Blekko has been on a crusade to destroy spam. This started at the inception of the company, where the human curation of “slashtags” allowed pre-screened results and tossed a number of results into a “spam” category. However, human curation isn’t the only element of Blekko’s strategy. Earlier this year, Blekko banned 20 “known content farms” from its index. Some industry analysts were skeptical, echoing complaints such as “Just 20? By these standards, far more should have been removed.” Well, Blekko has gotten around to removing to a few more sites – 1.1 million of them, to be precise – from its index.
This wasn’t a manual banning process like the 20 sites removed earlier. Rather, Blekko has updated its core algorithm. They’re calling the new version of their algorithm “AdSpam,” because it looks at the interaction of content and advertisements to determine whether a site is legitimate.
There are two parts of the algorithm’s strategy. The first is to examine the source of advertisements on the page, looking out for “self service advertising networks” that are popular amongst spammers. Second, if the user is utilizing these self-service networks, Blekko checks the correlation of keywords on the page with the advertisements. Sites that seem to be spamming to drive traffic for that ad content will be knocked out of Blekko’s index entirely.
With 1.1 million sites, there’s little doubt that Blekko’s net is being cast wide. Still, Rich Skrenta’s statements indicate that Blekko will only be looking for ad spam on sites that already seem to show low-quality content, based on the markers the company has previously determined.
[via Tech Crunch]