An interview with investor Paul Graham of Y Combinator

Paul Graham believes we’re not in a bubble, that startups shouldn’t worry about their business models and the best companies are the ones with potential to kill old monopolies. Graham is a partner at Y Combinator, a Mountain View firm that invests in very early tech startups. Companies that Y Combinator has funded include the social news site Reddit, online form creation tool Wufoo (our coverage), and the online calendaring startup Kiko, whose controversial sale on eBay last month ended with a quarter million dollar winning bid. The next Y Combinator funded company to launch will be loopt, a mobile presence service also funded by Sequoia. Y Combinator’s site says, “We’re the right choice for a group of two or three young hackers who have an idea, and want some money and advice to get it launched.”

Graham is known for making strong statements about innovation and economics in technology. He’s the author of three books, most recently Hackers and Painters, published by O’Reilly. His essays can be found at The picture of Paul on the left is from his keynote talk at this year’s RailsConf.

In the following short interview, we talked about the growing trend of small investments in tech companies, Paul’s thoughts on startup business models, Kiko, the bubble and Web 2.0.

Marshall: One of the dilemmas that some startups face is that VCs only want to make large investments and small startups aren’t always interested in that – Digg most famously. Some people believe that’s changing though. Do you think that small investments in very early stage startups represent a growing trend, and if so why?

Paul: There’s definitely a trend toward smaller investments, because it costs so much less to start a startup now. And if you take less money initially, you keep more options open. Once you’ve taken a VC-scale investment of two or three million dollars, you give up the option of an early acquisition.

You’re also under more pressure to grow fast, which can cause you to make design errors. It may not be a coincidence that both Flickr and avoided the usual VC route. Both had to get a lot of subtle, social things just right. You’re more likely to do that if you can evolve organically.

You mentioned Digg as an example of a company that took VC money. They took $2.8 million, whereas reddit, their closest competitor, has taken only $88,000. (I know because Y Combinator funded them.) And yet reddit is not only able to compete, but has a visibly more authentic, participatory feeling. I read that the top 100 Digg users submit 56% of the frontpage stories. The frontpage stories on reddit are much more widely distributed. And that may be because reddit grew organically, through word of mouth, like Flickr and did.

Whatever the causes, it’s an interesting data point that a company with $88,000 in funding can even compete against one with $2.8 million. That could not have happened before the web.

Marshall: You’ve said that having a good business model is not important for startups because the good ones are liable to change models several times anyway. Yet many people believe that the absence of viable business models is one of the primary indications that we’re in a bubble. Do you disagree with that?

Paul: What I tell founders is not to sweat the business model too much at first. The most important task at first is to build something people want. If you don’t do that, it won’t matter how clever your business model is.

Of course you have to have a business model eventually. But experience so far suggests that figuring out how to make money from something popular is a lot easier than making something popular.

I get a lot of criticism for telling founders to focus first on making something great, instead of worrying about how to make money. And yet that is exactly what Google did. And Apple, for that matter. You’d think examples like that would be enough to convince people.

Is this another Bubble? I don’t think so, not so far. There may be a lot of lame startups being started, but that’s not the definition of a bubble. A bubble is when a lot of money is being invested in lame startups, and that’s not happening yet. The reason so many new startups are getting started is that the cost has gone down, not that funding has gone up.

Marshall: What kinds of companies or particular technologies are you most excited about right now?

Paul: There are so many. I especially like things that take advantage of the power of networks. As Sun used to say, the network is the computer. Except now it’s the Internet plus the phone network, not just your local LAN, and that changes everything.

Frankly, even though I’m supposed to be an investor, the ideas that excite me most are not necessarily the ones that make the most money, but the ones that blow away evil old monopolies. For example, I love collaborative news sites not so much because they make a lot of money– though they might– but because they’ve shown what a bad job the “old media” were doing.

Most people don’t understand what a social force startups can be. There are a lot of changes that can only happen through companies. One startup I dream of funding is the one that kills the record companies. You know your business model is broken when you’re suing your customers. The new business model must be out there somewhere, and my guess is that the way to beat the bad guys is not through political action (or at least, not only that), but by inventing
whatever replaces them.

Marshall: You were an investor in online calendar startup Kiko, whose founders recently sold the company on eBay. When the company was initially put up for auction you said that the release of Google Calendar and its integration with GMail was what nailed Kiko. You wrote that the best lesson for startups was probably to stay out of Google’s way and that the best way to do that was to not start something that Google employees would use at work – have I got that summarized right? I know I’ve given some more consideration to my initial statement that Google killed Kiko in light of the things made possible by the Google Calendar API. Do you think the game is over for startups making online productivity apps?

Paul: Yes, that’s a pretty accurate summary. I wouldn’t advise competing with Google in things they’re good at. So what is Google good at? As a first approximation, making things their own developers use at work. So they’ll do a better job on an online calendar than a video sharing site, for example, because their employees are probably not supposed to be sitting watching videos at work.

Another weakness of Google’s is that they’re a big company. I’ve heard some real horror stories from hackers there about the bureacratic obstacles to getting stuff released. That means they’re unlikely to do anything so novel that it frightens the bureaucrats.

So a startup could compete with Google if they had an idea so wild that it would freak out the internal gatekeepers, no matter what area it was in.

Unfortunately for the Kikos, there’s probably not as much room for wild ideas in calendars as in other applications. We all knew Google Calendar was coming, of course, but we hoped it would be the Orkut of calendars. Unfortunately it seems more to be the GMail of calendars, and it’s clear now why– because Google Calendar and GMail are the kinds of applications Google hackers use at work, and Orkut isn’t.

Marshall: What’s the new startup you funded from the Kiko team?

Paul: I can’t say yet, but it is certainly the most entertaining idea we’ve ever funded. In fact, insane might be a better word. But it’s what they wanted to do.

Marshall: What’s your take on the question of whether most Web 2.0 technologies only appeal to a relatively small niche? If you could share any of your thoughts on what you think helps a technology stick and gain a viable foothold in the market, I’m sure that would be much appreciated by our readers.

Paul: To me “web 2.0” translates to “web.” And web technologies don’t appeal only to a small niche. Web-based email services have hundreds of millions of users. The network (in the broader sense of the Internet plus the phone networks) pervades everything now.

We’re pretty open about what we think makes a technology stick. We print it on T-Shirts: “Make something people want.” If you had to reduce the recipe for a successful startup to four words, those would probably be the four.

The easiest way to make something people want is to make something you want. What do you wish existed that doesn’t? For example, back in the early 90s a friend of mine wrote some software for converting voice to data so he could talk to his girlfriend in Taiwan without paying for long-distance phone calls. That would have been a great project to turn into a startup.

If you’re a little older and/or have some particular domain expertise, you could try making something that you yourself don’t want, but you think other people do. This is a much riskier path, though. Most of the great startups seem to have begun with something the founders wanted: Google, Yahoo, Apple, even Microsoft.