ICYMI: Google poisons SERPs since Sept/October to hit ChatGPT & Co.

Key Takeaways

Perkins from Nozzle analyzed 42.5 million SERPs and found that Google swapped results with ones of different but close enough topics so the SERPs still looked normal
27.6% are poisened
on November 2nd even 68.6% of SERPs were poisoned
Moretti stated that poisoned SERPs have the following in common:
introduce artificial SERP volatility (while normal searches show 80% similar webistes from run to run, poisened SERPs show roughly 20%
show more UGC
show fewer AI Overviews
advertising and shopping results appear stable
Semrush reported two forms of poisoned SERPs:
Results on a SERP do not match the query at all
a SERP that looks normal at first glance with the first results being legit but further done the SERP shows results that only loosely match the search - likely because the SERP was poisened by results matching only one word of the multi-word query
SerpApi reported similar behavior not only for Google Search but also for Google Shopping results
So, data is poisoned, less reliable. It takes more tries and checks to get accurate data. Volatility might be screwed. Costs for rank trackers and AI chats increase again.




Sources:
![ICYMI Metehan: Statistics & Fact Checking in AI Chats and beyond the Web [Experiment]](https://static.wixstatic.com/media/24c590_b0a8e1392c1c47f28a0a46a67d59df71~mv2.jpg/v1/fill/w_980,h_653,al_c,q_85,usm_0.66_1.00_0.01,enc_avif,quality_auto/24c590_b0a8e1392c1c47f28a0a46a67d59df71~mv2.jpg)

