01 May 2006

Paul Graham's advice at Startup School

Paul Graham at Startup School
Originally uploaded by ptufts.
Paul Graham gave some advice to would-be startups this weekend at Startup School at Stanford. Here are my notes from his talk. I also have notes for the presentations by : Joshua Schachter (del.icio.us, memepool), Om Malik, and Caterina Fake (Flickr).

Here are my notes on Paul's talk:

1. Release early. While users hate bugs, they're ok with a minimal v 1.0 site if there's promise of more.

For example, Reddit has 500k users, and doesn't know who they are. It would have been silly for them to design for users that they don't yet know.

2. Pump out features. Don't just do this for its own sake -- keep coming out with things that make the user's life better.

3. Make users happy. Startups can't force anyone to do anything, they have to persuade. Design your site for casual users. "The median user will arrive at your site with their finger poised on the back button." You need to make them pause and look at your site.

The main job of your site is to convert casual visitors to users. And you need to measure growth to make sure you are doing this.

4. Fear the right things. Disasters are normal for startups: the founder quits, servers crash, you hit insoluable technical problems. These are normal, so don't fear them.

Similarly, don't fear large companies. Sure, Google or Microsoft could do what you're doing, but they aren't as determined to do this as you -- if they fail, they have lots of other options. If you fail, you're dead. So you are far more committed to what you're doing than they are.

What should you fear? Other startups. Startups are like cornered animals -- it's succeed or die. Fear them, not Google.

Other real problems: internal disputes, inertia, and ignoring users.

5. Commitment is a self-fulfilling prophecy. It's not about intelligence, it's about determination. Invest in grad students, not professors.

What motivates VCs is not the hope of returns, but the fear of missing out.

Commitment means people pay attention to you. If you're committed, VCs will worry that they will miss out on your successs, because they know you will be around. A company that isn't committed does not inspire these feelings in potential investors, because the investors can see that you are not likely to pursue your idea.

In short, committment means people pay attention to you, because they know you'll be around later.

6. There's always room. No matter how locked up an area seems, there is always room for improvement. Look at Google -- do we really think search 20 years from now will look like this, and that it will come from them?

7. Don't get your hopes up. Treat optimism like the core of a nuclear reactor: as a source of power, and also as something that can burn you.

Be optimistic about what you can do, but shield yourself by being pessimistic about machines and other people. Assume the deal won't happen. This is not to save you from being disappointed, it's to prevent you from leaning your company against something that could fall over.

Deals are usually dynamic. There are a lot of subsidiary questions to be cleared up after the handshake. If the other side senses weakness, they will be tempted to screw you over. "20k dollars, we thought that was 20k lira!" VCs are trained negotiators. While they're often nice guys, they can't help it. Don't even try to bluff them. Instead, when you hear "we want to invest in you" think "don't get your hopes up." This puts you in a stronger position without having to negotiate or bluff.


Startups are stressful. Economically, a startup is about working fast instead of working forever. Making money isn't grand, what's important is the speed -- getting the chore of a career done quickly "to show respect for life."

Update: Paul Graham has put his talk online.

No comments: