I have and always will have empathy for asset allocators. It’s a privileged job, but by no means is it easy.
LPs primarily entrust third party managers to make intelligent, risk-adjusted decisions on assets they don’t have access to, or at the least are not situated to underwrite or manage to the same degree. This is quite difficult for an LP who allocates across assets. Consider finding the optimal balance in an ever-changing environment across multiple asset classes i.e. private credit, buyout, real estate, commodities, etc. Even for allocators deploying to a single asset class like venture capital — it can be hard. In the near-term, rising interest rates have caused folks to rightfully sharpen their pencils to invest in managers that can sufficiently exceed the risk-free premium.
I’ve always enjoyed reviewing Christoph Janz’s SaaS Funding Napkin. It’s on a napkin because it is a sketch — a mix of art and science. What is good enough to raise a round for a SaaS company in a given year? How about for the fund that invests in that SaaS company? Cambridge Associates’ benchmarks are a common and okay place to start, but even they say it’s less useful to compare in the early days as funds settle into their ultimate quartiles between 5.8 years and 6.8 years.
Additionally, according to two well-known VC FoFs I’m close to, ~50% of their profits accrue after year 7. Many managers — myself included — raise new funds in the last year of the deployment period which is typically in years 3–4 of fund life. Sometimes the timeline is shorter for higher velocity investors. What should an LP optimally focus on in the underwriting decision, then?
The best predictor of future behavior is past behavior / past performance in a similar situation. Assuming the strategy is comparable, we must look at past funds to determine what repeatable performance can look like. Too often I see LPs focus on activities within the current fund being raised. The newly marketed fund may have early momentum, perhaps even a mark-up or two, but so little has been “paid-in” to make conclusions off of in the first year or so.
Enter the Seed Stage Enterprise VC Funding Napkin (at year 4 of deployment of the prior fund) which I’ve sketched below. It too, is also a mix of art and science — based on Cambridge metrics, Carta data, proprietary FoF data, and general guidance which is subject to interpretation akin to the SaaS Funding Napkin. See below:
Seed Graduation Rates: This metric is fairly straightforward. High-performing VC funds have high follow-on rates to Series A’s. Taking it a step further — a leading FoF referenced to me that if a seed VC fund has 50% of the portfolio graduate to Series B, the likelihood of a fund achieving above a 3.0X DPI is over 80%.
Pro-rata Accessibility: Seed investors should desire the opportunity, whether contractual or not, to re-invest in their winners. Not all GPs are given the opportunity to protect their initial ownership stake from dilution or grow it. Having the opportunity to invest in follow-on rounds expands the investment universe for the GP — who should have better information on the asset than others — is a good thing. The follow-on check is more likely to have a modest multiple vs. the entry ticket, though it might still be accretive to the fund’s overall performance. If the GP wishes to have a more meaningful reserve strategy in their new fund(s), this is additionally important. GPs should aim to have pro-rata offered to them the majority of the time, if not always.
Entry : Reserve Ratio in Top 3 Companies: This ratio is a metric for executing that pro-rata in fund winners (top 3 positions). Investing $1.0M (entry) in a seed round and $2.0M (reserve) in the follow-on is a healthy ratio for Preface III, for example. Going beyond that can encroach on single asset concentration limits stipulated within an LPA (commonly 10% but I’ve seen up to 20%). A leading FoF highlighted an even higher ratio (1:3) in their best performing underlying businesses.
Weighted Blended Entry Cost (post-Series A): The average VC backed exit is ~$120.0M today. Too often LPs look at entry price / ownership and not the dollar-weighted blended entry cost between entry and follow-on, which is critical if the fund has a reserve strategy (many do). We all hope for asset outperformance — but having low blended cost will generate attractive blended return in “average” exit environments as well. GPs may sometimes put a small entry check in and “logo buy” at a high price once an asset gets a strong follow-on offer. This metric should illuminate the downsides of that behavior. Many FoF’s have done Monte Carlo analyses on their funds’ reserve strategy — a notable one mentioned that if their underlying VC funds just wrote the same check size evenly across positions with no reserve strategy — they’d outperform the normative fund performance 70% of the time. That might be impractical but bigger earlier for seed tends to be a good rule (particularly in boom times like 2021).
Cumulative Burn Multiple: This is calculated by summing all $ burned to date vs. current ARR across the portfolio. Many folks reference burn multiple i.e. dollars burned in 12 months vs. net new ARR which is a helpful efficiency metric. However it’s possible that a company with a healthy current burn multiple may have burned a ton of cash prior to the current year, correlating to low management ownership (which we don’t want — founder led companies tend to outperform). Technology companies should fundamentally be capital efficient and startups cost less than ever to launch. In 2023, a Series A investor would look at $2.5M burned to get to $1.0M in ARR as a highly refreshing and attractive foundation for a company’s future growth.
% Ownership in Winning Positions (top 10%) vs. Average Ownership (post-Series A): We calculate this metric by taking the average percentage ownership of a portfolio’s best post-Series A companies (top 10% of them) and dividing it by the average percentage ownership of the overall portfolio. For example, if the average ownership of a set of post-Series A companies is 4%, but the ownership of the top 10% of the portfolio is 5%, the ratio is ~1.3X.
Tail Ratio: Modeled after the “power law” — Sumeet Gajri’s ‘Tail Ratio’ metric offers LPs a tool for assessing what’s required to return a fund of any size. The higher the ratio, the less likely the GP will return the fund on any given investment based on their fund model. Another way to view this metric — can one position be “outlier enough” to return the fund?
TVPI / DPI: Hopefully most of us know what these metrics mean. While directionally helpful and worth including, I see too many folks overweighting them early on. For example, I’ve seen fund TVPI tempered with the advent of convertible notes (quite common today) that don’t change holding valuations. DX, a high-growth company with fund returning potential in Preface II, is a strong example which is still held at 1.0X MOIC. DPI is also hard to see given the holding period of private investments is 5.0+ years on average. Some FoF’s look at TVPI / DPI while excluding the top and bottom quartile of investments to see “repeatability”; however again, this is less useful to assess by year 4 of fund life which is what we’re attempting to accomplish.
% of Portfolio Led by Tier 1 VCs at Series A: Many seed stage VCs reference how they co-invest with established firms like Sequoia, Benchmark, Accel, or a16z. When that happens is critical. According to AngelList, having a Tier 1 fund (multi-stage) invest in a company at seed correlates to a neutral or even negative Series A, defined as a 3X+ mark-up at normative dilution led by a reputable firm. That correlation changes quite a bit when the Tier 1 multi-stage firm follows the seed investor in a Series A. A notable FoF mentioned if one of the aforementioned groups lead the seed manager’s Series A company, the likelihood of a “homerun” i.e 10X return increases by 4X.
Early Recycling: Highest performing seed funds tend to recycle and deploy ~105% into companies. At year 4 one would hope some dollars have been re-allocated to existing winners or in new companies i.e. “more shots on goal”.
I continue to believe that there is an imperfect marketplace for LP capital. Who is allocating, how much, what’s the process, and where it goes is largely suboptimal. I’ve seen this happen first-hand but mostly second-hand in the 16 years investing in venture, as LP capital has herded to too few groups. My hope is that leveraging objective metrics such as the above will be helpful to both LPs and GPs. LPs will hopefully not misallocate to managers with an aversion to measurement or worse, to managers that are more art than science (when you need both).
Thank you to Yuri Sagalov, Winter Mead, the Preface team, and other close, experienced allocators for their feedback and late night debates on this lively topic. Further feedback is very welcome.