Thursday, March 5, 2015

Google to Rely on Trustworthiness Instead of Site Popularity - Here Comes Knowledge-Based Trust (KBT) Ranking Score

We all have talked about links in the past but in future Google is set to consider trust (accuracy) of the content as a major ranking signal instead of site popularity which is solely based on links. A new ranking factor named Knowledge-Based Trust (KBT) score is set to affect all the webpages and they shall be ranked on the basis of high to low KBT scores. (Certain exceptions apply as the actual Google algorithm is based on so many other factors) but most importantly links will lose some of its importance to KBT scores.

What is Knowledge-Based Trust (KBT) Score?

Judging the quality and accuracy of information is a tremendous task. Traditionally Google evaluated it using exogenous signals such as hyperlinks and browsing history. These signals however weighted popularity of a webpage more instead of it's trustworthiness. Some less popular websites having very accurate information neveer get the same visibility because the main ranking algorithm relies of PR that is measure of popularity. 

To break the shackles, Google has proposed a way to distinguish errors made in the extraction process from factual errors in the web source by using joint inference in a novel multi-layer probabilistic model. This trustworthiness score is known as "the Knowledge-Based Trust (KBT). 

What's New In This Model?

Under this new model, sites having high PR scores or DA might have low KBT scores because none of the PR algorithm or Moz DA algorithm includes factual correctness of information while computing the webpage scores. Hence, this new model if proved to be correct in the longer run will provide new means to judge the quality of the webpage. 

KBT is different from all other previous models of source importance computation. PageRank and Authority-hub analysis consider signals from link analysis. EigenTrust and TrustMe consider signals from source behavior in a P2P network. 

Web topology, TrustRank, and AntiTrust detect web spams. But, KBT considers an important endogenous signal — the correctness of the factual information provided by a web source.

What Will Be the Effect of KBT Scores On My Site?

Sites that are having high popularity like the QA sites often have inaccurate information in them because those are directly answered by users and are not proofread by any editors. This is opposed to the information provided by Wikipedia which are proofread by editors. Hence, QA/Forum sites although have a score in popularity (means high DA, PR as well as high Alexa) but they might actually have a low KBT score. If your site does not provides accurate data, it might fall in the category of low KBT score and hence lesser rankings.

Low Page Rank Sites Might Have High KBT Scores 

and

High Page Rank Sites Might Have Low KBT Scores

Does High KBT Scores Means High Rankings?

The short answer is no :) 

The Google ranking algorithm is composed of so many factors and KBT is just one of them. So, there is no guarantee that a high KBT score might guarantee high rankings always. But yes, most of the times if other factors favor your site then you can enjoy higher visibility on Google.

Are Links Dead?

Not at all. Links still hold a lot of value because site popularity is still a factor in the ranking algorithm. Links might have lost some of its value but they are not dead.

How to Increase KBT Scores?

KBT scores is based on trust. Google compares the quality of the information presented on your site with several other pre-compiled resources and checks its accuracy. This means, having accurate information in your site is more important than ever. if the content facts presented on your site matches with those present in n number of high trust score websites then you will have a high KBT score. 


Post a Comment