Google asks users to help fight content farms

There’s been a lot of discussion about the quality of Google’s search results recently, particularly its ability to filter out “content farms” filled with low-quality articles that are written to appeal more to search engines than readers. Today, Googler Matt Cutts revealed a new weapon in the company’s attempt to block spammy content.

With a new extension for Google’s web browser Chrome, people will be able to block websites that they don’t want to see anymore, presumably because the content is always useless. (You can always click a button at the bottom of your search results to show what was blocked, and you can also edit your list of blocked sites, so these decisions aren’t permanent.) Google can then use the extension to collect data about which sites are being blocked, which it can feed back into its search results.

It seems like there are some big risks involved in letting the data shape general web search results. Google is famously secretive about the details of its ranking system, even when it announces improvements, because it doesn’t want companies to know how to game the system. By creating such an explicit way to influence results, Google might be opening the door to content farms that hire people to download the extension and manipulate the results.

Plus, there’s the inherent uncertainty about why someone is blocking a site. For example, there are a number of tech news sites that I might block because their content never interests me, but that doesn’t mean I’m telling Google that the content is spammy.

At least this system seems a little harder to game, since there’s no way for someone to say something should be ranked more highly. So spammers might be able to attack their rivals, but if a bunch of real users block their content, there’s no way to erase that. Plus, Google is still being a bit cagey about its exact strategy, saying “will study the resulting feedback and explore using it as a potential ranking signal for our search results.” So presumably it will find ways to avoid gaming and other irregularities, or it will just scrap the experiment if it doesn’t seem to be working.