Was the large rise in impressions without clicks due to bots, not AI Overviews? Fascinating question. I, like many others, assumed AI Overviews was the culprit. the narrative made, and still does, make sense. But this is a compelling alternative explanation.
SEO News: if you missed my post over the weekend, there’s been some massive fluctuations to data in GSC. Interestingly, this change may actually point to a broader underlying issue in our data that was originally thought to be related to the AI overviews rollout. Impressions going up but clicks not following. This has been the shared understanding of the SEO community based on recent studies, but we may be now discovering how flawed this methodology was in the first place. Google has been pushing back on these studies recently and the disabling of &num=100 may have motives outside of preventing competition from scraping their search results. The outcome? Well, it seems like desktop impressions in particular in GSC for many sites (due to desktop being the primary rank tracking surface) may have been heavily influenced from an influx in bots (due to new AI analytics entrants) scraping their search results. We can see this in the graph shown, where desktop impressions have dropped a whopping 200K impressions overnight for this site, forcing average position to move in a direction that is more close to mobile. So you’re telling me 200K daily impressions on desktop were from scraper tools!? That seems a bit ridiculous, but potentially not fully outlandish. There is likely more to this story. The reason for this is that by default, most users would have the top 10 search results shown in Search. An impression is only recorded in GSC if the result is loaded on that page and the link is within visible, but it doesn’t need to be scrolled into view. When the top 100 results are loaded, this will completely throw this dataset out of whack, causing the average position to increase dramatically. Many think that Google has made this change to prevent external services abusing their systems. While this may be true, I do wonder whether there are ulterior motives in putting “the great decoupling” theories to rest.
