Like a lot of people, I’m excited about crowdfunding, and specifically the crowdfunding of startups now that’s it’s legal in the US. Based on my own experience investing in startups, here are some thoughts and issues that come to mind regarding startup crowdfunding.
1. Startup financings tend toward the extremes of being very oversubscribed or very undersubscribed. If you graphed out investor interest, it would look like a “U”. This is primarily the result of signaling – once a few investors commit (especially high quality ones), other investors pile on. If investors don’t commit, other investors start to wonder what’s wrong. So when you consider startup crowdfunding, it’s important to distinguish the oversubscribed cases from the undersubscribed cases. (Although one counter to this is that the U is the result of an inefficient market – when the crowds get involved valuations will float to their market clearing prices).
2. Historically, startup investing returns have tended to obey power laws (Peter Thiel has a good discussion of this phenomenon here). The vast majority of the returns came from the breakout hits. And if you go back and look at the early financings of breakout hits, a lot of them were hotly contested and oversubscribed. If amateur investors had been trying to invest in those startups via crowdfunding sites, they probably would have been squeezed out. If those amateurs were part of a syndicate, the syndicate lead would have felt pressure to drop them, at least for those hot deals. (Counterargument: the power law is caused by the myopia of traditional investors looking for the next Google. The crowd will be able to find new investments that greatly expand the set of successful startups)
3. Crowdfunding works best when the backers have special knowledge about the project that leads them to fund things that otherwise would have been overlooked or undervalued by traditional investors. This happens, for example, in the Kickstarter video games category, where most of the backers are game enthusiasts. The most promising scenarios for startup crowdfunding are where the backers are potential customers of the product (e.g. HR managers backing new HR software). This could also solve the adverse selection problem, as the startup founders would probably favor these backers over traditional startups investors.
4. When you look at the biggest crowdfunding markets – publicly traded stocks on NYSE, NASDAQ, etc – you find that a) In general, non-professional investors lose money when they try to pick individual stocks. This suggests that something similar to mutual funds would be the best mechanism for amateur participation. b) There is a constant cat-and-mouse game between regulators and sketchy market participants. If this happens with private financings, and more and more rules and regulations get added, many of the advantages of being a private company could go away.
5. Most successful seed investors will say that it is mostly about investing in great people, and it is very hard to evaluate people even after multiple in-person meetings. If founders are going to be evaluated online without in-person meetings, great care has to be taken to make sure the evaluation mechanisms are sufficiently nuanced and reliable. (The counterargument is that this might be true when individual professional investors evaluate startups. In the aggregate, the crowd can outsmart individual professionals even with fewer direct interactions.)
6. One way to look at startup crowdfunding is as the first step in a process that includes additional steps that prevent adverse selection, sketchy behavior etc. For example, a startup I know raised money recently from a single lead investor and then found additional investors via a crowdfunding site. They ended up rejecting many of the interested investors but found a few useful investors that they otherwise wouldn’t have found. In this model crowdfunding looks more like LinkedIn for investors – extremely useful for connecting, but only the first step of a process that includes interviews, reference checks, etc.