Paul Graham is a serial entrepreneur and angel investor. Paul and Robert Morris founded Viaweb in 1995, the first online web store builder, and sold it to Yahoo in 1998. In 2005 they started Y Combinator, a seed stage investment company.
I got to know Paul personally just recently, after first engaging in a blog debate. Paul wrote an essay "Microsoft is Dead" to which I replied "Since when does growing $4 Billion a year = dead?". It is funny, some of my best friendships have come as a result of arguments and debates. I became friends with Mark Cuban in the same way.
Y Combinator provides seed stage capital and a 10 week startup boot camp for budding entrepreneurs. At the end of the 10 week session Y Combinator invites VCs and angels to an investor day / demo day. I wrote about the YC demo day a few weeks ago.
Paul is a great guy and has built a huge following in the startup community. I have a lot of respect for him and what he has done. Y Combinator is a great story. Paul agreed to share some details with me. Here it is live and uncut.
When did you start Y Combinator and how did you come up with the idea?
We started YC in March 2005, after I gave a talk to undergrads at Harvard about how to start a startup. I realized then that it had been 7 years since we'd sold our company and we still hadn't done any angel investments, which seemed kind of lazy. So we set up YC as a vehicle for angel investing. Initially we were just going to do the occasional deal, like other investors. We were only going to work on this part time. But once I started thinking about the project I got sucked into it, and our plans became more ambitious.
When did you do the first session? How many companies?
The first was in the summer of 2005. There were 8 startups in that batch.
How did you recruit the companies for the first batch when YC was unknown?
YC was not so unknown as you might think. When we started it, a good fraction of hackers had heard of me or Robert.
How many applicants were there for the 1st session? How many applicants for the most recent session?
There were a huge number of applications the first time because initially we pitched YC as an alternative to a summer job for college students. We got 225 total. After that we started discouraging students from applying unless they really wanted to drop out of school, and the number of applications went down to 96 in the next cycle. It has grown steadily since. There were 435 for this summer.
How many sessions have you done? How many companies are in the YC alumni?
This summer's was the 5th batch. We've funded 58 companies total.
Reddit and Zenter are YC success stories. It is still early, but are there others?
Actually Loopt is probably the biggest success so far. Reddit and Zenter are famous because they got bought early, but there are a whole bunch of companies that could do better eventually. Six have had series A rounds from big VCs, for example. Several others have made it to profitability just on angel money.
What kinds of help does YC provide? Legal advice, equity structure, VC intros, hiring, market definition, business models?
This is a long one to answer. The best place to look is our site.
http://ycombinator.com/about.html
But the two biggest things we do are work with people on their ideas, and help them raise money from investors. We also do a bunch of little things, like getting them incorporated, introducing them to lawyers, arbitrating disputes, helping them hire people, etc.
Does YC provide infrastructure during the session? Office space, Servers, Networks, Telephone, etc?
No. We could easily afford to, but we think this would be bad for them. Saving people from dealing with legal documents is good, but working in our building or using our servers would make them start to feel like employees. I think that may be why "incubators" were such a bust in the 90s.
You hold sessions in San Francisco and Boston each year. Any differences in the setup, approach, or nature of companies?
The main difference between the Boston and Silicon Valley cycles is that the Boston startups present twice, once to Boston investors, and then again a week later at our Silicon Valley office. Otherwise things are much the same.
Are YC companies typically ready for VCs after 10 weeks or are angel investors a better choice?
Angels are a better choice for most young startups, whether they were funded by YC or not. But there are usually a few in each batch that are ready to raise full series A rounds.
VCs like to invest $3M to $5M, but most YC companies don't need that much. Have you found VCs and angels to fill the gap?
Actually a lot of VCs are starting to invest less. VCs will do whatever it takes if they like a deal enough, no matter what their stated policies are. But of course in addition there are a whole bunch of angels that invest in the startups, and these probably account for the majority of the deals.
Has YC had any liquidity events yet? How do you keep score on successful investments?
Four startups have been acquired so far. Reddit was a merger of two, plus Zenter, plus another one we still can't talk about. For us, success = the founders get rich, roughly.
What level of involvement does YC have after the three month session? Do companies stay in the area, mentor new companies, come back for advice, stay connected via email and social networks?
We keep doing everything we've been doing, except we don't have dinner every week. The startups help one another a lot, and obviously this continues too.
I assume you do this because you love building companies, have had success building companies yourself, and want to help new entrepreneurs. Do you think this is a viable investment model? Is 5% to 6% equity enough to compensate for the overall investment risk?
Actually we're doing it more to help the world than individual founders. We think the world would be immensely more productive if the best hackers started their own companies instead of marching off to work in cube farms. Think how much more Larry and Sergey did as startup founders than they would have done if they'd gone to work for a big company. Imagine that multiplied by a hundred,
or a thousand.
We want to make at least enough money that we don't have to stop. It would be nice to make more, but so far we have no idea whether this would be worth doing from a purely financial point of view. Classic VC funding is a well-understood model. What we're doing is very different. We have no idea if it will work.
