With the talk of GoogleLabs (gmail) last week, it was obvious that Google was launching products based on actual observed consumer behavior. Compared to Microsoft, they also seemed to have a much much smaller development team. These are the result of being a completely hosted service. There are no legacy software to manage, no partnership / software alliances that have to be supported. The web is the platform and the it is a liveweb.
It's possible to look at this another way: Google's very broad base of users give it the capability of sensing and seeing user behavior in near-realtime. Google can also track the rate of "launch success" against other factors it now has at its disposal: "do early adopter lessons cross over to the main stream for this type of application"?
While everyone has being looking at GooData as being "the best advertising information in the history of the world", perhaps the behavioral and profiling information coupled with expertise in discerning "intent", have made Google the most formidable "product development company" in history. After all we all buy products and services "to achieve an intent", be that latent or overtly expressed.
Now before everyone goes crazy on me, (all two of you) and says the obvious ("Google is a one trick pony, with a a dodgy hind leg, i.e. applications), let me posit that Google is learning how to use this information to build products, because to date, Google has not been a "product development company", but an "advertising platform". It will launch and fail, launch and fail again, but it will learn better than anyone else, how to fail successfully and how to incorporate the information into the next iteration. It won't approach this like an old school Proctor and Gamble, and I ultimately, I think they will make these kinds of "use-case-scenario-analysis" available as an API. Why not?