This certainly is a direction of study that seems to be suggested. If we do understand what Google is doing, wouldn't the first challenge be to understand the type of taxonomies that Google has established to test various results against?
I see one significant challenge. There's no reason to think that every site gets this "traffic shaping" treatment we've been discussing. Google seems to roll it out only in certain cases - perhaps where they're having trouble "pegging" a website within one of their taxonomies.