Have you ever been skydiving, bungee jumping, or something similar (even public speaking for many)? Unless you have a disposition of steel, when you're at the point of jumping, your gut is screaming at you to not do it, even if you're rationally aware of the (low) risk and want to do it.

Comparing launching something on to the Internet to jumping out of a plane seems silly on the surface, but there are some strong parallels. The more you do it, the easier it gets (though in both cases it rarely becomes a breeze).

A common trick is to pull back on some features so you can launch with the "minimum viable product". And even if you do that, yes, it might fail. But at least you'll know sooner and can stop dragging your heels. Build up some velocity and even if your product/site is going in the wrong direction, you can more easily change course than by looping around procrastinating.

I say all this as a semi-reformed procrastinator. My fix is merely to keep up the velocity, even if it's in an unknown direction (examples - at the weekend I built http://tweetcanal.com/ - today I launched http://javascriptweekly.com/ - both procrastination activities of sorts but they fit into my plan). Changing course is easier than starting the engine in the first place.