Google is running a layout test in AI Mode, its conversational search experience, that moves the results count and the source favicons from above the citation and link cards to a position below them. Brodie Clark of SERP Alerts spotted the variant and posted screenshots on X. Search Engine Roundtable, where Barry Schwartz reported the test on June 9, 2026, noted the change appears on a subset of results with no confirmed rollout.
The detail sounds minor, but placement governs attention. In a zero-click result, where a generated answer often satisfies the query before any link is clicked, source attribution that sits above the answer earns more early eye contact than attribution pushed below it. A reader who gets a complete response has little reason to scroll past it. Moving the count and favicons under the cards puts them later in the reading path, after the moment many users decide they are done.
This fits a year-long pattern. Across AI Mode and AI Overviews, Google has repeatedly shuffled where links, citations, and source labels appear, testing inline citations, expandable source lists, right-hand panels, and varying card formats. Each iteration changes the odds that a user notices a publisher and clicks through. None of these tests arrives with a public explanation, and most surface only because independent trackers like Clark catch them in the wild.
The problem for publishers is measurement. Google’s generative-AI performance reporting in Search Console remains thin on the click and impression detail that would let a team isolate how a given layout affects referral behavior. Teams can see that AI surfaces exist and that traffic patterns are shifting, but attributing a change in clicks to a specific attribution-position test is currently out of reach with first-party data alone. That gap is why a single experiment like this one is weak signal. It is one variant, on some results, with no published click data, and it may never ship.
It is still worth logging. Small changes to where a source sits relative to an answer compound across the volume of queries Google handles. A placement that trims a fraction of a percent off citation visibility becomes a measurable traffic line at scale, and the effect lands unevenly: sites that depend on AI Mode and AI Overviews citations for discovery absorb more of it than sites with strong direct or brand demand. The teams that will understand the impact first are the ones already watching the right numbers.
The action is straightforward. Separate AI Mode and AI Overviews referrals into their own analytics segment now, before any attribution-position change reaches general availability. Tag the referrer patterns Google uses for these surfaces, baseline your current click-through from them, and annotate the dates when trackers report layout tests. When a change like this one does roll out broadly, a team with that segment in place can detect the shift in days rather than guessing after a quarter of eroded traffic.
Search Engine Roundtable (seroundtable.com), reported by Barry Schwartz on June 9, 2026.
