Understanding Knowledge-Based Trust
What is knowledge-based trust?
Knowledge-based trust is the idea that an algorithm can establish the level of trust a website deserves based on its accuracy. In February 2015 Google published an academic paper that proposed using trust as a means of ranking websites and determining the ones that were most likely to provide users with the best answer.
What did Google discuss in the paper?
Google had its engineers crawl millions of web pages to pull out 2.8 billion trusted facts. These facts were then compared to the data that was contained on 119 million web pages. They used these facts to determine how much these websites could be trusted. When the engineers then went back and manually checked some of the results, they found that the algorithm was accurate at ranking the trustworthiness.
What could knowledge-based trust mean for the future of SEO?
Since this is not yet being used as a ranking factor, it is difficult to know exactly how it will impact websites, but the implications could be huge. For example, it can be used to help establish domain authority and determine the credibility of certain sites. Google already has a few sites, such as The New York Times that are maintained as trusted websites. This factor would allow Google to employ this type of list to millions of websites, which can then impact the rankings of countless sites.
It is possible that these developments would encourage the rise of niche sites that specialize in particular types of content. It could also potentially lead to certain types of sites self-censoring, such as medical journals becoming more reluctant to publish controversial research.
The implications of knowledge-based trust as a ranking factor would be far reaching and impact most websites. Google has indicated an interest in the subject, however, which means that brands need to be ready. Fact checking your publications and creating trustworthy articles should be a priority for everyone, even if the update has not yet been added to the algorithm.