tedster - 7:00 am on Jun 4, 2009 (gmt 0)

I used the word "algorithm" in a purely mathematical sense - not as a reference to Google's ranking/relevance algorithm.

What it looks like to me is a machine learning algorithm on its second pass. It looks for a certain footprint and automatically assigns a -50 tag on all the search results from that domain. There's even a -50 on the "example" part of "example.com", but no penalty on the "example.com" navigational search, which still even retains sitelinks.

In both these new cases from May 28, there has been no link buying for several years, pretty much from the beginning of Google's media blitz about paid links. So it's not that they are using the same agency or something, they are not buying any links, period.

But both have backlinks that are laid out in the page template in a way that certainly LOOKS like a "sponsored link" when you open the page. It seems some webmasters have an ad insertion functionality in their CMS and they used that to make a backlink "gift" to the site they liked. If there's a shared anomaly here, it's this type of backlink triggering a false positive - at least that's my current operating assumption.

----

I also see some junk sites being floated to the top. That change is Google's ranking algorithm at work. We've seen this happen several times in the past and the junk usually gets skimmed off in a short while. Several members have guessed that it's a kind of statistics-based machine learning. That idea sounds as good as any to me. As long as they get "skimmed off", it's not a long term problem.