Sin, Secret, Series A

“How many secrets are there in the world? Recall that, reframed in a business context, the key question is: what great company is no one starting? If there are many possible answers, it means that there are many great companies that could be created. If there are no good answers,

“How many secrets are there in the world? Recall that, reframed in a business context, the key question is: what great company is no one starting? If there are many possible answers, it means that there are many great companies that could be created. If there are no good answers, it’s probably a very bad idea to start a company. From this perspective, the question of how many secrets exist in our world is roughly equivalent to how many startups people should start.”

“Do things that don’t scale.”

“Social networks do best when they tap into one of the seven deadly sins.”

There is a correlation between good companies and dirty secrets. Perhaps there’s a reason.

This pattern is striking. For company after company, especially in the social space, there’s some dirty secret — but in every case, it’s something the company did, not something they do.

A CEO poses next to his growth marketing strategy (left: six months after founding; right: month before IPO)

It may be the case that these companies are the exceptions, and that everyone else kept their noses clean. And there are dubious members of this club: did Airbnb spam Craigslist to get users? They say it was an overzealous contractor, and we’ll never know for sure. The founders seem like upstanding people. On the other hand, they seem like very driven people, and it’s entirely possible that they did some aggressive growth marketing in the early days when they were desperate for users.

This may be one of those behaviors Paul Graham is talking about when he says to do things that don’t scale. When a big company commits a little mistake, the cost of that mistake is multiplied by the value of their brand. But if you’re a nobody and you do something bad, nobody cares. As companies grow, they scrub their reputation. This process tends to accelerate once they start preparing the S-1, when every startup seems to hire a PR person whose full-time job is to sand their CEO’s personality down into the most boring conceivable version.

If this is the case — if nearly every company does something dubious early on, and then grows out of it — it gives us two useful conclusions. First: be very reluctant to condemn early-stage companies for their corner-cutting; it probably doesn’t reflect what the company will grow into. But second, if you’re at an early-stage company, pay very close attention to whether this kind of behavior is getting more or less common over time.[2]

The Other Exceptions

The products I’ve mentioned have something else in common: they’re all consumer-facing. And in a way, this is consistent with consumer products generally. A cynic would not be far off the mark in saying that every single consumer brand name is based on a lie. It’s usually a white lie — the clothes don’t really make you more attractive, the beer doesn’t make you popular, and putting Sapiens on your shelf won’t turn you an intellectual.[3] Fast food companies ostensibly sell food, but as David Perell and Chris Arnade have independently discovered, they fulfill the same role as churches: a focal point for the community and a universally safe and welcoming space.

Some consumer products are marketed as a way to better market yourself to others; some products are a way to market yourself to yourself (any upfront monetary commitment to a better lifestyle — a gym membership you won’t use, organic ingredients you won’t cook — fits into the self-deception framework).

Importantly, the lie is usually directionally true, and sometimes it’s coincidentally right. Clothes don’t make you popular, but they do signal membership of an ingroup, and if you’re a fit for that group the signal might turn out to be a valid one. In other words, it’s a lie that becomes a Schelling Point, functioning in the same way as an ideology. Monarchs don’t truly have a divine mandate, and voting doesn’t genuinely give you any political freedom, but believing in those systems coordinates behavior towards prosocial aims, and once you’re united in your shared belief in a falsehood, anyone who tells the truth is easy to identify as an enemy.

(If only Carl Schmitt had served in the Cola Wars.)

A social media site might turn out to be the reductio ad absurdum of the brand-as-lie/lie-as-Schelling-Point phenomenon, since the entire point of user interaction on the site is to make the lie true. If a site markets itself as the place where a certain kind of cool person hangs out, and says it boldly enough to the right audience, it becomes exactly that.[4]

At the opposite end of the spectrum is the kind of company whose success is a physical rather than a social phenomenon, where something is true or false, rather than in some weird superposition. Elizabeth Holmes and Bernie Madoff trafficked in secrets and lies, too.

And it’s surprisingly hard to find stories of B2B companies, hard science companies, biotech, etc., with charming stories about founders gulling early investors and fibbing to users. The worst you get is companies claiming to automate something when the founder is actually doing it by hand.

But the power of exaggeration is still powerful for these companies. They just have to exaggerate the effect of which their product is a cause.

Consider the Segway. It was going to be “maybe bigger than the Internet,” it was going to revolutionize transportation and reshape cities, and it — the Segway was literally code-named “It” — kind of flopped. The Segway sold okay. You still see them around. But cities look pretty much the same. Hard to say why. The technology worked, high-profile disasters notwithstanding. Maybe Paul Graham’s dork theory explains it. Regardless, it’s a cool invention, and the people who worked on it could have worked on something else. But there was really no other project that required the same skills and had the same level of hype.

Or look at Juicero. The plan was to save the American diet. The product was an over-engineered $700 juice press. (With WiFi.) There were probably other things those engineers could have worked on, but everybody else was making kitchen appliances, and Juicero was trying to fix obesity. A vision like that isn’t just inspiring. It’s a way to convince early employees and investors to ignore prosaic risks. Appliance companies worry about outsourcing. People on a mission to solve a pressing social problem don’t ever wonder if there’s somebody in Shenzhen planning to solve that social issue for 30% less money.

While Segway and Juicero exaggerated their prospects, it’s hard to say they lied. In fact, they delivered a product that did exactly what it said it would do. It just turned out not to be something their target customers needed done.

Importantly, this doesn’t mean that startups interacting with the real world can’t have secrets. They have to, or they’ll have competition. The difference in secrets between consumer startups and other startups is that consumer startups lie by commission — they make themselves appear to be what they’re not. Other companies lie by omission: they don’t want their prospective competitors to know what they’ve figured out until it’s too late.

Education and Healthcare: Big Lies

The promise of the US higher education system is that whoever you are and wherever you came from, you can get your union card to join the middle class. The healthcare system tells you that however sick you are, however much pain you’re in, you can get more time if you spend more money. And both systems have a combination of indirect payments and a subsidized, tightly-regulated financial system — we prepay for our healthcare consumption through Medicare and health insurance, but defer paying for education through student loans and paying income taxes on our post-degree earnings.

Fear of isolation, fear of death, and a payment system that strictly separates the person who lives with the outcome from the person writing the check. No wonder they’re eating the entire economy.

Both sectors do something useful but market themselves through a shared lie: in education, the lie is that we spend on schooling because we want to buy education; in healthcare, that we spend on drugs and surgery when we’re trying to buy additional healthy time. Spending on schooling eventually becomes an expensive form of procrastination, while spending on end-of-life healthcare is just paying five-star-hotel rates for a sub-POW quality of life.

And in both cases the magnitude of the lie is compounding faster than the value they produce.

For healthcare, the fundamental issues are:

  1. The pace of advances has slowed down dramatically in the last two generations, possibly because we’ve plucked the low-hanging fruit and possible because regulatory caution makes breakthroughs both riskier and more time-consuming.
  2. There is almost no limit to how much healthcare people are willing to consume on the margin, and as countries get richer, their marginal propensity to spend on healthcare rises. In fact, this arguably explains that the US is not an outlier in terms of healthcare spending, just an outlier in terms of consumption.

Adjusting for the confounders, the story of high healthcare spending is a story of Americans getting the best healthcare money can buy, and the “money” side of the equation growing faster than the “best” part.

Education is arguably a ticket to membership in the middle class, and elite education gives you a pretty good shot at joining the elite. But a lot of this is confounded; colleges get tuition, donations, and government aid because of the correlation between going to college and being successful, but that correlation is largely driven by who gets admitted and who finishes their degrees.

The ubiquity of college degrees means employers can use them as a first-pass filter, and as long as smart people still choose college over other options, that filter will work. But it imposes a negative externality, by causing students to waste time and money on four (realistically: six) years of education (realistically: partying), for a credential (which they may or may not get) whose explanatory value is roughly half that of their acceptance letter.

If you’re entering one of these fields, the lie dynamic presents an opportunity: frame your business around a truth that the incumbents can’t acknowledge, rather than competing with them to build a better-branded lie. Lambda School figured out that education is a combination of consumption and signaling, and that getting rid of the campus and aligning school and student incentives could get rid of most of the cruft. Healthcare is still pretty broken, and most of the innovators are on a micro scale, although there are promising developments abroad.

Sins and Lies, not Liars and Sinners

I hope this framework is a sort of meta-cynicism: rather than looking at the world and seeing a hopelessly broken and incomprehensible system, there’s a modest and healthy level of moral imperfection. Peccadilloes and white lies, not unforgivable moral crimes and total frauds.

It’s hard to talk about this stuff without implicitly making a blanket endorsement of what is, from a virtue ethics perspective, bad behavior. But the utilitarian Law and Economics wonk in me says that the right amount of any kind of badness is close to zero, but never zero.

This moderate level of badness provides the activation energy for a lot of goodness, and nearly everyone feels at least embarrassed when their old bad behavior comes to light. And if you find a company that seems totally earnest, scrupulously honest, and radically transparent, try to figure out what they’re hiding. If the answer is “nothing,” it’s bound to fail.

Every good company needs a secret, but not everyone has a secret worth keeping. If a company doesn’t start with a secret, it seems that one option is to do something worth covering up.

Thanks to Alexey Guzey for editing, and to John Backus for compiling many, many examples. And apologies to the founders mentioned.