Self-Promotional Listicles

A self-promotional listicle is a 'best X' or 'top X' article published by a company that ranks its own product first, usually without independent testing or disclosed criteria. The format spread as a GEO tactic because AI answers to 'best' questions often cite list articles, but since early 2026 there has been growing evidence that Google discounts it in search results and in AI Overviews.

The appeal is easy to see. When someone asks ChatGPT, Perplexity or Google's AI features for the best tool in a category, the sources behind the answer are often comparison and list pages (see citation content types). A vendor that publishes its own list, puts itself at number one and refreshes the year in the title can end up among those sources. Some companies produced these pages at scale, one for every category and use case they could think of.

The evidence that this has stopped working comes mostly from one researcher, so it should be read with care:

• In February 2026, Search Engine Land reported research by Lily Ray of Amsive on SaaS and B2B brands that lost 30% to 50% of their organic visibility within weeks in January 2026. The losses were concentrated in blog and guide sections holding dozens or hundreds of self-ranking listicles, many refreshed with '2026' in the title and little else.
• In June 2026 Ray analyzed 100 'best B2B software' queries in Google's AI Overviews. When a brand's own listicle was cited, the brand was left out of the actual recommendation 69% of the time.
• Her September 2026 update of the same query set found AI Overviews citing 38% fewer of these pages than in June, and the self-ranking brand omitted 83% of the time when its page was cited.

Other data points the other way. According to a PPC Land summary, Kevin Indig found in August 2026 that 62% of vendor-written list pages had gained estimated organic traffic since January, and a Scrunch study of about 10,000 URLs cited by seven AI assistants found that a cited self-ranking page lifted its author's recommendation rate from about 4% to about 7%. The studies measure different surfaces and outcomes, and none has been independently audited.

Google has not named self-promotional listicles in any policy. Two May 15, 2026 changes are relevant, though. Google's spam policies now explicitly cover attempts to manipulate generative AI responses, and its guide to generative AI features says seeking inauthentic mentions 'isn't as helpful as it might seem'. A brand that publishes hundreds of near-identical self-ranking lists also risks the scaled content abuse policy, and lists built from generic category facts are commodity content. Self-promotion is one of the tactics often grouped under AI spam.

The most useful finding for GEO is the gap between being cited and being recommended. In Ray's data, AI Overviews often used a vendor's list as a source while recommending its competitors, so the page did more for rivals than for its author. Tracking only citations would count that as a win. Promptwatch records citations and brand mentions separately for each answer, which is how that gap shows up in practice.

Comparison content is not off limits. Pages that disclose that the publisher is a vendor, explain the criteria, treat competitors fairly and show real testing are closer to what Google's review guidance asks for. The more durable route to appearing in 'best' answers is being included in lists other people publish, which is covered under third-party mentions and in our post on whether listicles still work.