{"@context":"https://schema.org","@type":"BlogPosting","headline":"Did GPT-5.6 Really Kill Listicles? Not in Hotel Search (2026)","description":"A publicly shared analysis (Tomek Rudzki / Peec AI; 1M prompts, ~180M sources, one week before/after the rollout) reported that after GPT-5.6: fan-outs per chat rose ~154%; site: search adoption jumped to 18.4% of chats (+20,018% relative); \"top\" fan-out share fell 75% and \"best\" share fell 56.4%; listicle citation share fell from 15.77% to 7.80% (-50.5%); comparison citation share fell from 9.08% to 6.17% (-32.1%); citations per chat rose ~25%. We re-ran the same before/after comparison on our own 616-prompt AI Hotel Landscape ChatGPT corpus, split at our own observed model-field transition (public.fanout_captures.model switches gpt-5-5 -> gpt-5-6 between the Aug 3 and Aug 10 weekly runs, bracketing August 6 -- ChatGPT.com's actual consumer-default rollout date; July 9 was the developer/API release only). Results, pre (Jul 27 + Aug 3) vs post (Aug 10 + Aug 17): fan-outs per chat 0.959 -> 0.965 (+0.7%, not +154%); \"best\" share of fan-outs 46.2% -> 79.7% (+72.3%, opposite direction from the -56.4% reported, confounded by our own templated \"best hotels in...\" prompt wording being echoed in fan-outs); \"top\" share 8.7% -> 8.3% (-5.2%, far short of -75%); site: operator adoption 0% -> 0% (no change, not +20,018%); citations per chat 9.85 -> 10.03 (+1.9%, a real but much smaller increase than +25%); listicle citation share (own title/URL classifier, not Peec's) 33.8% -> 36.4% (+7.6%, opposite direction from -50.5%); comparison citation share 0.31% -> 1.03% (up on a very small base of 14-76 rows/week, opposite direction from -32.1%). Hypothesis for the gap: hotel destination-discovery prompts (\"best hotels in [city]\") are inherently listicle-shaped and rarely invite a site: or vs-style fan-out, unlike the branded/comparison-shopping queries GEO-spam tactics typically target -- GPT-5.6's retrieval changes may be concentrated there, not in open discovery search. Read-only throughout: existing stored fields only, no new captures.","datePublished":"2026-08-24","dateModified":"2026-08-24","url":"https://nicolassitter.com/research/gpt-5-6-hotel-search-changes-2026","category":"research","keywords":["GPT-5.6","ChatGPT fan-out queries","ChatGPT listicles","GEO listicles","AI search hotels GPT-5.6","ChatGPT site search operator","AEO GPT-5.6","ChatGPT model update search behavior"],"articleSection":"Research","wordCount":1900,"readTime":"8 min","articleBody":"AI Search · Hotels\n\n# Did GPT-5.6 really kill listicles?Not in hotel search.\n\n**TL;DR:** A widely shared analysis (Tomek Rudzki / Peec AI, based on a 1M-prompt, 180M-source panel) found that after GPT-5.6, ChatGPT ran far fewer “best/top/vs” fan-out searches, adopted `site:` search at scale, and cut listicle citations by half. We re-ran the same comparison — fan-outs, citations, content type, before vs. after — on our own AI Hotel Landscape corpus (616 destination-hotel prompts, weekly, since December 2025). In this vertical, none of it replicated. Fan-outs per chat: flat. Citations per chat: up 1.9%, not 25%. `site:` searches: zero, before and after. Listicle citation share: _up_ 7.6%, not down 50%.\n\nNS\n\nNicolas Sitter\n\nPublished August 2026\n\n+0.7%\n\nchange in fan-outs per chat (claimed: +154%)\n\n0%\n\nchats using site: search, before and after (claimed: 18.4%)\n\n+7.6%\n\nchange in listicle citation share (claimed: −50.5%)\n\nAug 6\n\nconsumer GPT-5.6 rollout date our own data lines up with\n\n[Read the Report](#executive-summary)\n\n[Summary](#executive-summary)[1\\. When 5.6 actually arrived](#timing)[2\\. Fan-outs](#fanouts)[3\\. Citations & content type](#citations)[4\\. Why hotels might differ](#why)[Methodology](#methodology)[FAQ](#faq)\n\n## Executive Summary\n\nA real, widely cited effect — that doesn't show up in this vertical.\n\nIn late August 2026, an analysis based on Peec AI's data (shared publicly by Tomek Rudzki) circulated widely: after ChatGPT's GPT-5.6 rollout, fan-out searches per chat roughly doubled, `site:`\\-operator searches went from near-zero to 18.4% of chats, and listicle/comparison pages' share of citations fell by roughly a third to a half. The interpretation — ChatGPT retrieving more, more precisely, and going straight to source over GEO-optimized listicles — is a plausible and important story for anyone doing AEO/GEO work.\n\nWe have our own weekly, first-party ChatGPT hotel-search corpus running since December 2025 — a natural place to check whether the same effect shows up. It doesn't. Not in fan-out volume, not in `site:` adoption, not in citation content mix. If anything, several numbers move in the opposite direction. This isn't a rebuttal of the original finding — a 1M-prompt, 180M-source panel is a much bigger and more diverse sample than ours. It's evidence that whatever changed in GPT-5.6's retrieval behavior, it didn't change hotel destination-discovery search the same way.\n\n**One methodology note before the numbers:** July 9 is GPT-5.6's developer/API release date, not when it became ChatGPT's consumer default — that happened August 6. Our own capture pipeline's observed transition lines up with the consumer date, not the developer one. Worth checking which date a claim is actually anchored to before you slice your own data on it. See Section 1.\n\nSection 1\n\n## When GPT-5.6 actually arrived — in our data\n\nBright Data's ChatGPT scrape reports a `model` field directly in its response envelope — not something we compute, a value it observed. We checked it week by week across our hotel corpus, expecting to see a switch around the widely cited July 9 date.\n\nModel field on chatgpt captures, AI Hotel Landscape corpus, weekly.\n\nWeek\n\nModel observed\n\nn captures\n\n2026-07-13\n\ngpt-5-5\n\n616\n\n2026-07-21\n\ngpt-5-5\n\n614\n\n2026-07-27\n\ngpt-5-5\n\n616\n\n2026-08-03\n\ngpt-5-5\n\n616\n\n2026-08-10\n\ngpt-5-6\n\n615\n\n2026-08-17\n\ngpt-5-6\n\n582\n\nJuly 9, 2026 is GPT-5.6's developer/API release date — not when it became ChatGPT.com's consumer default. That happened August 6, 2026. Our own capture surface (which scrapes the consumer ChatGPT experience, not the API) shows GPT-5.5 through August 3 and GPT-5.6 from August 10 — a transition window that brackets August 6 almost exactly. Using the developer-release date as the pivot for a consumer-surface comparison would have compared GPT-5.5 against GPT-5.5 — a null result by construction, and a reminder to check which release a public claim is actually anchored to before slicing your own data on it. Every comparison below uses our own observed transition (Aug 3 vs. Aug 10).\n\nSection 2\n\n## Fan-outs: essentially unchanged\n\nTwo weeks of GPT-5.5 (Jul 27, Aug 3) vs. two weeks of GPT-5.6 (Aug 10, Aug 17), same 616-prompt hotel library, same destinations.\n\nAI Hotel Landscape corpus, ChatGPT, ~616 destination-hotel prompts/week. Peec AI figures per the publicly shared analysis, different corpus and methodology — not directly comparable, shown for contrast.\n\nMetric\n\nPre (GPT-5.5)\n\nPost (GPT-5.6)\n\nChange\n\nPeec AI claimed\n\nFan-outs per chat\n\n0.959\n\n0.965\n\n+0.7%\n\n+154%\n\n\"best\" share of fan-outs\n\n46.2%\n\n79.7%\n\n+72.3%\n\n−56.4%\n\n\"top\" share of fan-outs\n\n8.7%\n\n8.3%\n\n−5.2%\n\n−75% (share); −36.5% (per-chat)\n\n\"vs\" / \"comparison\" fan-outs\n\n0%\n\n0%\n\n—\n\nnot directly reported\n\nChats with a site: fan-out\n\n0%\n\n0%\n\n—\n\n18.37% (+20,018% relative)\n\n**An important confound, in the interest of full disclosure:** our prompts are templated — “best/luxury/affordable hotels in \\[city\\], for \\[traveler type\\] / with \\[amenity\\]”. When ChatGPT's fan-out sub-query closely echoes the original prompt (which it often does — e.g. a “best business hotels Gold Coast…” prompt fans out to “best business hotels Gold Coast Queensland convention centre business travelers”), our “best” share is partly measuring prompt design, not an independent ChatGPT search-strategy choice. That confound exists in both the pre and post period equally, so the _direction_ of change is still real — “best” share went up, not down — but the absolute numbers shouldn't be read as a clean measure of ChatGPT's organic term preference the way a more varied prompt panel's would be.\n\nThe `site:` operator result is the starkest: zero occurrences in any of the four weeks we checked, before or after. For a destination-discovery prompt (“best hotels in Bangkok”), there's no obvious single domain to target with `site:` the way there might be for a branded or comparison query — so this may be a query-shape effect as much as a model-version effect.\n\nSection 3\n\n## Citations: more of them, and listicles didn't lose ground\n\nListicle/comparison shares use our own title+URL heuristic classifier (see Methodology) — not Peec AI's classifier, and not directly comparable in absolute terms. \\*Comparison citations are a very small absolute count (14–76 rows/week out of ~6,000), so this percentage swing is noisy.\n\nMetric\n\nPre (GPT-5.5)\n\nPost (GPT-5.6)\n\nChange\n\nPeec AI claimed\n\nCitations per chat\n\n9.85\n\n10.03\n\n+1.9%\n\n+25%\n\nListicle share of citations\n\n33.8%\n\n36.4%\n\n+7.6%\n\n−50.5% (15.77% → 7.80%)\n\nComparison share of citations\n\n0.31%\n\n1.03%\n\n+237%\\*\n\n−32.1% (9.08% → 6.17%)\n\nReal examples of what our classifier calls a listicle, still showing up in GPT-5.6 citations: Tripadvisor's “THE 10 BEST Business Hotels in Gold Coast 2026,” Trip.com's “10 Best Hotels with Rooftop Pools in Beijing,” Latitude's “The 10 best luxury hotels in Bangkok in 2026.” These are exactly the page type the original analysis described as losing ground — and in our corpus, this category's citation share is slightly larger after GPT-5.6, not smaller.\n\nThe one number that moves in the same direction as the original claim: citations per chat, up 1.9%. It's a real increase, just an order of magnitude smaller than the +25% reported — and given everything else stayed flat or reversed, it's not obviously part of the same “search more, cite more precisely” pattern.\n\nSection 4\n\n## A hypothesis for the gap, not a conclusion\n\nThe most likely explanation isn't that the original finding is wrong — it's that hotel destination-discovery is a specific, narrow query shape that may not be where GPT-5.6's retrieval changes concentrate:\n\n-   **Listicles are the destination-discovery answer format, not a GEO workaround, in travel.** “Best hotels in Bangkok” genuinely is a listicle-shaped question. In software/B2B GEO — the world the original analysis's examples (“top 10 \\[brands\\]”, “\\[X\\] vs \\[Y\\]”) point at — listicles and comparison pages are more often a scaled tactic competing with a vendor's own site. ChatGPT reducing reliance on them there doesn't necessarily generalize to a category where the listicle _is_ the primary real source type.\n-   **`site:` search needs an obvious target domain.** “Site: marriott.com best hotels in Paris” isn't a natural sub-query for an open discovery prompt the way it would be for “\\[brand\\] pricing” or “\\[brand\\] reviews.”\n-   **Our prompt library doesn't include comparison-style prompts at all.** “X vs Y” fan-outs were 0% in every week we checked, before and after — there may be nothing for GPT-5.6's retrieval change to act on if the underlying query never invites a comparison framing.\n\nNone of this is proven — it's the most plausible reading of why a widely reported, real effect on a 1M-prompt panel doesn't reproduce on a 616-prompt hotel-only one. The honest takeaway for hotels specifically: don't assume the “listicles are dying” narrative applies to your category without checking your own vertical's data.\n\n## Methodology\n\n**Corpus.** The AI Hotel Landscape ChatGPT corpus: 616 destination-hotel prompts (56 destinations × ~11 templates), captured weekly via Bright Data since December 2025. Four weeks analyzed here: 2026-07-27 and 2026-08-03 (GPT-5.5, “pre”), 2026-08-10 and 2026-08-17 (GPT-5.6, “post”) — the two weeks immediately either side of our own observed model transition (see Section 1), giving the “week before/after and two weeks before/after” view.\n\n**Model field.** `fanout_captures.model`, a top-level field in Bright Data's own response envelope (not derived by us). Checked week over week from mid-July through late August to locate the actual transition.\n\n**Fan-out terms.** Each capture's `fanout_queries` array was scanned for whole-word matches of “best”, “top”, “vs”, “comparison”, and “review”. Share of fan-outs = matching fan-out queries ÷ total fan-out queries issued that week (same framing as the original analysis's “share of total” metric). `site:` adoption = share of captures where `has_site_operators` (an existing pipeline flag) is true.\n\n**Citation content type.** Our own classifier, not Peec AI's: a citation is tagged `listicle` if its title or URL matches a “top/best + number” pattern, and `comparison` if it matches “vs”, “versus”, “comparison”, or “alternatives”. Spot-checked against real citation titles before trusting it (examples in Section 3) — precise enough to trust the direction of change, not necessarily calibrated to match any other team's classifier in absolute terms.\n\n**External data.** The GPT-5.6 findings this article tests against are from a publicly shared analysis by Tomek Rudzki, based on Peec AI's data (a 1M-prompt panel, ~180M sources, tracked one week before and one week after the rollout the source dated to July 9, 2026). We did not have access to Peec AI's underlying data or classifier — the numbers attributed to them above are taken from their public post and are shown for contrast, not independently verified by us.\n\n**Access.** Read-only throughout: existing stored fields (`model`, `fanout_count`, `fanout_queries`, `citation_count`, `citations`, `has_site_operators`), no new captures, no writes.\n\n## FAQ\n\nNot necessarily wrong — untested in this vertical until now, and it doesn't replicate here. The original analysis is based on a much larger, more varied 1M-prompt panel across many categories; ours is 616 hotel-only prompts. The honest reading is that GPT-5.6's retrieval changes may be concentrated in categories (software, B2B, comparison-shopping) more exposed to GEO-driven listicle/comparison spam than hotel destination-discovery is.\n\n## Every number here is from our own AI hotel-search captures\n\nAI Hotel Landscape corpus, weekly since December 2025 — CC-BY-4.0. Using this data? A citation or link back to nicolassitter.com is always appreciated.\n\n[Live dashboard](/projects/ai-hotel-landscape/all-models)","author":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com/about","sameAs":["https://www.linkedin.com/in/nicolassitternolleau/","https://github.com/Nicositter88","https://hotelrank.ai"]},"publisher":{"@type":"Person","name":"Nicolas Sitter","url":"https://nicolassitter.com"},"image":"https://nicolassitter.com/api/og/gpt-5-6-hotel-search-changes-2026","mainEntityOfPage":{"@type":"WebPage","@id":"https://nicolassitter.com/research/gpt-5-6-hotel-search-changes-2026"},"tags":["AI Search","Hotels","ChatGPT","GPT-5.6","AEO"],"sameAs":["https://hotelrank.ai/research/gpt-5-6-hotel-search-changes-2026"],"alternateFormat":{"html":"https://nicolassitter.com/research/gpt-5-6-hotel-search-changes-2026","json":"https://nicolassitter.com/api/post/gpt-5-6-hotel-search-changes-2026","rss":"https://nicolassitter.com/rss.xml"},"datasets":[{"name":"summary","contentUrl":"https://nicolassitter.com/data/gpt-5-6-hotel-search-changes-2026/summary.csv","encodingFormat":"text/csv"}]}