My Startup Manifesto

2009/01/25

This is my startup manifesto.

1. Professional and/or financial success

Gauge the software market and look for a need, a use-case that is not being fulfilled properly. Build something useful and try to market it (financial success). If your bussiness fails, just release your software as open-source. Build non-trivial things, and make sure the architecture and code is examplary, so even if financial success eludes you, your software is valuable on its own and earns the respect of your peers (professional success). Use technologies that will withstand the test of time.

Professional and financial success.

2. Compete where your strengths are

Figure out what your strengths are relative to your peers, and compete in that area. Too many people around the world can set up a LAMP stack and program in HTML/CSS/JS/PHP/Python/Ruby, so try to compete in an another area, where you have a competetive advantage. It's okay to use the mentioned technologies, but if all it takes to duplicate your product is knowledge of easy-access technologies, you're competing with millions of people, and your chances drop accordingly.

Feynman, according to himself, was an average bongo drummer.

3. Focus on fundamentals

Look for fundamental problems and propose fundamental solutions. Shoot as deep into the software stack as plausible/possible.

Rocky Balboa focusing on fundamentals.

4. Charge money

To make money off advertising, you need insane amounts of pageviews, which is very unlikely. To make money off a company that does not make money, you need to sell the whole thing. This is hard, especially if you are outside Silicon Valley / U.S.A. Combined with the previous points, the best bet is to write software that you can either sell or realistically charge money for support for, software which is important enough.

Don't be part of another bubble.

Example:

(Not my startup.) Write software to make data-based recommendations based on user-data (see earlier post). Many websites collect user data, but only few are any good at using them. There may be a bussiness use-case here (financial success). Write high-perfomance code that will withstand the test of time (don't depend on libraries that won't be around in 5 years). This is an area that is non-trivial enough to get into that you're not competing with all programmers who can write a for-loop. You can either give away the software for free and charge for support and development, or charge money for the software. (In the latter case, you can't count on the hordes of hobbyists to download your software and blog about it.)

- Marton Trencseni