The New York Times has a great article about Yahoo's consistency versus Google's "wow" factor.
Saul Hansell reports that Google impresses with new products that go beyond baseline features, but it doesn't explicitly connect those products together. Meanwhile, Yahoo releases predictable products that may not be as cool, but which are all banded together to form a cohesive user experience.
I'm not sure that "cohesive user experience" matters so much. Google is more interested in the implicit connections users create while searching and using Google products. In other words, Google is trying to learn how its users search, click, research, consume, create, sell and buy. Then, after analyzing that data, Google will create products that conform to established patterns.
Not that Yahoo doesn't do that (Yahoo Answers is one way in which it does). But Google sees product creation as a math challenge, not a branding challenge. The Google engineers are interested in probabilistic systems and emergent behavior over a diverse set of data sets, both small and large. The fact that the data sets aren't unified under one home page is, in that view, pretty inconsequential.
The best example of this is Google Co-Op. As a stand-alone product, Co-Op is interesting. It allows you to personalize your search and benefit from the expertise of other searchers. That's cool. But what's cooler is the knowledge Google can gain from all those searches. Every time you drill down into results, you're showing Google what you need and how you need it. That, in turn, will allow Google to build a better product, for say, the health care industry.
I think this quote from The Long Tail explains the point well:
"The other thing that happens when consumers talk amongst themselves is that they discover that, collectively, their tastes are far more different than the marketing plans being fired at them suggest. Their interests splinter into narrower and narrower communities of affinity, going deeper and deeper into their chosen subject matter, as is always the case when like minds gather. Encouraged by the company, virtual or not, they explore the unknown together, venturing further from the beaten path."That said, I don't think Google has a master plan for banding all its products together. (Sergey admitted as much himself during the earnings call). We want to believe the company has a master plan, but that plan, like the systems it studies, is probably more emergent and less planned than we'd like to think.
But Google is building its services in such a way that will let it combine its products easily, should the need or financial incentive arise.