How does Distributed Models Encoding impact Google ranking after 2014?
Has anyone seen any ranking changes tied to Google's shift to distributed models encoding since 2014? I'm wondering if this affected ecommerce product pages more than others, especially pages with duplicate or thin content. Any examples or data from tools like Ahrefs or SEMrush?
Asked by ehsanulhaq
1 Answer
Yeah, Google's move to distributed models encoding like word2vec and the stuff that led into RankBrain definitely changed how thin and duplicate content get filtered, especially after 2015. For ecommerce, I saw in SEMrush and Ahrefs that a lot of thin product pages with almost identical titles and descriptions dropped visibility between 2016 and 2018. Looking at SEMrush Sensor/visibility graphs for big ecommerce sites, the ones who rewrote product copy to be more unique and used better internal linking clawed some of it back. The biggest drops were in pages with template content or just manufacturer snippets. It's worth running Site Audit in SEMrush to find thin/duplicate pages and focus on making those unique or at least consolidating SKUs. Duplicate meta tags show up hard in these tools too. If you have access to log files, you'll also see Googlebot crawling less on those near-duplicate pages after the changes.