Expert + Reasonable + Crazy Idea = Crazy Good

The amazing Paul “PG” Graham came out with an essay this month on crazy new ideas. And the thing I’ve learned over the years, being in Silicon Valley, is if PG writes, you read. In it, one section in particular stood out:

“Most implausible-sounding ideas are in fact bad and could be safely dismissed. But not when they’re proposed by reasonable domain experts. If the person proposing the idea is reasonable, then they know how implausible it sounds. And yet they’re proposing it anyway. That suggests they know something you don’t. And if they have deep domain expertise, that’s probably the source of it.

“Such ideas are not merely unsafe to dismiss, but disproportionately likely to be interesting.”

I’ve written a number of essays about crazy ideas. Here. Also here. The last of which you’ll need to Ctrl F “crazy”, if you don’t want to read through all of it. And also, most recently, here. But that’s besides the point. The common theme between all of these is that crazy ideas are not hard to come by. Crazy good ideas are. Good implies that you’re right when everyone else thinks you’re crazy. When you’re in the minority. And the smaller of the minority you are in, the greater the margin on the upside. Potential upside, to be fair.

As investors, we hear crazy pitches every so often. David Cowan at Bessemer even wrote a satire on it all. For the crazy pitches, go to episode five. The question is: How do we differentiate the crazy ideas from the crazy good ideas? But as PG says, if it’s coming from someone we know is a subject-matter expert (SME) and they’re usually grounded on logic and reasoning, then we spend time listening. Asking questions. And listening. ’Cause they most likely know something we don’t.

That was true for Brian Armstrong, who recently brought his company, Coinbase, public. He worked on fraud detection for Airbnb in its early days prior. And he knew he was getting into the deep end with crypto back in 2012. But he realized how unscalable crypto transactions were and how frustrated he was. Garry Tan, then at YC and part-time at Initialized, saw exactly that in him. A reasonable SME with a crazy idea. Garry just released an amazing interview between him and Brian too, if you want to tune into the full story.

What if some of the variables in the equation are missing?

But most of the time the founders you’re talking to aren’t subject-matter experts with deep domain expertise. Or at least, they haven’t left an online breadcrumb trail of whether they’re a thought leader or if they’re reasonable human beings. So subsequently, in the little time I have with founders in a first or second meeting, I look for proxies.

For proxies on domain expertise, I go back to first principles. What are the underlying assumptions you are making? Why are they true? How did you arrive at them? What are the growing trends (i.e. market, economic, social, tech, etc.) that have primed your startup to succeed in the market? Does timing work out?

To see if they’re “reasonable” under PG’s definition, I seek creative conflict. How do you disagree with people? If I brought in a contrarian opinion you don’t agree with, how do you enlighten me? How do you disagree with your co-founders?

In closing

To be fair, we’re not always right. In fact, we’re rarely right. On average, in a hypothetical portfolio of 10 startups, five to six go to zero. One to two break even. Another one to two make a 2–3x on investment. That is to say, they return to the investor $2–3 for every $1 invested. And hopefully, one, just one, kills it, and becomes that fund returner. Fund returner — what we call an investment that returns the whole fund and maybe more. Of course, every time a VC invests, they’re aiming for the fences every time. As a VC once told me, “it’s not about the batting average but the magnitude of the home runs you hit.” And even in those 10 investments, it’s a stretch to say that all of them are “crazy” ideas.

But the hope is that even if we’re wrong on the idea, we’re right on the people.

Written by David Zhou