The accuracy of a site’s information, rather than its popularity, may one day help determine how it ranks.
That’s according to a paper by research scientists at Google, reports the New Scientist. The paper, Knowledge-Based Trust, presents a new and different way of assessing the quality of Web pages by gauging their accuracy rather than previous ranking factors.
Typically, sites have been ranked on “exogenous” signals, such as hyperlinks, according to the Google scientists. But they have suggested using “endogenous” signals, or the veracity of the facts on a page, for ranking. The fewer false facts on a page, the more trustworthy the page is deemed and the higher it is ranked.
After building an extensive database of known facts, Google introduced its Knowledge Graph in 2012. According to the paper, a Google search could check the facts that appear on Web pages against known facts stored in Google’s knowledge base. Web pages that check out with a high degree of accuracy would earn higher rankings, while pages with less veracity would sink lower.
Although the scientists say this is a promising approach, they concede that it wouldn’t work across the Internet. Many pages don’t attempt to share facts or contain information that’s in the Google knowledge base, for instance.
For pages that don’t have enough of their own facts for checking, the scientists said that evaluating additional pages from the same websites could help assess their trustworthiness.
Although preliminary tests of knowledge-based trust have been promising, according to the scientists, the process might work best as a way to augment and improve current search practices rather than totally replacing them. It might be used “in conjunction with existing signals such as PageRank,” they say.
For more on this topic, get the full story here.