Long time, no wild conjecture

So here's the first of two posts based on purely wild conjecture. In a lengthening chain of trackbacks, Ryan Eby mentioned Christina's observation that Springlink has started displaying Google ads, presumably to supplement their subscription and pay-per-article income. Ryan goes on to wonder:

Will vendors continue with the subscription model or move towards a pay-per-view only model since they already have it for search engines in the future and can probably make more money? Will a search engine convince some of them to end the subscriptions and subsidize the change in order to try to pull more people into using their product instead of libraries? Will Google integrate Google Checkout with Scholar for easy shopping or perhaps personal subscriptions?

For a few months now, I've been wondering why Google hasn't made a move to purchase the Elseviers, Springers, and the rest of the ilk of the commercial research journal publishing world. If they could pull it off (and potentially avoid anti-trust measures by making the content available for "free"), I thought that ad revenue would be their public rationale for the move, but that the real reason for making that move would simply be keeping tabs on all of the research interests at both aggregated and granular levels. Think about the implications of being able to track all of the search patterns emanating from within a given company (quite often correlated with an individuals Gmail account, if they happen to have authenticated against Gmail during the same Web session). It would essentially amount to opt-in corporate espionage.

Or, for that matter, how about tracking the research interests of a given country. Iran, perhaps?

But now I've realized that Google doesn't even have to purchase the publishers. If the publishers accept Google Ads, then Google gets to track all of the searches or item views performed on that site. No purchase required.

Luckily for us, Google isn't evil, and their privacy policy doesn't leave too much room for selling that information; they can really only use it to develop new services. Hmm.