In the future, Google may value the accuracy of your content more than the quality of your backlinks, according to a paper (PDF) recently published by researchers within the company.
New Scientist reports that Google is working on a system where it can determine the trustworthiness of a page not by who is linking to it, or how many incoming links it has, but by the number of facts it contains.
A score, called a Knowledge-Based Trust score, would be computed for each page by cross referencing the content with facts stored in Google’s Knowledge Vault. The Knowledge Vault is a database of 2.8 billion facts extracted from the web, and is the primary source of information behind theboxes that appear on the right side of some searches.
The more facts contained on a page, the better it will rank. In instances where few facts are found on a page, Google will check the accuracy of other content contained on the site to determine how well it can be trusted overall.
In early tests, the research team says the Knowledge-Based Trust score has been able to reliably predict the trustworthiness of millions of websites. This sounds impressive on paper, and I’m sure the SEO community would appreciate an alternative to links as a ranking signal, but this concept leaves me with a lot of questions.
For example, not every website exists to report facts, so how will trustworthiness be determined in those cases? Well that’s when the research paper says Knowledge-Based Trust isn’t necessarily a replacement for current ranking signals, but a supplement to them.
I’m also concerned about pages written around new technology and new discoveries, with information that hasn’t yet been entered into Google’s Knowledge Graph. If Google started to rely on Knowledge-Based trust to rank web pages, would it then focus additional effort on revising and updating the Knowledge Graph?
That question, and many others, aren’t answered in the report — but I suspect more information will surface as Google continues its testing.
Freelance Writer at MattSouthern.com Search Engine Journal