Listicle Citation Rate in AI Search: 2026 Report

Our research team measured how often AI search engines cite listicles, the ranked "best X" and "top 10" pages, compared with other page types, so marketers can see how much citation share lists earn by engine, query intent, and industry, and how that share moved during 2026. We ran 2,400 prompts across 12 industries on ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Gemini, and Microsoft Copilot every two weeks from February 9 to September 11, 2026, producing about 2.06 million citations. A listicle was defined as a page whose main body ranks or enumerates at least three named options with commentary on each.

Listicle Citation Rate in AI Search by Engine

In the table below, we show the share of each engine's citations that pointed to listicles across the full study window, alongside three other ways of expressing the same rate.

The four columns explain most of the disagreement in the published record. Measured as a share of all citations, listicles account for 16.5% of what the six engines cite. Measured as a share of the 1,000 most-cited URLs per engine, the figure nearly triples to 47%, because a small number of heavily updated roundups get cited again and again across related prompts. A researcher who samples only the most-cited pages will report a figure closer to the 63% Evertune published for top-cited URLs, and one who counts every citation will land much closer to ours.

Perplexity shows why the per-answer view matters. Its listicle share of citations is below average at 14.2%, yet it has the highest share of answers citing at least one list, at 52%. Perplexity attaches an average of 17.6 sources to each answer in our data, against 5.1 for ChatGPT and 4.3 for Copilot, so a list has more slots to fill even when it wins a smaller fraction of them. Google AI Mode sits at the bottom on share because it leans more heavily on brand homepages and product pages than any other engine we tested. The gap between AI Mode and Google AI Overviews, 9.4% against 22.8%, is one of the larger splits in the dataset, given that both products draw on the same index. AI Overviews tends to summarize the ranked pages for a query, and those ranked pages are frequently lists; AI Mode runs more follow-up searches and ends up closer to the vendors themselves.

Listicle Citation Rate by Month, February to September 2026

The table and chart below track the monthly listicle share of citations for ChatGPT, Google AI Overviews, and Perplexity, and for all six engines combined. September covers a single collection cycle.

• We found that ChatGPT's listicle share held between 14.6% and 15.9% for six months, then fell to 8.6% in August, a relative drop of 42% from July. Nearly all of the decline appeared in collection cycles that ran after the August 6 model change, and the lower level held through the September cycle with no sign of recovery.
• Our data showed no matching drop on Google's side. Google AI Overviews stayed within a two-point band all year and peaked at 23.8% in June, which points to a change specific to ChatGPT. Across the other five engines combined, the listicle share in August was within half a point of the July reading.
• We observed a slow, steady erosion at Perplexity, from 15.1% in February to 13.2% in September. The citations it withdrew from listicles went mainly to discussion threads and publisher reviews, and the decline was gradual enough that no single month accounts for more than a third of it.

Listicle Citation Share Compared With Other Content Formats

In the table below, we set listicles against the other page types the six engines cite, pooled across all engines and the full study window. Shares are calculated on all 2.06 million citations, so the rows and the unlisted remainder sum to 100%.

• We found that listicles ranked second among all page types, behind articles and explainers at 19.8%, and ahead of product and category pages at 15.1%. On Gemini and Google AI Overviews, listicles ranked first. The remaining 8.4% of citations went to video pages, encyclopedia entries, documentation, and a long tail of other formats.
• Our data showed that listicles win on volume more than on placement. Their average position of 6.8 in an answer's source list trails brand homepages at 4.2 and articles at 5.1, meaning lists are often cited as supporting evidence after the engine has already named its primary source.
• We observed that comparison pages punch above their share. They drew only 5.6% of citations, but when an answer cited one, it cited it at an average position of 6.1, ahead of listicles, and usually alongside at least one list covering the same products. In 71% of answers that cited a comparison page, the same answer also cited a listicle.

Listicle Citation Rate by Query Intent

The table below breaks the listicle rate out by the intent behind each prompt, with the ChatGPT change after August 6 shown separately for each group.

Intent moves the listicle rate more than engine or industry does. Commercial investigation prompts, the "best project management tool for a 20-person team" kind, sent 35.4% of their citations to lists, and 81% of those answers cited at least one. Navigational prompts barely touched the format, because the engine already knows which site the user wants. The pooled rate for all prompts works out lower than a prompt-weighted average of these groups would suggest, since informational answers carried more sources apiece (10.4 citations on average, against 8.4 for commercial answers) and pulled the overall share down.

The ChatGPT column is the most useful for anyone planning content around lists. The August change cut listicle citations for commercial investigation prompts by 52%, the steepest decline of any group, while local prompts lost only 14%. Our reading is that ChatGPT now reaches past generic software and product roundups to cite vendor pages and reviews directly, but still relies on lists when the underlying options are restaurants, hotels, or service providers with thin websites of their own. That resilience goes further in Nicolas Sitter's weekly tracking of 616 hotel prompts, in which listicle share rose from 33.8% to 36.4% across the same transition. It also suggests that the lists most exposed to future model changes are the generic software and product roundups that currently earn the largest share of citations.

Listicle Citation Rate by Industry

In the table below, we show the four industries with the highest listicle rates and the four with the lowest, out of the 12 in our prompt set, along with how many listicle citations went to lists published by a company that ranks itself.

• We found a sixfold spread between the most and least listicle-heavy industries, from 31.8% in B2B software to 5.1% in legal services. Industries where buyers compare many interchangeable options draw the most list citations, while regulated fields such as healthcare and legal services lean on institutional, government, and bar association sources, along with the websites of individual providers.
• Our data showed that self-promotional lists, where the publisher ranks its own product or firm, made up 16% of all listicle citations across the 12 industries. B2B software led at 27%, consistent with Ahrefs' finding that 67.6% of Google results for software "best" lists came from publishers ranking themselves first.
• We observed that a citation to a self-promotional list rarely translated into a recommendation for its publisher. In B2B software, engines named the list's author among their top picks in 36% of the answers that cited it, echoing Lily Ray's June analysis, in which Google AI Overviews recommended other companies in 224 of 323 such citations.

Conclusion

Across the full study window, listicles accounted for 16.5% of all citations from the six engines we measured. That rate rose to 35.4% for commercial investigation prompts asking for the best option in a category, and fell to 1.2% when users asked about a specific company by name. ChatGPT's listicle share dropped by about 42% after OpenAI's August 6 model change, while Google AI Overviews held steady. For publishers still investing in ranked lists, the practical reading is that listicles remain a major citation format on Gemini and Google AI Overviews, especially for commercial and local prompts, and that generic software and product roundups are the lists most exposed when a model change cuts list citations.

Requesting a Copy of This Report

If you would like a PDF copy of this report, or the full prompt set and page-type classifications behind it, you can reach out here.

Sources

• Listicle Citation Rate in AI Search Study, Independent Research Team, September 2026, New York, New York.
• The Listicle Window Is Closing in AI Search: 30% Decline MoM. Seer Interactive (Nick Haigler), February 2026, Philadelphia, Pennsylvania.
• AI Search Loves Listicles: What 25,000 URLs Reveal About Citations. Evertune via Search Engine Land, May 2026, Redding, Connecticut.
• Listicles Are the Most Cited Content Type in AI Search. Qvery (Vlad Shvets), March 2026, Tallinn, Estonia.
• The Pages Your Agency Built to Win AI Citations Just Lost Half of Them. The State of Brand (Outlever), August 2026.
• Did GPT-5.6 Really Kill Listicles? Not in Hotel Search. Nicolas Sitter, August 2026.
• Do Self-Promotional "Best" Lists Boost ChatGPT Visibility? Study of 26,283 Source URLs. Ahrefs (Glen Allsopp), December 2025, Singapore.
• Why Calling Yourself the "Best" Could Be Helping Your Competitors Win in AI Search. Lily Ray, June 2026.
• Listicles for AI Search: Why Marketers Love and Hate Them. Wix Studio AI Search Lab (Kiera Carter), April 2026, Tel Aviv, Israel.