I am sure all of you read the NY Times piece from the weekend named The Dirty Little Secrets of Search. It talks about how major realtor, J.C. Penney, had unbelievable Google ranking for many keywords that brought them tons and tons of e-commerce income for several months - but how those rankings were achieved through link spamming.
Of course, J.C. Penney management claims they had no clue what their outsourced SEO company was doing and fired them immediately after speaking with the NY Times. But all those months, they had number one rankings for keywords like “skinny jeans,” “home decor,” “comforter sets,” “furniture," "dresses," and the list goes on and on. They reaped the short term reward and now find themselves publicly humiliated and penalized in Google.
If you want to learn more of exactly what J.C. Penney's SEO team did on a technical level, read Vanessa Fox's explanation at Search Engine Land.
There is, as you would imagine, tons of discussion around this in the SEO community.
Here is one quote from WebmasterWorld from Tedster, but the forums are definitely worth a read:
I thought it was interesting that the algorithmic change is described as taking so long to roll out. We have been noticing for months how big algo changes first show up and then seem to bounce around for days. This article seems to point in the same direction. I'm thinking statistical sampling and feedback - machine learning of some kind.
It seems that "how we count backlinks" has been a common description of algo changes for quite a while now. Makes me wonder if Google's address to content farms will also be in this area. And also if a site is tagged as hosting a low value outbound link, might that roll over to other outbound links they host.
I know that Google can negate link equity flow at the level of individual link, the whole page or the whole domain. But I wonder if enough "individual problem links" could spell bigger trouble for the hosting site.