freejung - 2:48 pm on Nov 2, 2010 (gmt 0)

They have a lot of work to do... sometimes the best thing to do is to do nothing and let it sort itself out

In a way the second part follows logically from the first. It's pretty clear that Google is in the testing and development phases of significant changes in the way they handle user behavior, taxonomy, and profiling. There also seems to be pretty broad agreement that they haven't got it quite right yet, indeed that they may well be making things worse at the moment. However, that doesn't mean they're going to stop -- you don't get to work for Google by giving up easily.

So assume that there will continue to be major changes until they achieve a significantly higher level of quality. Therefore, any results you see in the short term from changes you make now may not be predictive of future outcomes.

Therefore, either do nothing, or better yet do things which are likely to be a good idea regardless of Google changes. Sgt_Kickaxe's suggestion of making changes that increase repeat visitors and lower bounce rate is a perfect example. That's always a good idea regardless of search engine behavior. The steps I outlined earlier for my situation mostly fall into that category as well.

If Google is performing the kind of multivariate testing we're talking about here, sending large groups of profiled users to different categories of site to test the outcomes, presumably they are doing this with the goal of improving the quality of results. Some might suggest that they are simply trying to maximize short-term profit, but I think they're smarter than that. To maintain their edge long-term, they have to take user satisfaction with search to another level. If that's the case, then as a webmaster you want to keep your eye on your long-term goals as well. What is your site _really_ about? How does it provide value to the searcher? How can you improve on that?