It helps to step back and take a look at the big picture. Companies have a certain number of advertising dollars. Companies that are really big in the consumer space spend large chunks of their revenue on advertising: consumer goods (soap, etc.), durables, autos, pharma, etc. This money is not going away anytime soon.
Now, if you're an advertiser, you want to advertise where there are people. This is why you see the MetLife blimp over big football games. The decals on the NASCAR cars. The funny ads in the SuperBowl.
For an advertiser, the holy grail is to achieve laser-like targeting: to be able to advertise to exactly the person who's most likely to buy his product. 80 years ago, the best option for advertisers was the radio. But it's a broadcast medium; you have no idea who's listening. So you make some intelligent guesses based on the program content and hope for the best. What's the ROI on radio ads?
20 years ago the best was the TV: the content was varied, so you could make better guesses. Want to appeal to the male 18-24 market? Advertise in football games, NASCAR events, WWF "Rumbles".
Then came the Internet. An advertiser could make even better guesses about the user; and more importantly, the advertiser could keep track of every impression, every click and every action. This is good, but still not good enough.
Broadly speaking: if you (advertiser) are willing to pay $X for a 10% chance of reaching the right audience, then chances are you'd be willing to pay $10*X for a 100% chance of reaching the right audience. Right?
And finally there's Facebook. The genius of FB is that the users have, on their own accord, given FB detailed information about themselves. Thanks to their social connections, they can't lie much either. I mean, it would be hard for me to claim I'm a female and single when my girlfriend is a connection on FB.
People keep thinking that FB will make money from the pageviews. IMHO, that is not the case. Yes, they have a bazillion pageviews and minutes spent online; they could make a billion or five per year easy with their pageviews.
The goldmine here is the "like" button. It will allow FB to spread its influence far and wide; and more importantly, it will allow FB to influence advertising on 3rd party sites. Instead of buying advertising on, say, CNN and paying $2/CPM, an advertiser will be able to hyper-target a user and advertise to that specific user on CNN (or wherever that user is). User + Content = More Value for Advertiser. So while CNN may get $2/CPM for its pages currently, by using the "FB Ad Network" they will be able to make more, and give a large cut to FB too.
Anyways: that's where I think the value lies with FB. Sorry about the long comment. :-)