TL;DR

Summary of this article:

• Correct facts and sources within your content increase your trustworthiness therefor increasing trust rank
• Having only few facts does not affect trust unless those few are wrong but you lose the advantage if your competitors are more trustworthy
• The Paper published by Google suggests a better solution for data extraction that decreases and almost eliminates the errors of previous data extraction methods allowing it to show accurate results which are verified by human subjects afterwards with greater success rate.
• KBT (Knowledge based trust) method calculate Trustworthiness accurately 85% of the times even without any external signal “such as links” or PR.

Insights to use practically

• Best practice for acquiring high KBT – educational or informational 10x content with many facts that is based on large amounts of data that can be verified will provide the highest amount of trust rank.
• Content which is not gaining KBT – content that has few correct facts or facts that cannot be confirmed, Can be blog post of 1,000 words or less without any real value.
• Losing KBT and considered as low KBT – content with wrong or false facts.

Important to remember

These two are easy to mix up, “Buzz” doesn’t equal Trust, high link or share rate doesn’t necessarily mean trustworthiness.

KBT Analysis

As you all remember I published an article about how Google calculates trust factor of entities. And in that dissection of Google’s patent we learned the mechanics behind Google’s method of trust calculations for entities. I even mentioned the trust factors which are based on external signals such as proximity to “The seedlist” (through links) and “Link juice” from high trustworthy entities, but something was missing from that picture.

We learned that Google knows about the trust level of an entity by the way the user engages with it (trust button) but it didn’t shed light on how Google knows that the knowledge presented by the entity is actual trustworthy and how does that come into play when calculating trust rank.

Correctness of facts as a signal of trustworthiness

Not Long ago a group inside Google published a full Paper named “Knowledge based trust: estimating the trustworthiness of web sources”. This full paper describes the algorithmic method for calculating trustworthiness without the need of external signals such as “links”, this method is based on the extraction and analysis of facts written in web sources. In layman’s term it means the more truthful and accurate your source is, the more trustworthy it is.  

The Implications of this document on rankings are huge especially if it is already been implemented into search rankings calculations. This document incorporate AI systems into its methods.

Google has many tools that are using its machine learning methods such as “Google prediction” which is part of its machine learning algorithm, and it is incorporated in parts of the method of data extractions. This Document is suggesting a sophisticated multi-layer system named “Knowledge Triples” (Subject, predicate, object) to overcome former weaknesses in calculating Trust based on KV (Knowledge Vault).

In the opening of the document you can see the following distinction:

Quality assessment for web sources is of tremendous importance in web search. It has been traditionally evaluated using exogenous signals such as hyperlinks and browsing history. However, such signals mostly capture how popular a webpage is. For example, the gossip websites listed in [16] mostly have high PageRank scores [4], but would not generally be considered reliable. Conversely, some less popular websites nevertheless have very accurate information.

Text as shown in the actual paper

This distinction between trust and popularity is crucial especially for our industry, Google works differently with each query depending on the “reasons” of the user to search it and the context within it was searched. Therefore if our content answers all the different “reasons” one might type it in the first place while using very in depth content based on facts we will be able to gain enough trust and authority over time to climb to the top results and get pinned there.

High PR = High Trust in most cases

PR algorithm has been claimed to be “dead” and discarded by Google (which is completely Wrong but that is another story) it still provides a strong indication on the “overall strength” of websites in SERPS when compared to their type of industry and query.

If you ever read the original PR formula you’ll see that the algorithm is based on links, the more links you have the more PR flow they give you, the Higher your PR gets.

High trusted websites has high PR in most cases unless it is faked – PR has an important role in this trustworthiness calculations because the researches had no way of calculating “quality” of a web source so they Used “PR” as a benchmark, and found that their data is almost the same expect in several cases.

High PR but Low KBT (Knowledge Based Trust)

In some cases you see a lack of alignment between the KBT and PR. That means that in some of these cases they will need to improve their algorithms combining them with many different signals (authority, spaminess of websites etc.) to reach more accurate results although the algorithm was very accurate by itself, meaning websites such as Forums receive low Trustworthiness because they are full of inaccurate information.

High KBT and Low PR

In this case it shows that their new method of calculating KBT is so accurate that even with lacking of “external” signal they managed to calculate trustworthiness correctly in more than 85% of the times.

KBT is just one of Google suggestions for future development, in the end of the paper you can see the suggestions for future development so this KBT method can get even better.

Final Words

We must strive to provide the best value for our users in everything we publish while incorporating true knowledge in order to enrich the world around us, spreading our wisdom and knowledge forward. That is the only way we will be able to survive in the future of online marketing.

“Expressing your true knowledge in everything you create internally and externally”