As it does every year, Google has released its annual spam report in searches. However, this edition has had a peculiarity, since in the last year Google launched its new algorithm called SpamBrain. This is a system that automatically identifies spam, thanks to machine learning techniques, and its influence has been noticed in the results shown in the report.
SpamBrain is Google’s algorithm capable of automatically identifying low-quality pages and spam in search results.
In 2021, SpamBrain worked six times more than in 2020. what does this mean? According to Google’s report data, in 2021 the algorithm identified and reduced hacked spam (web pages containing scams and scams) by 70% and “gibberish” spam (web pages and sites that automatically copy other well-ranked pages) and in links by 75%.
By being able to recognize the different types of harmful and malicious behaviors behind these spam pages, Google’s new anti-spam algorithm was able to maintain the 99% of searches spam-free.
One of the main advances of SpamBrain is that it is able to identify on legal pages any link that directs to a website classified as spam and block it automatically. In this regard, during 2021 Google has been able to block 40% of these links and expects that by 2022 a figure of over 90% will be reached.
In this regard, Google wants to expand the effectiveness of SpamBrain so that it is also able to detect malicious content or content that inciting online harassment when performing name searches, thus aiming to reduce abusive content removal practices.
Eliminate manipulation of Google positioning
Google’s algorithm not only seeks to eliminate spam, but aims to minimize low-quality content and manipulation of SERP rankings. To do so, they will fight against behaviors that attempt to violate quality guidelines.
One of the areas where SpamBrain has already improved has been in identifying fake reviews (both positive and negative). Also on product review pages, which often have low quality content, merely rephrasing product descriptions without actually testing the products.