strangely, it seems to make a difference whether the ads send traffic off facebook or they stay on it.
When we run ads sending traffic off FB we have seen high levels of clicks which never arrive at the site, or arrive looking like bots.
When we run ads targeting people for likes, or for interactions on the facebook page, the increase in likes/interactions does seem to be roughly around what we're paying for and we keep an eye on the quality of the people we get - although you sometimes get the usual facebook randoms who seem to do nothing but play farmville and answer questions, they look to be real humans not bots - if you were going to fake a life you would NOT fake some of these..... and at least the standard percentage come back and interact later.
I do track every action from FB to site, and yes if you want comp entries or to drive traffic to an external site via your timeline these can be high enough to see a decent ROI imho, although not as good as ppc for us (but a hell of a lot cheaper). What I can't prove uis many direct sales from ad clickers. Since we now send almost everything via the timeline I'll never be able to pin that down - I can see/confirm actions from ads, and I can track online sales and attribution from FB, but unless they open up some extra functionality/stats you can't ever fill the middle gap I guess.
Very confusing. Can't quite get my head around why there would be such a difference.