TLDR
Google's AI Overviews excluded brands that ranked themselves first in their own 'best of' listicles 83% of the time in September 2026, up from 69% in June, per Lily Ray's 100-query study. In 70% of cases the self-ranking page actively helped a competitor earn the recommendation instead. Two other studies find organic traffic and citation odds still rising, making the tradeoff between search clicks and AI recommendations the question Australian publishers now face.
KEY TAKEAWAYS
What Ray's September re-run found
Ranking yourself first in your own 'best of' page was a standard SEO play for years. Google's AI Overviews (the generative answer boxes Google displays above standard search results) have flipped the incentive. Amsive's Lily Ray re-ran 100 'best B2B software' queries in September 2026 and found brands that put themselves at number one were left out of AI Overview recommendations 83% of the time, up from 69% in June.[1]
That is a 14-percentage-point move in three months. Ray said: "Compared to June, Google's AI Overviews now cite 38% fewer self-promotional 'best of' listicles, and they leave the publishing brand out of the recommendation 83% of the time, up from 69% in June."[1]
In 70% of cases, the self-ranking page did not simply get ignored, it actively helped a competitor earn the AI recommendation instead of the brand that published it.[1] Publishing a 'best of' page with yourself at the top has become, in the majority of cases, a subsidy for the rivals listed below you.
Where the three studies disagree
Two other data sets point in a different direction. Kevin Indig's Growth Memo study, published 24 August 2026, found that 62% of vendor-authored listicles gained estimated organic traffic since January 2026, with a median page gain of 38%.[2] Organic search clicks held up even as AI Overview citations fell.
Scrunch's analysis, published 24 September 2026, found that when an AI assistant does cite a self-promoting listicle, the brand's recommendation odds rise from roughly 4% to about 7%.[3] That is a real relative lift, sitting against an 83% exclusion rate, so the upside is narrow.
Ray measures AI Overview citation behaviour. Indig measures organic click volume. Scrunch measures recommendation probability once a citation has already happened. A publisher can be winning clicks from Google's ranked results while simultaneously being excluded from the AI-generated answer box sitting above them. Which one matters more depends on where buyers are actually looking.
Google's tightening update cadence
The intervals between Google's three spam updates in 2026 shrank from 92 days between March and June, to 55 days between June and August, to 37 days between August and September, a compression of roughly 60% over two consecutive cycles.[1]
The September spam update was still running as of Ray's 4 October 2026 analysis, the third in that tightening sequence.[4] On 1 October 2026, Google separately updated its guidance on generative AI content, stating it is critical to manually fact-check and review all AI-generated content for accuracy and trustworthiness before publishing, and extended that requirement to titles, meta descriptions, structured data and alt text.[5]
Prompt the Market covered Google's updated helpful-content guidance when it landed, noting Google now grades pages on effort, originality, talent and accuracy.[6] A self-promotional listicle that ranks its own product first without independent testing scores poorly on all four dimensions.
What changes for Australian publishers
Australian SaaS vendors, digital agencies and comparison sites have relied on 'best X in Australia' pages as a standard acquisition channel for years. Ray's September figures apply to B2B software queries globally, and the same mechanism runs for any vertical where a publisher controls both the listicle and the product it recommends.
The practical question is whether a page is generating search clicks, AI Overview recommendations, or both, and whether that answer has changed since January. Indig's data shows organic traffic held up for most vendor listicles through August 2026, with clicks potentially still arriving even as AI Overviews route buyers to competitors.[2] Publishers who have not separated those two measures are optimising for a mix they cannot see.
Ray's June baseline had the exclusion rate at 69%.[1] Her September re-run put it at 83%. With Google's spam update still running and the next core update due inside a window that has been shrinking by roughly a third each cycle, any Australian 'best of' page that has not been audited against Ray's exclusion findings is carrying materially more risk than it was carrying in June.
SOURCES & CITATIONS
- Lily Ray, Prediction: The Next Massive Google Update
- Kevin Indig, Has Google Demoted Listicles? (Growth Memo)
- Scrunch, What Happens When You Crown Yourself Number 1
- Prompt the Market, Google's September Spam Update Could Run for Two Weeks
- Google, Using Generative AI Content (Search documentation)
- Prompt the Market, Google Grades Pages on Effort, Originality and Accuracy
FREQUENTLY ASKED QUESTIONS
What is a self-promotional listicle?
What are Google AI Overviews?
Does being cited in a self-ranking listicle help a brand's AI recommendation odds at all?
How fast is Google updating its spam policies in 2026?
James Dore is Strategy Director at Bushnote, a leader in SEO and AEO in Australia. He writes on search, AI and the technology decisions shaping government and industry.
