1. Get big enough to build and scale an ad network and other services
2. Get bought by an existing scaled network
Option 2 is far easier which is why startups/VCs optimize for it. FB benefits by letting them do all the experimentation and then acquiring when they have proven traction, effectively gaining control of the new users for less money than competing outright, while the founders get a nice payday and move on.
What gets really interesting is when the same VCs behind FB are the ones kicking off these new experiments, a nice little money machine once you're in the loop.