly, I’m puzzled why. I suspect that these folks (all of which are far smarter than I am) are responding to the data they are seeing (the returns of previous seed funds, which were amazing) and are concerned about some of the fund size dynamics as mega-fund consolidate. Perhaps this works, but what seems more likely is that AVERAGE seed returns will be remarkably bad for the AVERAGE firm and investors / allocators need to be heavily disciplined in entering the market.
Back at our AGM in 2023, I highlighted a bit of data that suggested that the seed stage market was the least attractive it had been in at least a decade. Historically, the valuation premium of late-stage to seed-stage companies ranged between 20-25x. However in 2021, this ratio propelled upward, peaking at 40x in 2021. Today, that premium has dropped to below 10x, the lowest we’ve seen on record. As we factor in the increased dilution from larger A/B rounds (today’s funding rounds are significantly larger than in years prior to 2021’s bubble), this makes seed investing the least attractive it’s been as far as back as I have data available. This data is a bit old now, but I believe the point still holds.
Yet, the interest in early-stage investing rages on. I returned to the latest data to dive in further and fell upon an interesting observation about early-stage investors in the 2014 vintage (unfortunately, I only had the data for “early-stage”, not “seed”, but I believe this is a reasonable proxy ). If you were an early-stage investor in 2014 and saw your companies raise a growth round in 2021, you benefitted handsomely as the average valuation premium between those two cases (i.e. 2014 early and 2021 growth) was over 68x. What a time to be alive! The dynamics for seed stage managers were simply incredible and I believe that has proven out in the performance of many of those funds (see the Cambridge Associates data - the quality of the returns for this vintage and its surrounding years are the best of the last 25 years).
As the early 2010’s vintages are returning boat loads of capital to the LPs and GPs of those funds, LPs are doubling down on those funds even as they scale to much broader strategies. This is amplified by LPs AND GPs pursuing more exposure to what has been working (this is referred to as recency bias). This, however, could prove dangerous.
To achieve that same degree of favorability for investors in 2024 as those had in the 2014 vintage a decade ago, the average valuation of growth stage companies in 2031 would need to be a staggering $11.54 billion. Yes, the average growth stage company would need to be a decacorn. Possible? Yes. Likely? No. This seems like a disconnection in the market.
What does this mean for VCs? The overall basket of early-stage investing is nowhere near as attractive as it was in 2014. The AVERAGE manager is likely going to produce far less exciting returns in today’s environment, than those of these previous vintages. Early-stage investors need to figure out their “right to win more” than ever and practice discipline to ensure they are breaking away from the crowd. Those that do are likely to deliver returns that resemble those that look far more like the 1999-2000 vintages.
What does this mean for LPs? They need to look back at managers from prior vintages and determine whether past performance was driven by skill or luck. Michael Mauboussin often writes on the entanglement of these and the danger of focusing on outcome > process. Turns out a lot of prior managers benefitted from a bull market (some of which were more lucky than skillful), so it may be worth considering new managers capable of addressing the needs of today’s market rather than assuming the prior returns will persist.
Venture is a slow business. But those who look back to the past hoping to replicate it without evaluation or the present and future conditions seem destined to fail. The market changes much faster than our business does, so we can’t simply rely on best practices of what produced success - we (VCs and LPs) need to develop our own best practices for the needs of the market and where it’s presenting the best opportunities.