Google Developed New 'Algorithmic Solution' to Detect Merchant

Amit Singhal on Google blog responded to the recent New York Times article: "By treating your customers badly, one merchant told the paper, you can generate complaints and negative reviews that translate to more links to your site; which, in turn, make it more prominent in search engines. The main premise of the article was […]

Amit Singhal on Google blog responded to the recent New York Times article: "By treating your customers badly, one merchant told the paper, you can generate complaints and negative reviews that translate to more links to your site; which, in turn, make it more prominent in search engines. The main premise of the article was that being bad on the web can be good for business."

"In the last few days we developed an algorithmic solution which detects the merchant from the Times article along with hundreds of other merchants that, in our opinion, provide a extremely poor user experience. The algorithm we incorporated into our search rankings represents an initial solution to this issue, and Google users are now getting a better experience as a result," Singhal said.

Consider the obvious responses we could have tried to fix the problem:

  • Block the particular offender. That would be easy and might solve the immediate problem for that specific business, but it wouldn't solve larger issue in a general way. Our first reaction in search quality is to look for ways to solve problems algorithmically.
  • Use sentiment analysis to identify negative remarks and turn negative comments into negative votes. While this proposal initially sounds promising, it turns out to be based on a misconception. First off, the terrible merchant in the story wasn't really ranking because of links from customer complaint websites. In fact, many consumer community sites such as Get Satisfaction.com added a simple attribute called "rel=nofollow" -- that in general is a mechanism that allows websites to tell search engines not to give weight to specific links, and it's perfect for the situation when you want to link to a site without endorsing it. Ironically, some of the most reputable links to Decor My Eyes came from mainstream news websites such as NYT and Bloomberg. The Bloomberg article was about someone suing the company behind Decor My Eyes, but the language of the article was neutral, so sentiment analysis wouldn't have helped here either.

    As it turns out, Google has a world-class sentiment analysis system (Large-Scale Sentiment Analysis for News and Blogs). But if we demoted web pages that have negative comments against them, you mightn't be able to find information about many elected officials, not to mention a lot of important but controversial concepts. So far we've not found an effective way to significantly improve search using sentiment analysis. Of course, we'll continue trying.

  • Yet another option is to expose user reviews and ratings for various merchants alongside their results. Though still on the table, this would not demote poor quality merchants in our results and could still lead users to their websites.

[Source]