Google Discloses How It Uses Ratings To Improve Search Quality

Google recently overhauled and released a copy of its Search Quality Raters guidelines. It's the first time Google has released the full 160-page version of the guidelines for search engine optimization. Some might argue the rating system is somewhat analogous to the Media Rating Council accredited services guidelines for viewable display and video ad impressions.

While Google's document supports those evaluating the quality of search results to give in return feedback on the company's experiments, it also gives SEO professionals insight into better understanding of how they should optimize Web site pages and content. The guidelines focus on the use of mobile.

Mimi Underwood, senior program manager, search growth and analysis at Google, writes in a blog post that the ratings do not actually influence search engine rankings, but they are used to help evaluators understand Google's experiments.

One of the more interesting rating criteria points to whether the content meets the needs of those searching for the information. The rating system covers everything from the most favorable, "Fully Meets" expectations, to "Fails to Meet," which means the query results that are unhelpful for zeros or very few people searching on the topic.

The Fails to Meet rating means the results completely fail to meet the intent of the user, such as a lack of attention to an important part of the query or user location that satisfies user intent. It is used to describe results that are extremely low quality, serve stale or outdated information, or nearly impossible to use on a mobile device.

Google also offers a section in the report detailing how to rate queries with multiple intents and interpretations such as apple the fruit or Apple the company, or Cougar the car vs. cougar the animal. There may numerous user intents for an iPhone query such as research iPhones, buy an iPhone, go to the iPhone page on Apple's Web site, but all users are basically referring to the same thing: the phone made by Apple. The different meanings are interpretations.

A helpful result for a dominant interpretation should be rated Highly Meets, because it is very helpful for many or most users. The rating system on queries ranges from a dominant interpretation of FullyM result, to one very unlikely that will provide a valid interpretation. Those are rated FailsM. 

Next story loading loading..