Wrong But Coherent

One of the best business parables goes like this: A man walks into a bar, and he tells the bartender “I bet you $10 I can bite my eye.” The bartender says “You’re on.” He pops out his glass eye, puts it in his mouth, and bites down. The

One of the best business parables goes like this:

A man walks into a bar, and he tells the bartender “I bet you $10 I can bite my eye.”

The bartender says “You’re on.”

He pops out his glass eye, puts it in his mouth, and bites down.

The bartender slides $10 across the bar.

“I bet you $20 I can bite my other eye.” The bartender saw him walk in, so he knows the guy’s not blind. He agrees to it.

The man removes his false teeth, and uses them to bite his other eye.

$20 changes hands.

“Okay, one more. I’ll bet you $100 I can piss into an empty beer bottle and not spill a single drop.” The bartender is a savvy man, so he has his doubts, but his doubts about urinary precision overcome his doubts about his ability to spot a con, so he agrees.

And he’s right! The customer stands up on the bar and — makes a colossal mess, soaking the bar and the bartender alike. He hands over $100 to the grinning bartender. But he’s smiling, too.

“What’s so funny?”

The customer points down the bar “I bet that guy $500 I could piss all over your bar and you’d be happy about it.”

Obviously this is one of the greatest jokes of all time, but I think about it a lot in another context: it pays a lot to know what problem you’re solving, and it’s hard to guess in advance.

Among startups, the “pivot” goes in and out of fashion — for a while, it seemed like every great company had radically changed either what they did or how they did it. Then there was a period starting about half a decade ago when people got more skeptical. Marc Andreessen:“When I was a founder, when I first started out, we didn’t have the word pivot, right? We didn’t have a fancy word for it. We just called it a fuck up.”

That’s completely true. If you didn’t mess up the first iteration of your model, you wouldn’t have to pivot, so in a sense a pivot is a failure. The real problem is a problem of terminology: there are two distinct kinds of pivots.

Some pivots are random (“We still have money in the bank, we all own computers, and we like each other. So, let’s start a tech company! Again!”). These have a deservedly bad rap. If a team gives up and does something completely new, there’s a good chance they’ll be just as wrong the second time. The graceful way to handle this is to do what Odeo did, and return capital to all the investors who didn’t want to take a flyer on a ridiculous idea like Twttr.

There’s another variant of the pivot, a sort of jiu-jitsu move where a startup translates its single biggest weakness into the core of its model. The problem with every startup is that it’s so delicate that any one problem can kill it — and if you’re doing something new, there will be lots of unexpected problems.

Tobias Lütke wanted to sell snowboarding equipment, but it turned out that the lack of good e-commerce software was nearly fatal. Stewart Butterfield wanted to build a cool game, but his team had so much trouble communicating internally that they ended up rolling their own instant messaging product. Genii ran into the same problem: it’s hard to know what everybody at a company is working on past a certain point, so they built Yammer.

In each case, it turned out that the million-dollar idea was adjacent to a billion-dollar opportunity.

A good question to ask is: why didn’t a bigger company solve the same problem internally? That’s one of the curses of bigness: you have enough resources that you can throw more resources at the problem. If a big company’s project is two months behind schedule because Product isn’t talking to Engineering often enough, the project launches two months late. If a startup is two months late on a launch and they have one month of cash in the bank, they just die.

As it turns out, the delicate nature of a startup turns them into a truffle pig for adjacent opportunities. Pivots that move one step up or down the supply chain are inordinately successful; it’s hard to think of many counterexamples. This model applies beyond tech, too; one of the best career pivots of all time was when Ray Kroc, a milkshake machine salesman, wondered how one hamburger joint could possible need to make that many milkshakes.

It happens more often with tech companies, because a tech startup starts with a coherent vision of the future: things will be different, but only if we make it happen. And what happens next is the rude awakening. The reason the future hasn’t happened yet is that some aspect of it is surprisingly hard to implement.

Having a coherent vision means saying “I believe A and B are true, which implies that C is possible — since C hasn’t happened yet, the first person to make it happen will have a monopoly.” Actually execution means learning that A is mismeasured and B is just false, but that A or B should be true. Most predictions about the future are false, because there are so many more ways to be wrong than right. Those mistakes happen because of ignorance, and a little experience cures a lot of ignorance.

So, to judge a pivot, you have to fiercely interrogate the idea that it was a “learning experience.” Did you learn that you’re a bad manager who can’t code to save your life? Maybe returning capital is the honorable choice; anyone who has been in the workforce for a long time knows that there are plenty of management jobs available to bad managers. But if you learned something specific, something nobody else knows, then you accidentally dug up a good idea.