Google has just confirmed a data problem many search marketers have been feeling for months. Search Console still treats an AI Overview as one big block, and that single decision is warping the rankings, impressions and traffic reports you show clients.
Search Advocate John Mueller told a Reddit thread that tracking classic rankings for AI features is “hard to do in a way that makes it useful.” He confirmed that Search Console counts the entire AI-generated overview as a single listing. The practical result: squeezing AI answers onto an old 1-to-10 scale no longer works.
Your average position was already lying
Even before AI, the average position metric blended wildly different SERPs into one tidy number. A page can rank 1 when an AI Overview appears and 19 when it doesn’t. If both happen equally, Search Console reports an average position of 10. That false sense of consistency hides the two extremes real users saw.
Average works when numbers cluster together. Search results do not. They shift by location, device and result type, so one blended number rarely matches what a real visitor experienced.
Block flattening makes it worse
In AI reports, Search Console uses block flattening. Instead of tracking where a link sits inside the summary, Google assigns the entire overview box one rank. Since AI Overviews sit above traditional listings, every link inside gets position 1—including links buried behind a “Show More” toggle.
That means an eye-catching link card near the first sentence and an almost invisible link hidden in an expandable section both show as position 1. The data calls them identical even though one is far more likely to be clicked.
Impression rules add another distortion. A default link counts as an impression the moment the page loads, whether or not anyone scrolls. A hidden link behind a toggle does not count until someone clicks to expand it. So unread links get full credit, while useful collapsed links look nonexistent.
What the traffic data really shows
Put those quirks together and standard reports become hard to trust. Search Engine Journal points to data showing an AI Overview CTR decline of 61%, with Seer Interactive research finding cited websites’ click-through rates dropped partly because automatic impressions skewed the math.
Meanwhile, a field study covered by SEJ found that having an AI Overview at the top cut clicks to standard organic results by around 38%. AI Overviews appear on 21% of searches and show up more often on question-style informational queries. The old link between top ranking and real visitors is breaking.
Measure what happens, not where you rank
The fix is not another AI visibility tool or a patched-up position metric. Mueller asked the SEO community for practical ideas, because modern SERPs no longer look like a simple list.
A clearer reporting approach looks like this:
- Treat AI visibility as a yes-or-no question: is your brand being cited?
- Track outcomes that impression rules can’t distort: actual visits, inquiries and conversions.
- When placement matters, inspect the live search layout directly or use automated live-screen tools.
A neat average position of 1.2 in AI reports usually just means your link appeared somewhere in the overview box. That says little about whether a person saw it, trusted it, or clicked it. The familiar ten blue links worked for 25 years, but forcing modern search layouts into that reporting shape now obscures what your audience actually does.
Source: Search Engine Journal
