I’ve written at length about VC-istan, its poor performance and its bigotries. What, however, is VC-istan’s “original sin”? Why is it so dysfunctional? Is there a foundational reason for its pattern of across-the-board moral and financial failure? I think the answer is obviously, “yes”. There’s a simple root cause: it’s solving the wrong problem. This requires two investigations: what problem should venture capital be solving, and what is it actually doing?
What’s the purpose of venture capital?
This is an easy one. The purpose of venture capital is to finance endeavors that require substantial backing in an all-or-nothing transaction. A biotechnology firm that requires $100 million to develop, and put into clinical trial, a new drug or device would be one example of this. With $10 million, it produces nothing salable; with ten times that, it has a chance. Others exist around infrastructure and in more deeply industrial pursuits like clean energy. Venture capitalists do invest in these spaces, and that’s outside of what I’d call “VC-istan”. Not everything that venture capitalists do is ugly, of course, and not all of it is VC-istan.
Venture capital, in a way, was originally intended as the “capital of last resort” for high-risk, capital-intensive businesses that would never qualify for more traditional financing. Why? Because when the proper way to invest is all-or-nothing, that has (unavoidable) negative consequences for all sides. It means that most people won’t get funded, and it’ll be extremely competitive to get capital, and dilution of founder equity will be severe. It’s not ideal, but if all you have is an idea, your product is 3 years away from the market in the best-case scenario, and you’re asking for $125 million to get started, those are the terms you have to take. It is, of course, quite a noisy process. The best ideas might not get funded, because there is literally no one able to assess what the best ideas are.
Venture capital for biotechnology and infrastructure has its own rules and culture. I’m not an expert on that, but it’s not what I consider “VC-istan”. From my perspective, which may be limited, venture capitalists in that space are trying to act in good faith and invest in viable businesses. To be blunt, I don’t think the “cool kids” nonsense (see: TechCrunch) matters so much in those sectors, because the science has to be sound. If you’re trying to turn algae into diesel fuel, Mike Arrington’s half-baked opinion of you matters a billion times less than the chemistry inside your lab.
What problem is VC-istan solving?
VC-istan is a subset of “all venture capital”, and focused on the “hip” stuff that can be likened to “reality TV”. To explain this analogy, ask this question: why are “reality TV” shows so prolific? It’s not about their quality. Cheap production is often cited, but it’s not just amount the numbers in the accounting ledger. The reality show formula is one that admits perfect commoditization. Writers and actors, at high levels of talent, resist commoditization. They won’t work on shows that are bad for their careers, and they have agents whose full-time job is to represent their interests. This makes them non-fungible. At the highest level of talent, labor may be able to push back against commoditization of itself, because there are few enough people at the highest levels to make the market discrete rather than continuous– or, in other words, illiquid. Reality TV does away with those “prima donna” creatives and celebrities: the writing demands are minimal and can be fulfilled with a mediocre staff, and the actors are nonentities. This enables the production studio to iterate quickly with half-baked concepts without needing to concern itself with the career needs of the parties involved.
VC-istan loves social media, and consumer-web marketing experiments, which are like reality television in that they can be produced with mediocre, “commodity-grade” inputs. To launch a biotech firm, one actually needs to have a strong grounding in science. Assessing founders for scientific literacy is hard, and private equity people are rarely up to the task. But any idiot can come up with, and hire someone good enough to implement, a Snapchat or a Clinkle. In the soft space of marketing experiments using technology, as opposed to the much harder sector that is technology proper, commodity founders and engineers suffice, and because the space is a gigantic bike shed, every investor feels entitled to have strong opinions. If genuine technical talent is needed for “scaling” down the road, it can be hired once the company has been covered by TechCrunch and appears legitimate.
Ultimately, the purpose of VC-istan’s “tech” companies is not to innovate or to solve hard problems. It’s to flip teams that have been validated by three to six rounds of venture funding, and possibly by success in the market (but that’s optional). Occasionally there’s an IPO, but those are about as common as big-company spinoffs. More typical is the “acqui-hire”, whose purpose can only be understood in the broader context of corporate dysfunction.
M&A has replaced R&D
A company’s need for top talent tends to be intermittent or subtle, and most often both. An example of the first (intermittent need) is around a short-term crisis that only a small percentage of people will have the insight, creativity, experience, or work ethic that is necessary to surmount it. The second pertains to the long-term existential need for innovation; if the company doesn’t have some engine that produces an occasional positive-impact black swan, it will be torn to shreds by the bad kind: crises that no amount of talent or effort can resolve. While every company pays lip service to its need for top talent, the truth is that most companies don’t need top talent for their day-to-day operations. If they did, that would be irresponsible design: a dependency on something that is somewhere between a highly volatile commodity and an outright non-commodity. The need for top talent tends to be a long-term issue.
Top talent is difficult to truly employ; one merely sponsors it. Old-style corporations understood that and invested in R&D. When the rare crisis that was truly existential would emerge, talent could be borrowed from the R&D pool. Additionally, while R&D could focus on basic research that was of general benefit to the world, and not necessarily in the firm’s immediate, parochial interests, the proximity to that research the corporation enjoyed gave it such an edge in practical innovation to pay for itself several times over.
Unfortunately, basic research was one of the first casualties of the private equity invasion that began in the 1980s. The old R&D labs that built C, Unix, Smalltalk and the internet weren’t scrapped outright, but reduced to a fraction of their former size, and forced to take a next-quarter focus. Conditions weren’t actually made so bad as to flush existing talent out, but positions became scarce enough that new talent couldn’t get in. The executives of those companies weren’t all short-sighted idiots, though. They knew that the high-autonomy, R&D-oriented work was the only thing keeping top talent in place. With corporate R&D near obliteration, that was threatened. So they knew they needed a solution that talent-intake problem. What did private iniquity propose as a solution? More private equity.
Enter venture capital, formerly a subsector of private equity that was generally avoided by those with other career options, due to its infuriatingly intermittent performance. What would it mean, however, if venture capital could be made less “venture”, by filling a need created by the disintegration of another part of the economy? Companies shutting down the blue-sky, high-autonomy R&D work had to get talent somehow. Explicitly paying for it proved to be too expensive, except in investment banking, due to hedonic adaptation– people who are performing at a high level, if their needs for autonomy are not met, require 25-50% per year raises to be content. Tapping high-talent people for managerial ranks proved fruitless as well, because many of these people (while exceptional as individual contributors) had neither the desire nor the ability to manage (and, additionally, middle-management positions were also cut during the private equity invasion). The remaining solution to the talent problem became one that private equity men found extremely attractive, given the premium they collect on deals– companies must buy it.
I don’t intend to insult the low-level employees of the Googles and Yahoos of the world by saying that those companies have “no talent” at the bottom. That’s clearly untrue. Companies don’t acqui-hire (which is far more expensive than internal promotion) because they have no top talent in their ranks. They have plenty, but they acqui-hire because they have lost the ability to discover what they have. It’s a malfunction of the middle-management layer. These companies are like hoarders that buy new coats every winter not for a lack of coats, but because their houses are so out of order that a new purchase is preferable to sorting the old place out.
Moreover, a company cannot, in general, adequately commoditize its highest levels of talent. The best will always seek their own career goals foremost, and perform at their highest only when there is coherency between their long-term personal goals and the work assigned to them. There are also, to put it bluntly, not enough such people to merit any explicit managerial correction to this problem. An executive focused on the career-coherency issues coming out of the most talented 5% is ignoring the day-to-day work completed by the other 95%. Two (problematic) solutions end up emerging. The first is for the company to ignore the high-talent problem and treat its top 5% like everyone else: closed allocation, low autonomy, etc. Then it loses them, plain and simple, and becomes dysfunctional after a few years of brain drain. The second is to leave them alone and effectively let them work on whatever they want. That’s great, in the short term, but it can be politically messy; others (who may belong in the top 5%, but haven’t been recognized) may resent them for their higher level of autonomy, or that set of people may lose sight of their need to continually market themselves and justify their favorable conditions, and then be crushed (not for a lack of value to the organization, but because it fails to market itself) when there is a management or market change.
So what is the “problem” that VC-istan exists to solve? It’s there to commoditize top talent. Although a specific company cannot commoditize its top 5%, the conceit is that an army of dedicated specialists– a mix of investors, corporate biz-dev executives, and “tech press”– can do so. In the consumer web space, venture capitalists have become a sort of high-end headhunter, but one that follows different rules.
For one major difference between the old corporate ladder and the acqui-hire system, employers are not allowed to explicitly discriminate on age, pregnancy status, health issues, race or gender. Investors can. Middle managers are too busy to conduct invasive “back channel” reference checks that, in truth, constitute civil harassment and would admit blacklisting charges if they ever interfered with employment (thus, risk-averse companies prefer not to do so). Investors can do so, and in such a way as to work through people who will keep their secrets (preventing lawsuits). This is a wet dream of the new right wing, an Uberization of executive hiring. The old system, with decades of regulation thrown into it because those rules were actually necessary, has been supplanted by a premium, rule-breaking, and vicious new one. The people who need the regulations imposed by the old system (i.e. women, minorities, people with health problems, people over 40, people with kids) are simply judged unfit to compete.
Here’s a question: how well is VC-istan actually doing, on its own terms? The first question is: what does it mean to “commoditize top talent”? While that sounds like something I might be against, I can’t actually say it’s a bad thing– not even for top-talent people. When something is commoditized, a fair price (that may fluctuate, but is fair relative to published market conditions) is established and it’s very easy to buy or sell it near that price. Currently, the compensation for top (2.0-level) engineering talent swings between about $75,000 and $10+ million per year– there is a clear uncertainty about what it is worth– with a median around $150,000. If that level of talent were adequately and fairly commoditized, that range would be more like $300,000 to $500,000– which would give most of them a hefty pay bump. The truth about the commoditization of labor is that labor generally finds it unobjectionable when the terms are fair. In fact, one effect of labor unions is to explicitly commoditize labor while attempting to ensure fairness (while professions, in general, oppose the commoditization regardless of terms). The murky issue in technology is that “top talent” is very hard to detect, because the people with the requisite skill have better things to do. Those who can, do; those who can’t, evaluate others’ work.
VC-istan, then, is built on the record-company model. Founders and engineers are treated as commodities (and generally, for reasons I won’t get into here, don’t get fair terms) but there is a hope that, thanks to the law of large numbers, top talent will be detected and validated by the outside market.
Where VC-istan went wrong is that it never figured out what top talent might look like, so the resources were thrown behind those who were either best at self-promotion or (increasingly, over time) those who could pull inherited connections. As a mechanism for detecting the rising generation’s top marketing talent, it might not be doing so bad. For picking out the best technical talent, especially as pertains to long-term R&D, it’s worse than abysmal. It’s doubtful that it’s picking up any signal at all. Companies that have a genuine need for R&D talent will be poorly served if they source it through acqui-hires.
VC-istan exists to commoditize top talent, but it has also erected a feudalistic reputation economy in which investors hold the cards. Founders hold few, engineers hold none. The highest levels of technical talent have been rendered, by this new economy, effectively irrelevant, depriving it of any leverage whatsoever. So, the terms are made bad– so bad that top engineering talent is rarely delivered. Whether this will strip VC-istan of credibility in the long run is something that remains to be seen.
The point I’ve made here is that it’s “solving” an ugly problem in a bad way.
What can venture capital do for technology?
Venture capital’s purpose is to build companies that, if successful, will become massive corporate behemoths. On a fundamental level, it’s stuck in the 20th-century mentality where a gigantic organization is the only acceptable prize for winning. Startup life is sold (by founders, and rarely by investors directly) to talented, usually clueless, engineers as an antidote to the ills of “MegaCorp” when, in truth, the explicit purpose of the VC-funded startup is to become exactly that: a new MegaCorp, but usually with crappier health benefits, longer hours, and faster firing.
What the best engineers actually tend to want is high autonomy so they can deliver exceptional work. They’d prefer ownership over it, all else being equal, but as long as they’re fairly compensated, they’re generally happy whether they work for a 20,000-person company or for themselves. When corporate R&D was sold for parts, venture-funded startups were proposed as the solution, the new way forward. Don’t like what happened to your old job? Create a new job for yourself! The lie here is that founding a VC-funded company provides the autonomy associated with true ownership. In truth, venture capitalists become full owners (de facto, if not de jure, due to the power afforded them by VC’s feudal reputation economy) of the company even when they hold a minority stake. Working for VCs is not fundamentally better than working for a boss; in many ways, it’s worse because the social distance is greater. Most bosses don’t consider themselves inherently superior based on favorable birth;
There are several critical misses that have become evident as venture capital has attempted to replace more traditional venues for innovation. One is that it has proven not to be a valid replacement for internal R&D. Nothing that VC-istan has coughed up is anywhere near the order of magnitude of Bell Labs or Microsoft Research. The second is that it has failed to be an engine of small-business generation, which is necessary for economic growth. It hasn’t connected top talent with the autonomy that comes from ownership. Rather, it has abandoned top talent in the pursuit of commodity startups run by commodity founders and commodity engineers. Over time, one might hope top talent to abandon it. That trend seems to be emerging, but I have no idea when or how (or, even at this stage, if) it will mature.
There is a fundamental technical flaw with VC-istan, additionally. That I’ll focus on, because it might lead us in the direction of a solution. If we consider the risk/reward profile of businesses, we see an underserved middle of the spectrum. Low-risk businesses can take bank loans, but those require personal liability, so it’s not wise to use them for anything that might actually fail. High-risk gambits with above-90% chances of failure, but that are capable of returning 20-50x on success, are what VCs love. The mid-risk/mid-growth space– targeting 15 to 50% annual growth, with a low but nonzero chance of business failure– is inappropriate for bank loans (too risky) but unpalatable to venture capitalists (not risky enough). Unfortunately, I don’t see an easy fix for that. Venture capital could become very profitable by funding the 15-50% range, but investment decisions aren’t driven by profits so much as the career needs of the investors. Returning a steady profit (say, 25% per year, with a bit of variance) by investing in a number of solid but moderately-sized businesses is not career-making; having been in on Facebook (even as a minor and late investor) is. The name-dropping world of Sand Hill Road cannot be expected to change, and if it does not, the focus will be less on building quality businesses and more on taking insane risks in the hope of hitting a career-making blockbuster.
This is problematic because the mid-growth/mid-risk space is exactly where true technologists live. They do not become machine learning experts or compiler gurus in an overnight episode of “virality”, and whether Mike Arrington or Paul Graham owes them a favor is irrelevant to whether they can actually code. They get good (and, if possible, rich) slowly. In terms of abstract value-added capacity, 15 to 50% per year seems to be about the natural rate (although most engineers would be thrilled to have salary increases at even half that rate). Technologists are extremely good at delivering these 20 and 40 percent per year improvements. What lies outside their interest (and, usually, their ability) is engineering the social conditions that admit 100x “viral” growth (or, far more often, abysmal failure). It’s just not where they live; they weren’t born in casinos.
VC-istan is not about to die, any more than recording labels have ceased to exist. As a method of shaving 15 years off a rich kid’s corporate-ladder climb via “acqui-hire”, it will persist. As a machine that produces commodity startups run by commodity entrepreneurs, it will persist and probably be profitable for quite some time. As a way of enabling companies to discriminate on age, health, pregnancy status, and other illegal factors at upper levels (filled through acqui-hires, while rendering internal promotion rare) while keeping the discrimination off their books, it will hold that niche quite well. How relevant will VC-istan remain to true top talent? On that one, VC-istan’s lifespan may be limited. In that territory, it’s “ripe for disruption”.
So what shall be built to bring the disruption?