There are actually three scenarios:
1. G tilts sharply to the small sites to force the big brands to spend their $100+ million budgets on Adwords. The big risk with this is that an anti-trust suit is filed, as those big players can also afford fancy lawyers
2. G tilts to the big brands to force the small sites to buy Adwords. the risk is that the small guys have very small budgets and often no advertising budget at all, and by trying to get revenue only from the small guy G goes bust.
3. Simply rank the SERPs based on quality - sometimes the small guy will win and the big brands will be forced to buy Adwords (nice profit for G), sometimes the brands win and the small guy is forced out (no profit for G, but at least they can point to these types of SERPs to protect them from an anti-trust suit).
P.S. In my Smashwords-Kindle example, I think G was going for trust and quality - the Kindle store has a mess of rubbish, whereas Smashwords stringently approves works to check for originality as they then redistribute to the Apple bookstore and Barnes and Noble (who are also picky). However, if a listing then gets overwhelmed by links, G must assume that it must be OK (and given Amazon only send links to books that do well - i.e. have customer support this is rational.
So I think what is going on is they assess and weight on an initial trust factor, and then let backlinks do their job.