(late to the party again am I?)
SPAM is precisely one of the problems with behavioral metrics in ranking systems. These signals have the potential for crippling spam.
Imagine that all those web spammers, currently auto-generating sites and auto-link spamming... started to hammer Google with query spam? The resulting cost on the system, as far as performance, would be staggering.
This is why, to me, Google is big on personalization. You can't spam yourself... well, you could... but what would be the point? One would have to think that any kind of ranking signal based on user behavior would be, at best, a weak signal in the regular SERPs...
BUT, the world of Google personalization is growing all the time and so understanding behavioral metrics and how they apply to rankings is becoming an important area of
SEO...
Was on about the Google Empire just yesterday;
The many faces of personalized search
As you can see, they really are a smart bunch and are building some strong data collection methods... which lead to knew angles for personalization of search...
And how about this funky angle on 'social search' from Redmond;
Alt Search Engines --- U Rank, oh yes U Do!
So, for me, peeps may rightly feel it is 'logical' for search engines to use these metrics (behavioral) but that doesn't mean they are....or at least not as definitively stated in the SEW post - this happens when the blinders are on to the downside of an approach. Many in the
SEO game are more interested in sounding professional, getting a competitive advantage or simply glory hounding, that considering ranking signals from a more subjective mindset is lacking.
I can say I have been recently reconsidering some of the approaches my company uses, but that is never ending really. I do believe that behavioral signals are important, just not sure how much and what resources should be allocated to nurturing them ultimately...