Google is putting on the gloves for another round against offensive content, according to Search Engine Land. The strategy is to “teach” its search algorithms the difference between good content versus bad or offensive remarks, and factual information versus false facts. The goal is to keep the wrong material off the front SERP.
This effort starts with the 10,000+ contracted quality raters who check search results daily based upon informational cues provided by their employer, Google. These raters check the quality of top search results, based upon a 200-page guideline book. Based on rater’s opinions, the search algorithms change to be much less favorable toward the indicated search result.
The newest rating tool recently added to the rating guidebook includes a new addendum on “Offensive/Upsetting” content. Such content falling under these designations includes anything that promotes hate or violence against people based on ethnicity, religion, gender and other such differentiating characteristics. Racial slurs are a definite cast off, as are graphic violence, animal cruelty or child abuse. Detailed info about “how-to” engage in harmful actions.
Any rating demotions in the search algorithms are generalized and do not affect direct requests. For instance, if a searcher asks for information on a neo-Nazis website (for academic research or other reasons) the engine will provide his information.
Google has been experimenting with these updates since December, giving only a fraction of their raters the new guidelines. The results showed a small percentage of success. Now that the new plan is fully underway, searchers can expect improvement in the SERP content.
