Gmail creator Paul Buchheit: Build something 100 people love

And this is what we preach to the startups is like, figure out what you can launch today. And so to jump forward to Google, it was kind of 2001. There was sort of a big reorg where they got rid of all the managers and just gave engineers more of a free range engineering thing, gave us all projects. And so they were just like, I should build an email thing. Like, they knew I kind of had an interest in email, Robert. But it was, it was just like, we want to do something in email, we don't really know what exactly. Like, just go build it. So having learned from my prior experience, I decided to. I kind of know I have a short attention span. So the very first version of Gmail I wrote in a day and it wasn't very good. But I had previously been working on Google Groups, which was like a Usenet search at the time. Now it's something different. But I took the code that I'd been working on for that and then I just took all of my email and shoved it into that database, into that indexing engine. And then I kind of had an email search.

So your first thing was just an email you could search?

Yeah. And so then I emailed that out to the engineering team. I'm like, hey, I built this email search. Let me know what you think. And then people reply like, yeah, it's okay, but it would be better with my email instead of yours. So I'm like, all right, I have a feature request. So version 0 only searched my email. And then like the next version, I made it so it would search other people's email.

But there's a stunning lesson in there, I think, which is the point wasn't so much that whose email was being searched, it was that email was being searched. Like, you guys got to realize back then you couldn't really search email. All you could do is find the email you wanted and click on it. So searching like a thing.

Yeah, And I figured, I mean, a lot of our email is in common, so it sort of works, right? We're on a lot of the same mailing lists or whatever.

Everyone's cc'd.

So yeah, so then version two is just like, it indexes other people's email. So I send that out and then people are like, all right, this is cool. Now I want to be able to reply to one of the messages. All right, feature request, right? And so the whole thing was just iterating to step by step trying to make something that would make people happy. And so we actually had, as our kind of Like Launch Gate, this idea that we needed to have 100 happy users inside of Google because we hadn't launched it to the world, obviously. And so I actually embedded in the interface, this thing would pop up and be like, are you happy? Yes or no? It was just like a single question, are you happy with Gmail? Or it was at the time known as Caribou, yes or no? And it would get a list, then I would look at of who said yes and who said no. And then I'd go to all the nos and be like, what's it going to take to make you a happy user? Right? And actually just get them one at a time. And some of the people, they're like, it basically has to be a clone of Outlook. I'm like, well, you're never going to be a happy user. Actually, I did end up getting a lot.

You are dead to me.

I know. I did end up getting those people. And the way I would get them is that Outlook, at the time when you would hit 2 gigabytes in your mail file, it would corrupt the database and they would lose all of their mail. And then I had the only copy of their email. And so then they would become a Gmail user. Not always a happy user, but they would become a user anyway. Other people, it would just be like a minor thing where, like, I just need to add one feature or, like, fix a bug or something, and then they'll be happy. And so we kind of slogged through, like, just one user at a time until we got to like, 100 happy users.

Is that the genesis of, you know, what we call yc, the Bukheit Rule? Maybe you can just tell these guys what, like, it's better to have 10 or.

Right? So one of the ideas is, like, when you're starting out building something new, especially if you're kind of going into an established category like email, like, literally, you know, email was like 30 years old when we started on it, right? So there was a lot of history and a lot of opinions about how email should be. People would sometimes angrily tell me, I'm doing it wrong because we made the reply on top instead of the bottom, or weird stuff like that. There's all this history. It's pretty much impossible to enter a space like that and make a thing that appeals to everyone. And if you try to do that, what you end up making is just like a mediocre product that nobody really loves. And so my philosophy and what we try to get all the startups to do is figure out a thing that will just have really deep appeal, even if it's to, like, a tiny fraction of people. If you can make that small fraction of people just, like, obsessively love what you're doing, it's easier to then grow that group, because there's always people at the margin where if I just make something slightly better, they're going to join into that group. So it's easier to start with that deep but narrow appeal and then broaden it over time than it is to start with just sort of, like, broad meh, and then try to, like, convert people from meh to, like, loving your thing en masse.

Yeah, I've been thinking about that lately. Just in terms of retention. If you get 100 people who never leave, that's a great place to start. Then if you have thousands of people, all of whom churn eventually.

Yeah.