VC-istan 3: Performance

No analysis of VC-istan would be complete without confronting the performance of venture capital as an asset class, which is, by all accounts, terrible.

Recall that the VC-istan Hypothesis asserts that the leading venture capitalists have assembled something that, while its disposable companies operate separately and may compete with each other, effectively functions as one postmodern corporate body, of which the leading investors (not founders, who are middle managers given better titles) are the true executives, and in which the nominally separate companies (whose investors and boards all know each other) are more accurately framed as departments. It seems to be, then, a poorly-performing company. It’s worthwhile to analyze that, since I’d argue that the cultural problems with VC-istan are bidirectionally related to its poor performance.

Normally, poor performance of a company is felt in layoffs. In VC-istan, it’s experienced by the shutdown of its subcompanies. This would be tolerable (people go into startups knowing there is risk) except for the fact that the failures are correlated. They happen in waves. Job loss usually isn’t a big deal in Silicon Valley, but is devastating (some people go down 50% in salary when the market tanks) when it happens to a large number of people at once. Indeed, if there’s one sin committed by the VCs against their counterparties, it’s not that they mislead about fair valuations (they don’t) but that they knowingly engage with people who misprice correlations. Savvy investors understand that, to price an asset, one must consider its correlation to others– a portfolio that is likely to tank when the economy tanks (and, quite possibly, you lose your job) is worth less than one with the same return, but that is independent of broad market conditions. The latter is more likely to have the money when you need it.

Venture capital is a fair-weather friend. There’s lot of opportunity when everyone (especially landlords) is getting rich, but it’s very easy (especially considering the age discrimination problem) for a whole generation to be tossed out like yesterday’s dogshit. Venture capital returns are dismal, most venture-funded companies fail, and most of these so-called “startups” are terrible places to be employees. Let’s hope that VC-istan isn’t the new economic regime that it claims to be; it’s not a very good one.

Root causes of poor performance

Why does VC perform so poorly? One might argue that they pay “too much” for companies, but I don’t think that’s true. If anything, they pay too little, because they’re mostly funding crap. After a typical Series A, option pools are so limited that even high-level hires are lucky to get more than $20,000 per year at-valuation– in exchange for an equal or larger pay cut. Startups become stingy not only with pay and meaningful benefits, but also with equity. This makes it hard for them to get the best people.

To be fair, I would argue that a company with outside investors shouldn’t be focused on hiring the best people. Hiring excellent people isn’t a business objective; turning a profit is. It’s quite possible to make a profit without staffing up with top talent; Wal-Mart does it. However, the get-big-or-die startups funded by the most stylish investors often do need the highest levels of talent, and they make two mistakes on that front. The first is that they underpay, not only in terms of salary, but also in terms of the equity that’s supposed to make the wage cut tolerable. This still works sometimes because there are some clueless people (typically, young ones) who are technically talented in spite of being extremely gullible, but it doesn’t scale because the pool of talented suckers is a small one. The second and, in my opinion, more devastating factor is that they grow too fast and overhire. Talented people will work for startup wages for a chance to own, but if you’re offering dimes (0.1% slices) and pennies (0.01%) while underpaying, you will find few talented people, keep almost none, and– if you have to hire fast, because your investors mandate rapid growth and aggressive risk-taking– you’ll flood the company with mediocrity.

Of course, paying more doesn’t guarantee getting good people. If you don’t know what you’re doing, it can have the opposite effect. I’m not going to cover that issue, and my point isn’t necessarily that VC firms give too little money; the problem is that the companies (perhaps due to investors’ demands, perhaps on their own accord) themselves are stingy even with equity, which is the only reason people put up with the shortfalls in wages, benefits, and autonomy. As to whether that’s the fault of the VCs– to whom founders are eager to shift the blame for their terrible employee options grants, but who are rarely in such discussions to defend themselves– or a more general problem with VC-istan culture (whose worst malefactors, in my observation, aren’t themselves VCs) I have no idea.

How could venture capital win?

Venture capital would have two ways of returning a real profit. One would be to buy low: to be better than anyone else at assessing fine-grained variables, such as the talent of the team, and get great companies cheaply. Trust me, they don’t have that skill; I’ve seen plenty of idiots get funded. They’re no better at picking talent than an average college-educated person. In fact, they’re worse, and I’ll get to the reason for that, later on. The other would be to add value to the businesses they fund, and then sell “high”. I don’t think, in general, that they do. On the whole, it seems pretty obvious that they detract. Why? Because the venture capitalists are the true executives of this postmodern entity that I’ve named VC-istan. They succumb to a personal need to manage.

Many day traders and poker players have, despite knowledge and superiority over their domain of confidence, been destroyed by trader boredom– that need to be active at all times, even when the additional work (outside of one’s domain of competence) will, at best, add zero-expectancy variance (noise) and, at worst, cancel out some of the gains. I think this is the case of VC-funded startups. It would be better for investors to give money and advice, but let the companies operate almost completely autonomously. This might not fulfill the emotional need of investors to be decision-making executives, but it would make those businesses a hell of a lot more successful, and that would improve their returns. By removing themselves from active management and taking a strictly advisory role, venture capitalists would do less work and get better returns.

This wouldn’t solve everything, of course, and it probably wouldn’t be enough to fix the VC-funded ecosystem. Founder quality is a much bigger problem than anything having to do with investor involvement. The investors are responsible insofar as they back the wrong sorts of people, but they aren’t directly causing the toxic cultures seen in most of these companies.

Founder quality

What are the causes of low founder quality in the VC-funded world? I think there are a few that deserve mention, and I’ll address them from least to most potent.

First, VC-funded companies are constantly fighting for survival. This might be said of all small businesses, and to some degree it’s true of all, but not in the same way. A non-VC business can grow happily at a “pokey” 32% per year, quadrupling every half-decade, and that’s considered healthy. Hell, most people would the thrilled to have their personal income grow at a mere 10 percent per year. That doesn’t cut it in VC-istan, where 5% per week is expected. That is, of course, utterly unsustainable at scale, and requires aggressive risk-taking for the small.

Additionally, per-week growth is hard to measure for companies that (a) might not have substantial revenues, and (b) aren’t publicly traded. The result is that companies are expected to post these increases in key metrics like “eyeballs” or usage, but before a company is profitable, that actually means that there is an exponential growth in costs (and, therefore, often losses) that keeps the business dependent on external capital, and therefore under VC-istan’s control.

Most small businesses must negotiate the market to survive. In VC-istan, the firm must battle not only the market, but demanding investors (who won’t accept 40% year-on-year growth) as well. What kind of person would sign up for this? In general, it attracts overconfident, exceptionalist people who, given the freedom to do so, will always overpromise (and underdeliver). It’s not that they do so maliciously; that’s the temperament that VC-istan’s way of doing things attracts.

Most investors would agree, privately, that they’re looking for people with the talent to make more with less; in other words, those rare people who can start or run a company on a shoestring budget and, against the odds, give it life. Typically, you won’t be able to hire a competent programmer at 50% of the market wage without offering serious (founder-level) equity. If you did luck into a golden hire, you wouldn’t be able to repeat it, and the bad programmers would sink the enterprise. It’s just very hard, and not at all typical, for a company that poor to have any life, much less rapid growth potential. Venture capitalists want to work with the atypical.

From first principles, I don’t think there’s anything wrong with that. Some people (perhaps 1 in 100,000) do have that sort of personal charisma– the kind that could convince a $200,000 per year programmer to take $80,000 and 0.05 percent– and I can’t blame VCs for wanting to fund those natural arbitrageurs. However, for each of them, there are hundreds of slimy hucksters who use deception (most typically, about the career benefits of the startup job) to bring talent in, and extortion (threats of reputation damage, negative reference, and exclusion from future ability to fundraise) to keep it. Just as 99.99+ percent of people selling $100 bills for $50 should be expected to be counterfeiters, and not charitable; most of these “more-from-less” operators are not preternaturally charismatic, but deceptive shysters.

The most devastating cause of VC-istan’s founder quality problem, however, comes from the fact that the VC-funded world is a feudal reputation economy. The people who excel in such settings tend to fall into two categories. The first are those who use extortion (including, often, hired physical intimidation) to protect and launder their reputations. In larger companies, they’re most likely to hide behind HR and offer generous settlements to wronged employees, but in the VC-funded environment where every company must operate on a shoestring, they have only the stick, and frequently take the extortion route. The second class are those about whom no one can say anything negative. They don’t piss anyone off, because they don’t really do anything, and they tend to thrive in reputation economies by attrition: everyone else is out working and (inadvertently) stepping on toes, while they’re smiling and making nice. These people make great figureheads and can be tapped to lead more established companies, but they’re terrible small-company founders. Those people are risk-minimizers, not excellence-maximizers. If you’re trying to launch a high-risk business, you require excellence-maximizers, but it’s impossible to find one who can thrive in a reputation economy. In a reputation economy, the first rule is never to overperform, but who wants that kind of person at the helm of a high-risk business?

Fixing the problem

There is a way to fix VC-istan’s poor performance, and it’s tied to an issue I’ll keep coming back to: collusion. Namely, before funding decisions are made, the VCs talk to each other and decide, as a group, who they like and who they don’t. This means that the funded people are the ones selected by a committee rather than individuals, giving the advantage to the reliable, affable risk-minimizers who are, to put it plainly, unsuited to lead high-risk technical enterprises.

Given that the flaws of design and selection by committee are well-known, why do venture capitalists collude? Shouldn’t there be a rogue investor or few out there who realizes the flaws of VC-istan, goes off and makes better choices, and shows up all of the established guys? Why doesn’t that happen?

The first thing to note is the distribution of payoff on investments. Stocks in established companies have a bell-shaped and approximately normal distribution in returns. Some will go up 20% in a given year, and others will drop by 20%, but for an established public company to change in value by a factor of 100 will virtually never happen. About half of the choices will outperform the market trend. The venture capital game doesn’t have the normal distribution, and it’s all about the outliers, or “black swans”, because far more than half of the investments are losers, but the winners hit big. These are the Facebooks and Googles of which two or three might exist per decade.

Being an early-stage investor in such a company makes an investor’s reputation and career in a way that a portfolio with more, smaller wins (even if the overall return were higher) wouldn’t. Becoming a partner in a venture capital firm nearly requires being part of a big-ticket item. It’s not about maximizing returns on capital to shareholders. It’s about the needs of the individual venture capitalists, if they want successful careers in that industry, to be involved in the two or three most important deals of the decade. A handful of smaller wins could be equally or more lucrative, but wouldn’t have the same career-making effects for individuals. Not surprisingly, this means they’re not interested in the small, less-risky, wins.

Because the venture capitalists need to know where the top few deals of the decade are coming from, there’s a premium on social access. Founders (much less engineers) are pretty much interchangeable. The talent-centric attitude that VC-funded companies feign is pure marketing, because the truth is that no one can reliably predict the biggest successes– not at that level of risk, and with such a long feedback cycle, it’s almost impossible for anyone to acquire the skill, data being already stale once collected. The best bet is to be near the blockbuster successes as they emerge, and be close enough to the gatekeepers to get involved. That is the source of the collusion.

In other words, VCs don’t collude because they’re exclusive snobs or bad people. They’re not. They collude because their careers mandate it. They don’t win by picking the best people or funding the most viable businesses and, in fact, operate at risk levels that make such predictions useless; rather, they win by being in access to the evident winners as they emerge. Those who can get into access may be able to escape the generally dismal returns that characterize the asset class, while those who remain out of access perish.

The alternative regime would be one in which VC collusion were banned outright. No comparing notes, no deciding as a group who gets to be a founder, none of this co-funding horseshit. It would represent a return to individual decision-making and beneficial (as opposed to degenerate) risk-taking. A legal ban on collusion would make venture capital fundamentally more decent and, quite likely, substantially more profitable on the whole (if disadvantageous to a few prestige investors)  on account of improved prospect selection. It may sound like a drastic move, but a collusion ban just might be the only thing that can save venture capital from itself.