I didn't actually go ahead and implement it - I just couldn't make the math work. In my opinion, there are too many unknowns with this model - you're quoting the price of a project without knowing all features, talking about design iterations, supported browsers, and a million other things. And if you substantially increase the price after getting the requirements, clients won't like that, because the most important number for them is the first number you tell them - in this case, $2,345.

I could go on, but I don't want to discourage you. You should try it, just be careful not to lose money. Also, if you're in the western world, make it $10,000 or $20,000, that's a more realistic number :)

That's usually how it works.