P16: a blog by Matt Kangas home archive
25 Apr 2008

A recipe for web-startup success

This week several big-name pundits spoke up about what, if you're running a web startup, one should do to prosper:

Fine, do those things. Buy a fancy espresso machine and cheap desks and nice chairs and 2nd monitors. Lots of distractions gone. Then what? What do you focus on?

Marc Andreessen has claimed, "the only thing that matters is getting to product/market fit". I agree -- but how do you achieve this?

You have an idea for a web startup. You want to build it, and you want it to be a big success. Where do you start?

I'm not as famous as these guys, but I've seen many web startups up-close. I've had my skin in the game. I even just had an old friend ask for my opinion about it this week.

So here's my answer.

One recipe for web-startup success

Take your awe-inspiring, world-changing idea. Apply Occam's Razor ruthlessly, stripping away every aspect that is not essential. Keep cutting until you arrive at a core value proposition for users -- the reason anyone will care. Now design the simplest possible website that actually delivers this promise. Describe the whole website as closed-circuit -- no undefined link destinations.

Pare it down again and again. Can you see the entire web site in your head yet? Every navigation element, every piece of content? Nope? Then keep cutting.

You want a complete, coherent design for the simplest-possible website that delivers on your value proposition.

When you can finally describe the whole damn product in a handful of page-sketches, and how every UI element should behave -- only then should you:

Why? Because you want to save money! And you want to get to product/market fit as quickly as possible!

Writing code takes time. If it's not your time, it's someone else's and you're paying the bill. Save time and money by writing less code. Debug less code by having less code. Get to market fast by knowing what to build. Verify your idea in the marketplace faster using a real implementation in a real timeframe.

Nearly every startup I've ever seen tries to avoid this process. It's easier to just start building something. Hire a team, let them figure out the details. Action feels good! The problem is, I've never seen this work -- not once. The missing pieces never fall into place on their own.

Plant the wrong seed, grow the wrong roots. Simple as that.

If you're having trouble getting to a radically stripped-down vision of the product, you could try:

But these are just "planning tools". They're only good helping you when you hit a dead-end mentally. They are not a substitute for the end result -- a radically actionable plan.

And... If you never get to the radically stripped-down plan that is actionable and would be valuable to users on day #1 -- maybe it's just not worth building. Seriously.