Key Takeaways
- A self-promotional listicle is a “best tools” or “top agencies” article where the publisher ranks itself at or near the top.
- The tactic spread because it worked for influencing AI answers that pull from ranking pages.
- Google said in April 2026 that it is aware of self-serving listicles and working to combat low-quality abuse.
- On May 15, 2026, Google updated its spam documentation to state that its spam policies apply to AI Overviews and AI Mode.
- Lily Ray’s research found sites heavily using the tactic lost significant visibility starting in January 2026. Honest comparisons that genuinely help buyers are still fine.
Why this tactic took off
AI answers often lean on pages that already rank. If someone asks an assistant for the best project management tools, it may summarize a few “best project management tools” articles. So companies started publishing their own versions of those lists, with themselves at number one. Some did it at scale. Lily Ray reported finding one company with around 2,000 of these pages.
For a while it worked. Search Engine Journal described these lists as one of the most effective ways to influence AI answers, at least in the short term.
What Google has said
In April 2026, after The Verge reported on the spread of self-dealing lists, Google said it was aware of the low-quality abuse and working to address it. Search Engine Roundtable covered the warning on April 14.
Then, on May 15, 2026, Google updated its Search spam policies documentation to clarify that all existing spam policies apply to generative AI features in Search, including AI Overviews and AI Mode. Coverage from PPC Land pointed out that this turned an earlier informal warning into documented policy.
What the data shows
Lily Ray has tracked this closely. In her research, Google appears to have made an adjustment in January 2026 that substantially reduced the organic visibility of sites leaning heavily on self-promotional listicles, with the biggest hits in the subfolders where those lists lived. In a later interview she said she had spoken with about 20 affected companies and identified around 40 more in her own research, and that she was seeing early signs ChatGPT was moving in the same direction.
Her read is that the outcome now depends heavily on the strength of the brand doing the publishing. A well-known, well-recommended brand may still get away with it. For the average site, it has shifted from an asset to a liability.
It is worth being fair here: whether every AI platform is actively discounting these lists is still unsettled. But with Google’s spam policies explicitly covering AI answers, and a core update rolling out in November, this is not a tactic to scale right now.
What is still fine
Comparison content is not banned. Buyers want comparisons, and good ones help them. The problem is bias dressed up as an objective ranking. Content that tends to hold up:
- Honest head-to-head pages that say clearly who you are best for and who should pick a competitor.
- Lists with disclosed methods, real testing and specifics, where your own product is clearly labeled as yours.
- Third-party lists written independently by publications in your space.
What to do if you published a lot of these
- Inventory them. Find every list where you rank yourself first.
- Keep the useful few. Rewrite them as transparent comparisons with clear disclosure.
- Merge or remove the rest, especially near-duplicate lists created for keyword variations.
- Earn the recommendation instead. The durable version of this tactic is getting real publications to mention and link to you. That is what AI systems treat as consensus, and it is the heart of our managed link building service.
Artemis covered this in the “Google’s spokesperson warns against listicles?” episode of our SEO + AI News series. If you want a hand cleaning up risky content before it costs you, our done-for-you SEO team can help.






