Dharmesh Shah on Everything He Knows About Startups

Just watched Dharmesh Shah's talk at the Business of Software 2008. He writes over at On Startups.



Interesting points:

• Building a reasonable ($30-40 million) business is perfectly respectable
• VC money is not a requirement
• If you're not embarrassed by your initial release, you've waited too long to release
  
• No one reads business plans. Write a blog instead, to get the idea out there and commented on as early as possible (don't worry about competitors finding out your idea; it's hard enough to get someone to fund you, never mind convince a competitor that your idea is better than theirs)
• Toy companies try out the advertising, then build the product. So sell pre-alpha versions of your product and figure out what people want
  
• Don't bother doing partnerships: they don't work
• Don't buy ad words: concentrate on getting a good page rank. That way "advertising" is free, and you can't be immediately displaced by someone paying a higher price
• Software as a service is great: but the margins are small. You now have to do all "tech. support". But you get detailed information as to which features your customers use, and what retains them
• Figure out what the impact of each feature will be on your customer retention/happiness
  
• VC syndication (where multiple VCs get together and fund you) is like price collusion. You need other players who will help to set the right price (a market of one means they set the price arbitrarily)
• Pricing your product is very hard. Pick a number, and then change on the basis of experience.