Google Told Brands to Stop the "Top 10" Listicle. We Checked 40 Searches, and Lists Still Held 95 of 332 Top Spots
A Google ads executive says generic "Top 10 Things to Consider" posts are why brands disappear in AI search. On November 22 we checked 40 Google searches: numbered lists ranked in the top 10 for 35 of them and made up a quarter of the sources AI Overviews cited.
Google is telling brands to stop publishing generic "Top 10" listicles, but in 40 Google US searches we ran on November 22, 2026, numbered listicles still held 95 of the 332 top 10 organic spots, so the listicles SEO question is about content, not format.
At least one numbered list ranked in the top 10 for 35 of the 40 searches, and a list ranked first for 12 of them.
38 of the 40 searches showed an AI Overview, and 35 of those 38 cited at least one numbered list; lists were 94 of the 357 web sources cited, and 23 of those 94 were YouTube videos.
The "stop the listicle" advice comes from Think with Google, the ads side of the company; Google Search's own guide to helpful content does not mention lists and says Google has no preferred word count.
Our view: keep the list format where readers scan, and rewrite every item that a competitor could have written without testing, pricing or using the thing.
Google is telling brands to stop publishing generic "Top 10" listicles, yet its own results are still full of them. On November 22, 2026, we ran 40 Google searches in the US and counted the numbered lists: they held 95 of the 332 top 10 organic spots, ranked first for 12 of the 40 searches, and made up 94 of the 357 web sources cited in AI Overviews. For anyone weighing listicles and SEO, that means the format is not what Google penalizes. Generic content is the risk, in any format.
What did Google say about listicles?
The advice comes from Think with Google, the part of Google that speaks to advertisers. In an October article on brand authority in AI search, Brendon Kraham, Google's VP of Search and Commerce in Global Ads Solutions, tells brands to build content on "unique expertise." His example of what to replace: "Instead of publishing a generic 'Top 10 Things to Consider,'" a brand should share deeper material that only it can make. The article also tells marketers to measure plans "based on real business goals like leads or sales," not proxy metrics.
The line spread this week after Think with Google's account on X turned it into a short instruction to CMOs, as reported by Search Engine Journal on November 19 and 20: stop the "Top 10" listicle, lead with first-hand expertise. We could not load the X post itself, so we rely on that report for its wording and on the Think with Google article for the substance.
Google Search's own guidance says less. Its page on creating helpful, reliable, people-first content does not mention lists at all. It asks whether content offers "original information, reporting, research, or analysis" and first-hand expertise, and on length it is blunt: "Google has a preferred word count? (No, we don't.)"
How we tested
We used the DataForSEO SERP API to run 40 Google searches on desktop, US, English, between 1:40 and 1:42 p.m. Eastern (18:40 to 18:42 UTC) on November 22, 2026, with AI Overviews loaded. Two sets of 20:
"Best" searches, where lists are the expected answer: best CRM software, best SEO tools, best running shoes, best travel insurance, best payroll software and 15 more.
Advice searches, the "things to consider" kind Kraham's example describes: things to consider when buying a house, how to choose a web designer, how to choose a dentist, tips for selling a house, Google Ads tips and 15 more.
We counted a result as a numbered listicle when its title promises a count of items: "Top 10", "12 Steps to Buy a Home", "The 15 Best Running Shoes of 2026". Counts of time ("in 10 Minutes") and counts of things tested ("400+ Beds Tested") did not count. We judged titles only, not the pages behind them. From the AI Overview sources we left out 22 Google product cards, which link to shopping listings rather than web pages.
How many top results were listicles?
95 of 332 top 10 organic results (29%) were numbered lists: 47 of 168 on the "best" searches and 48 of 164 on the advice searches. The advice searches, the ones Google's example targets, had as many lists as the shopping ones.
35 of 40 searches had at least one list in the top 10: 19 of 20 "best" searches and 16 of 20 advice searches.
A list ranked first for 12 searches, 5 "best" and 7 advice, and lists held 32 of the 120 top three spots.
5 searches had no numbered list in the top 10: best accounting software for small business, how to choose a marketing agency, things to consider when buying a laptop, how to choose a wedding venue and how to choose a web hosting provider.
The lists that do well are not all generic. RunRepeat's "7 Best Running Shoes in 2026" ranked first for its search, and AI Overviews cited two Wirecutter lists (four headphones, two password managers) and Zillow's "14 Tips for Selling Your Home Fast and for More Money", from publishers that test products or have their own housing data. But plenty of template lists rank too: on "best payroll software", one software directory held five of the ten spots, four of them "20 Best" pages that differ mainly by category.
Do AI Overviews cite listicles?
Yes, often. 38 of the 40 searches showed an AI Overview (best robot vacuum and best standing desk did not), and 35 of those 38 cited at least one numbered list. Every one of the 20 advice AI Overviews cited a list. In all, lists were 94 of the 357 web sources cited, about 26%: 27 of 126 on the "best" searches and 67 of 231 on the advice searches.
23 of those 94 list sources were YouTube videos, titles such as "Top 5 Project Management Software in 2026". The others came from a mix of large brands (NerdWallet, Zillow, Indeed, Rocket Mortgage, Salesforce, American Express, WIRED) and small business blogs, credit unions and colleges. AI Overviews cited lists as readily as Google ranked them.
Are listicles good for SEO now?
Our view: the format is fine, and Google's advice is right about the content inside it. A numbered list is how people scan "best" and "how to choose" answers, and Google's results on November 22 show that clearly. What Kraham is pointing at, and what Google Search's helpful content questions say in other words, is a list anyone could have written without doing anything. Those pages have the weakest claim to be cited when an AI answer can summarize ten of them at once.
What we would do with a site's lists this week:
Keep the format, rewrite the items. For each item, ask whether a competitor could have written it without testing, pricing or using the thing. If yes, add what only you have: a measurement, a price you paid, a photo you took, a result from your own work.
Put the method near the top. "We tested 12 for two weeks" or "picked from 300 client calls" tells readers and AI answers why your list is different. Only say it if it is true.
Use a number only when it is real. "7 Best" should mean seven you can defend, not a count chosen to beat "6 Best".
Don't delete lists in a hurry. The November 2026 core update has been rolling out since November 10 and was still listed as in progress on Google's Search Status Dashboard on November 22. Rankings are still moving; judge pages after it finishes, against your own Search Console data.
Check your own searches
This script runs the same check on any searches you choose. It needs a DataForSEO account (each search costs a fraction of a cent) and Node 18 or later.
// node listicle-check.mjs "best crm software" "how to choose a crm"const isList = (t = '') => { if (/\btop\s+\d{1,3}\b/i.test(t)) return true; for (const m of t.matchAll(/(^|[^$\w.])(\d{1,3})\+?(?![\w.%"”])\s+((?:[\w'&-]+\s+){0,3}?)([A-Za-z'-]+s)\b/g)) { const before = t.slice(0, m.index + m[1].length).toLowerCase(); if (Number(m[2]) < 2 || /(tested|tried|reviewed|over|than|of)\s*$/.test(before)) continue; if (/^(minutes?|seconds?|hours?)\b/i.test(m[3] + m[4])) continue; if (/^\s*tested\b/i.test(t.slice(m.index + m[0].length))) continue; return true; } return false;};for (const keyword of process.argv.slice(2)) { const res = await fetch('https://api.dataforseo.com/v3/serp/google/organic/live/advanced', { method: 'POST', headers: { Authorization: `Basic ${process.env.DATAFORSEO_API_KEY}`, 'Content-Type': 'application/json' }, body: JSON.stringify([{ keyword, location_code: 2840, language_code: 'en', device: 'desktop', depth: 10, load_async_ai_overview: true }]), }); const items = (await res.json()).tasks?.[0]?.result?.[0]?.items ?? []; const organic = items.filter((i) => i.type === 'organic'); const refs = (items.find((i) => i.type === 'ai_overview')?.references ?? []).filter((r) => !/google\.com\/search\?/.test(r.url ?? '')); console.log(`\n${keyword}: ${organic.filter((i) => isList(i.title)).length} of ${organic.length} top results are numbered lists; ` + `${refs.filter((r) => isList(r.title)).length} of ${refs.length} AI Overview sources`); for (const i of organic) console.log(` ${isList(i.title) ? 'LIST' : ' '} ${i.rank_group}. ${i.title}`);}
DATAFORSEO_API_KEY is the Base64 of your API login and password. Change location_code (2826 is the UK, 2356 India, 2036 Australia) to check another country. When we reran "how to choose a crm" a few minutes after the main test, it showed 2 lists in 8 results and 1 in 10 AI Overview sources, close to but not the same as the first run.
What we don't know
Titles, not pages. We counted what titles promise. A titled list can be thorough and first-hand, and an untitled page can be generic.
One afternoon, one country, desktop. Results change between searches, users and countries, and we ran each search once.
Correlation, not cause. Lists ranking well doesn't show Google rewards the format; many lists rank because their publishers are strong.
The core update. With the November 2026 core update still rolling out, these results may shift. We will rerun the same 40 searches after Google says it is complete. For the update itself, see our core update coverage, and for what AI answers quote, how to rank in AI Overviews.
Kiran runs Kihan Marketing, a web design, SEO and ads agency, and manages dozens of websites across real estate, home services and healthcare. He writes about SEO, AI search and paid ads news, and what it means for the people who run websites.