#NotesOnStrategy | Seed-stage Venture Capital Portfolio Construction

  1. It is my belief that the seed-stage technology venture capitalist’s only goal is to benefit disproportionately from uncertainty. To do this the best seed-stage venture capitalists seek startups that fit an investment thesis, and make investments before other investors would normally invest.
  2. When I think of uncertainty, I am thinking of a state of affairs in which I have limited information and knowledge, and must make a decision whose outcome I can’t predict because the future is unknown and unknowable. I do not think one can measure uncertainty quantitatively.
  3. When I think of risk, I am thinking of undesirable future outcomes some of which I can enumerate quantitatively.
  • VC has always been a power law business, so big hits drive portfolio returns . . . and the big hits are getting bigger, but on the other hand, pricing going up is going to cut returns, not only of the big winners, but also of the middle OK part of the portfolio.
  • Remember: Opportunity = Value — Perception, and the industry is so good at blowing up perception, but true Value is more fleeting — and, if we’re being pedantic, is the discounted value of future (positive) cash flows. [My comment; The dichotomy between Value and Perception that Chris is referring to explains why the data from Correlation Ventures seems so jarring at first glance.]
  • But everyone’s bought into the power law dogma, so unicorns are getting bid up, often with pricing for perfect execution, following winds, and fair seas . . . any hiccup (systematic or idiosyncratic) will lead to a lot of stranded unicorns, or as Bryce (of Indie.VC) calls them, “donkeys in party hats” . . . Speaking of which, I think his efforts over at Indie.VC have been a creative and thoughtful search for opportunity in the context of Value — remember, Opportunity = Value — Perception.
  • At the end of the day all that matters is Moolah in da Coolah — the distributed to paid-in-capital multiple that a fund ultimately achieves. Here was my effort at thinking about some of these portfolio construction issues in the context of valuation environment: All About the Benjamins.
  • But I remain really nervous about the environment. As Henry McCance at Greylock told me in 2001, VC works well when time is cheap and capital expensive. When that relationship is reversed, trouble ensues.
  • This means you should do a LOT more deals, unless you pick better than Sequoia. Of course depends on dealflow, selection, and stage, but if you start investing at seed-stage, most GPs with portfolios with N < 50 companies are playing Russian roulette. If large outliers of > 20x happen only 1% — 2% of the time, basic math would suggest a portfolio size of N > 100 is more rational.
  • [My comment; Yes, The basic math certainly supports that conclusion. Though, I wonder if there are nuances the basic math doesn’t capture. Do you think there are conditions under which one might justify deviating from that prescription?] Well if you’re a subject matter expert and/or have excellent access to dealflow or an established brand, you might choose to build a concentrated portfolio — but again you’d have to convince yourself, hopefully based on data, that you’ll generate a higher percentage of outliers than the average VC.
  • [My comment; Got it . . . Though, one has to wonder if there’s such a thing as a subject matter expert when it comes to predicting how the future will unfold. But, I see why that approach would make sense — in some rare cases. Who would you say does that really well? Anyone? CVCs?] Well for specific IP-related areas, people who are scientists/PhDs/professors might have an advantage. For industry verticals, maybe experienced business or technical folks. Famous people and/or VC firms might also have an advantage. Not sure about CVCs, unless specific IP perhaps.
  • [My comment; That reminds me of one of the points Richard Zeckhauser makes in Investing in the Unknown and Unknowable; collaborate with other investors with superior knowledge of specific industry verticals, among other things.]
  • In my experience the skew is almost systematically massive. With one fund returner — your fund is likely to be fine, with two — it’s likely to be great, etc. The top 2 to 4 exits will likely more than 2x+ the fund, the next 5 to 8 exits together will return 1x the fund, and the rest might lose some money.
  • Hence I’m always thinking — never invest in anything that can’t return the fund — which is, of course, a function of ownership and upside, with upside uncapped if you want to have a shot at a “glimmer of greatness”.
  • Tie decision making to absolute rules (eg I need a moat in every investment) and you’re introducing a systematic bias / hard screening criteria. That may be fine — but don’t confuse a disciplined decision-making process and method with fixed decision-making “rules”. The only rule that should be fixed in my opinion is: the potential for unlimited upside exists within the fund’s duration. [My comment; I agree. It’s less about absolute rules, and more about ensuring I have thought about the issues and have consciously decided one way or the other. I think it’s important to do that when things are uncertain and and failing to think through the issues is costly. As you point out, it’s more about process than hard rules . . . Especially — at the stage at which I’m doing this there are no moats yet, but I need to consider the possibility that they can develop over time, and why that may happen if things work out. But, point taken re systematic bias.]
  • I wouldn’t go as far as to say the data is completely useless, though I see the argument Arjun and Jerry are making. If one assumes that the two sets of data suggest possible probability distributions, then I think the interpretation should be that, First, any portfolio construction that assumes less than 50% of the portfolio going to zero is almost certainly naively over-optimistic.
  • In my work on modeling portfolio outcome scenarios for REFASHIOND Ventures’ first fund, I have gone as far as assuming 90% of the portfolio goes to zero. I have not modelled 100% of the portfolio going to zero because that’s obviously the trivial case — if that happens we’ll most certainly have bigger problems to worry about.
  • Correlation Ventures has not made public a version of this analysis that presents the results in a manner similar to that done by CBInsights. It is likely that they were unable to get to that level of granularity given the source data, or they may prefer to keep that specific version of the analysis as a trade secret.
  • To your specific question; Given skewed outcomes what’s the right strategy for a small fund? in general smaller funds will have less opportunity to consolidate ownership in outperforming companies. Thus I think the right strategy is to seek more exposure — place more bets. [My comment; Thanks, Seth. There seems to be a tension between seeking more exposure AND getting as much ownership % as early as possible, and reserving capital to follow on in later rounds. Especially, in the context of a fund 1 with say $10M — $20M of AUM. Thoughts? If you had to choose?]
  • There’s definitely that tension. In an ideal world you’d have enough capital and enough early, but predictive, data to consolidate into your best companies. Larger “Series A” funds do that regularly, but with a seed fund you have challenges with both capital and information.
  • You have a small fund and by the time you have to make a follow-on decision, or more ideally preempt a round, you don’t really have that much more data about the opportunity. That’s really the argument for placing as many early bets as possible.
  • The ones that run on you drive value and you’ll have plenty of exposure to the potential for upside. It’s easy to say in hindsight that you “knew” something was going to be great, but how often do you have that reliable insight between Pre-seed, Seed, Seed+, and Series A? [My comment; I know that’s a rhetorical question, but there’s so much pressure to seem prescient, all knowing, and fully certain about the future . . . But yes, it’s generally hard to know. So it seems things point back to the decision-making process, and a bit of luck, as others have alluded. I’ve stopped paying attention when people tell me to “sound more confident about what you’re saying” . . . How can one be confident when decisions are being made under extreme uncertainty? I’ll stop bloviating.]
  • Re: Luck, that reminded me of this ancient post of mine (2005 — I was an associate) about what makes a good venture capitalist and David Cowan’s comment, which has always stuck with me. [My comment; I couldn’t find the blog post Seth is referring to, but I found a discussion thread elsewhere in which David Cowan, of Bessemer Venture Partners says the correct answer to the question “What do you think is the most common trait among successful venture capitalists?” is “Luck.” This reminded me of a blog post I wrote about Fab.com around the time I was studying economic moats; Vcs often rail about startups raising too much money at valuations that are too high to sustain, but VCs too sometimes make investments that assume perfect knowledge, perfect execution by the team, perfect market adoption . . . etc, to the point Chris Douvos made earlier. I wonder how the early stage venture funds that invested in Fab.com have fared. I have not looked that up yet.]
  • To maximize the probability of success, fund GPs raising a first fund should perform the portfolio construction exercise based on their investment thesis, their knowledge of the market in which they plan to operate, and how much capital they need to prove that they can execute the thesis. Although, macro considerations are always a concern for managers raising their very first fund, this exercise should be performed independent of considerations from prospective LPs who make comments such as “You should not raise more than $X for your first fund.” In other words the dog should wag its tail, not the other way around. He discusses that topic here: Is there ideal portfolio construction for seed funds?
  • Although many will pejoratively speak of large portfolios as “spray and pray”, doing so ignores years of probability and statistics, and likely over-weights skill versus luck.
  • That said, larger portfolios do come with some challenges: Scaling value add to portfolio companies. Requiring larger exits (given smaller initial checks/ownership vs. concentrated portfolios) for definition of outlier (fund returner). Tougher to make follow-on decisions.
  • When pitching LPs, at a minimum expect to discuss
  • Target Fund Size,
  • The stage at which the fund will make its initial investments; Pre-seed, Seed, Series A, Series B . . . or later,
  • Average Initial Check Size,
  • Average Initial Ownership Target — which establishes the valuation bands within which the fund should invest,
  • Total Number of Startups in the Portfolio — a range is the norm,
  • The fund’s Follow on Reserve Percentage, and
  • The Fund’s Investment Period.
  • Demonstrate that the new manager has a unique investment thesis based on knowledge about a market that is under-appreciated by other investors,
  • Demonstrate an ability to source deal-flow in a manner that is both efficient and proprietary — given the rule of thumb that the best VCs see 100 startups for every 1 investment they make,
  • Demonstrate the ability to pick startups that have a high probability of returning the portfolio within the duration of the fund,
  • Demonstrate an ability to manage the portfolios losses in a way that maximizes the likelihood that the fund will meet LPs’ expectations, and
  • Demonstrate an ability to execute the fund’s started portfolio construction and portfolio management strategy under real world scenarios.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store