Don't take my word for it. Here are the stats:
While this is an early market, it's clear that even in this FB-hyped up era, the per user value is nowhere close to what a user in a standalone (free?) app on the WWW gets:
So to sum up what we have so far, if you put the highest dollar per Facebook-app user figures next to the lowest dollar per WWW-app user, you still have Facebook-app users going for a 90% discount to regular users.
Looking at these numbers, it doesn't make a lot of sense for a lot of people with "great" ideas to be laying down their bets on Facebook's poker table.
Why would this make sense? There's always a flip-side, right? Well, despite these numbers there reasons to create Facebook applications. First of all, at scale, low $/user figures don't mean anything at all. If you can crank out app after app and get 10X more users than a successful WWW app, then all the more power to you (hello, Slide and RockYou!). O'Reilly showed that these are not great expectations.
As a smaller app builder, there are still some benefits, but only if you actually create technological value for Facebook or other major app providers. Let's say Brooklyn's Kaltura proves that their Facebook video app is sooo leaps and bounds better than Facebook's own app, from a technological perspective, that Facebook acquires it for its technological value rather than someone else acquiring it for the user value. This is an obvious win for Kaltura and proof that in some cases it pays to build a single app for Facebook.
Otherwise, it's hard for me to imagine your life is 10X (or more!) easier building an app on Facebook. Sure, the immediacy of getting 10,000 users feels awesome, but those users are "worth" less than 1,000 in the "real world."
And if you have to value your users 10X less on Facebook, that means they're going to value you 10X less as well. Go have fun.