Could Facts Become Google Ranking Signal?

Google has recently published a paper that could look to drastically shake up the way they rank their index in the future. While the level of data handling in the paper is still somewhat hypothetical, Google is looking into using the validity of content on a webpage as a primary ranking signal. In basic terms, they want to rank sites based on how truthful they are, not just keywords, links and on-page optimisation. It’s a noble goal, as content credibility is one of the biggest issues online. However, when one considers the multiple reasons people go online besides research, one has to question the desirability of an algorithm that would place more value in objective truths over searcher relevance. To have a look at the full paper, click here. We warn you though, it’s not light reading.

Joining the Consensus

For a little perspective, Google initially became a market leader with its PageRank algorithm, that placed value on the external link network rather than the meta information and the keyword density of a site. This made ranking a question of authority ever since. In recent years, the increasing complexity of its algorithm allowed search ranking to shift away from PageRank and back to content, as well as 200 other ranking signals. But links still remain one of the foundations of its ranking practices. However all the SEO signals I’ve mentioned are external factors (even keywords), they relate to how people online interact with you or the search engine. What this will hope to achieve is make ranking equally as endogenous (internal), as Google puts it. Based on the perceived objective validity of the content, it will assign web resources a Knowledge Based Trust (KBT) score (not to be confused with TrustRank, which relates to historic data and SEO behaviour). What you may be asking yourself now though is, how exactly does an algorithm judge facts?

Likely Built on Knowledge Graph and Semantic Functionality

At this point Google has one of the largest databases in the world, so it’s got a lot to refer to when deciding what a factual error is. They’ll extract all the facts on any given site (likely using their Knowledge Graph extraction model that has an estimated 2.8 billion facts) and then, using what they call “joint inference in a novel multi-layer probabilistic model,” will determine if these facts are in fact, factual. The less errors there are, the higher the site will rank for trustworthiness. This isn’t entirely uncharted territory for Google. They’ve been creating automated synthetic data (data that’s a product of inference rather than measurement) for some time now, for instance:

• The Knowledge Graph answers your search queries directly, although there have been some problems with this graphing function.
• Semantic Search give you results based on synonyms to your search term and how you phrase the query.
• Google Health, a very recent addition, gives people highly accurate medical advice on the SERP based on the symptoms entered in the search query.

As you can see, Google is already able to determine the validity of facts, ensuring its accuracy, and basically understanding how language works, assuring the flexibility of this kind of function.

Trust vs. Popularity

It’s important to note that this is just a paper and merely speculates on a new primary ranking signal Google could explore in the future. SEO efforts are still safe, for now. It will also only form part of over 200 ranking signals, and it seems like it will only negatively affect those who rank due to their link structure. But this does beg the question, what are the priorities of people on the internet? On one hand, results like Yahoo Answers and popular forums rank very well due to their popularity. But the information from these resources is notoriously unreliable, but ranks due to traffic and people’s need for quick answers. But it could affect opinion sites and blogs- such as Thought Catalogue- that, while filled with subjective knowledge, have content that people genuinely value. But of course, we rarely search for opinions, (those are far too easy to find). And because Google is the biggest search engine, we sometimes confuse it with the informational democracy of the web itself. The fact is the internet is a shaky informational resource at best, which has always been part of its charm, but if there’s tool or function that can cut through all the spurious webpages and take you to the one that’s actually useful to you, there probably won’t be too many complaints.