A client asked me recently: “If I publish ‘The 10 best CRM tools’ on my own blog and put my CRM at number one, will ChatGPT start recommending us?” He had seen competitors doing exactly that, and he had read that AI tools love “best of” lists. It sounded like a cheap shortcut.
It is a fair question, and the honest answer is: sometimes it works, often it doesn’t, and it can quietly hurt you on Google. In 2026 we finally have real data on both sides. Ahrefs ran a four-month experiment with self-promotional pages, and Lily Ray documented big visibility drops on sites that relied heavily on self-ranking listicles.
Let me walk you through what the evidence says, when a self-ranking list is reasonable, when it backfires, and how to write a comparison page that features your product without misleading anyone.
Why so many brands now rank themselves first in “best” lists
The tactic is old. Software companies have published “best X tools” posts with themselves on top for years. What changed is AI search.
In December 2025, Glen Allsopp at Ahrefs studied 750 top-of-funnel prompts in ChatGPT and found that “best X” blog lists made up 43.8% of all page types cited. In ChatGPT software answers where the list’s publisher was recommended, a self-promotional list appeared in 34% of those responses. On Google, 169 of 250 “best X software” searches (67.6%) showed a list where the author ranked itself first.
Put those numbers together and you can see why marketers got excited. If AI tools cite listicles so often, and many of those listicles are written by the brands themselves, then writing your own looks like a direct line into AI answers.
There is a technical reason behind this, too. AI tools that search the web run several related searches for one question and then keep only a few passages. A page titled “Best CRM tools for small businesses” matches many of those searches neatly, so it often makes the shortlist. I explain that process step by step in retrieval-augmented generation explained.
But being cited and being recommended are two different things. That is where the story gets interesting.
What the Ahrefs experiment found when a brand promoted itself
Between February and May 2026, Mateusz Makosiewicz at Ahrefs published 34 self-promotional pages (listicles, reviews and opinion pieces) on five domains. They promoted two things: Ahrefs’ Brand Radar tool, which is part of a well-known brand, and Ahrefs Evolve, a newer conference with little existing coverage. Ahrefs then analyzed 9,886 AI answers across ChatGPT, Gemini, Perplexity and Copilot.
Here is a selection of the results that matter most for a business owner:
| What was measured | New brand (Evolve conference) | Established brand (Ahrefs tools) |
|---|---|---|
| New mentions in previously empty prompt slots that cited the experiment’s pages | 82% | 6% (94% came from third-party content) |
| Answers that cited the page but did not mention the brand | 43% | 11% |
| Best single-query result | 66.4% of “best SEO conferences 2026” answers | n/a |
| Weaker-fit query | 15.8% of “best marketing conferences 2026” answers | n/a |
A few other findings stood out to me:
- Platforms differed a lot. Copilot picked up Evolve steadily (39% of answers in March, 57% in April, 65% in May). ChatGPT never named Evolve for the tracked query.
- Citations were short-lived. About a quarter of citations appeared once and never came back. Between first and last citation, a page was cited on about one in three eligible days.
- Query fit mattered. The brand did well where it was a natural answer (SEO conferences) and poorly where it was a stretch (general marketing conferences).
The author is honest about the limits: it was a small test, over one four-month period, with two brands both connected to Ahrefs. I treat it as strong direction, not proof.
Cited is not recommended: the gap that matters most
If you remember one thing from this article, make it this line from the experiment:
“Being cited and being recommended turned out to be two very different things.”
— Mateusz Makosiewicz, Ahrefs, Self-Promotional Content Works—Until It Backfires
In 43% of the answers that cited a conference page, the AI used the page as a source but still didn’t mention Evolve. It read the list, took some facts, and then recommended other events. For the established Ahrefs tools that figure was only 11%, because the AI already had plenty of independent evidence that Ahrefs belonged in the answer.
In plain words: your own list can get you into the room, but other people’s opinions decide whether you get picked. The tactic worked best when two conditions were true:
- The AI had a gap in its coverage, so there was little else to cite (a new event, a niche category).
- The brand was a plausible fit for the question, so naming it didn’t look odd.
When a category is crowded with independent reviews, your self-ranking list is one voice among many, and usually not the one the AI trusts most. That is why I always pair this with work on getting into third-party best-of lists.
Which citations you can actually influence
It helps to sort AI citations into two groups. Some come from pages you control or can shape, such as your own site, your profiles and directories. Others come from independent editors, reviewers and communities. Your own list only moves the first group. For an established brand, the second group usually carries more weight, and even unlinked mentions there count. We cover that split in our posts on which AI citations you can influence and AI mentions without links.
How Google’s reviews system looks at your own “best” list
Now the Google side. Many owners don’t realize that a “best X” post is a review in Google’s eyes. Google’s documentation says the reviews system covers first-party content that recommends, gives an opinion on or analyzes something, including ranked lists and head-to-head comparisons.
“The reviews system aims to better reward high quality reviews, which is content that provides insightful analysis and original research”
— Google Search Central, Reviews system
Two details from that page matter here:
- The system primarily works page by page, but sites with a substantial amount of review content may be evaluated site-wide. Publish 300 self-ranking listicles and your whole blog may be judged as a review site.
- It evaluates your own editorial content, not user reviews on product pages.
Separately, Google’s spam policies describe scaled content abuse as many pages generated mainly to manipulate rankings. A template that spits out “best X for Y” for hundreds of variations, each with your product first, is the kind of pattern that invites trouble. Our guide to how Google’s reviews system works explains the background in more detail.
The January 2026 drops: what happened to heavy listicle publishers
In February 2026, Lily Ray published an analysis of several SaaS and B2B sites that lost large amounts of organic visibility in January, shortly after the December 2025 core update. She used the Sistrix US Visibility Index and looked at what the affected sections had in common. A sample of her findings:
| Site type (as described by Lily Ray) | Visibility change | Self-promotional listicles in the affected section |
|---|---|---|
| $8B B2B brand | -49% | 191 among about 30,000 blog articles |
| SaaS company | -43% | 228 among about 2,780 guides |
| B2B SaaS | -38% | 267 among 2,790 blog pages |
| Popular SaaS product | -34% | 340 blog listicles |
Source: Lily Ray, “Is Google Finally Cracking Down on Self-Promotional Listicles?” (February 3, 2026).
Search Engine Land’s coverage added that many of these articles had been lightly refreshed with “2026” in the title with little real updating, and that Google had not confirmed any update. Both Lily Ray and Search Engine Land are careful to say other factors, such as scaled AI content and artificial date refreshes, may have contributed.
So this is correlation, not proof. But Lily Ray makes a point I agree with completely:
“Using “best” in the title implies an objective, independent evaluation.”
— Lily Ray, Is Google Finally Cracking Down on Self-Promotional Listicles?
When the publisher is also the winner, the evaluation isn’t independent, and both readers and algorithms can tell.
When a self-ranking list makes sense: a decision matrix
Based on the evidence above, here is how I would decide. It isn’t a law, just a sensible starting point.
| Your situation | AI upside | Google risk | My suggestion |
|---|---|---|---|
| New product or niche category with little coverage | High | Low if it’s a single, honest page | Publish one clearly disclosed comparison |
| Established brand in a crowded category | Low (third-party content dominates) | Medium | Skip the self-ranking list; invest in reviews and third-party lists |
| Dozens of “best X for Y” variations planned | Short-lived | High (site-wide review evaluation, scaled content) | Don’t. Consolidate into a few genuinely useful comparisons |
| You can’t name a fair reason a competitor would beat you | Low | Medium | Don’t publish until you can |
| Local service business (“best plumber in…”) | Low | Medium, plus reputation risk | Focus on reviews and local listings instead |
How to write an honest comparison that still features your product
If you decide to publish, do it in a way you’d be happy for a competitor’s customer to read. Here is the process I recommend:
- Drop “best” from the title if you rank yourself. Use “X vs Y vs Z” or “Alternatives to X compared by the X team.” It sets the right expectation.
- Disclose in the first paragraph. “We make Product A. Here is how it compares with B and C, and when we think you should choose them instead.”
- Pick real competitors fairly. Include the ones buyers actually consider, not weak ones that make you look good.
- Say who each option is best for. Give every competitor at least one honest “choose this if” scenario. AI tools love clear fit statements, and readers trust them.
- Show your method. Explain how you compared: features tested, pricing pages checked on a given date, sources used.
- Fact-check every competitor claim and note the date. Wrong competitor prices are the fastest way to lose trust.
- Update for real, not for the date. Change the year only when you’ve actually re-checked the content.
The Ahrefs author’s advice fits in five words I like a lot: “Fill the gap accurately; don’t manufacture a ranking.” Our guides to building comparison pages and listicle SEO go further into formats.
A worked example: two ways to write the same page
Imagine a hypothetical bookkeeping app for freelancers. Version one is titled “The 12 Best Bookkeeping Apps for Freelancers in 2026,” ranks itself first, gives competitors one line each, and was last edited 18 months ago apart from the year in the title.
Version two is titled “Bookkeeping apps for freelancers compared: us vs 4 alternatives.” It discloses ownership, compares five tools on six criteria in a table, says plainly that a competitor is better for freelancers with employees, and shows the date each price was checked.
Version one may get cited for a few weeks. Version two is far more likely to be trusted by readers, survive a reviews-system evaluation and give an AI tool a clear, accurate fit statement: “best for solo freelancers without staff.” That is the sentence you actually want repeated.
Measuring whether your list helped, without fooling yourself
Self-promotional pages are easy to “prove” with a lucky screenshot. Here is how to measure them honestly:
- Track mentions and citations separately. Count how often you are named, and how often your page is linked. A citation without a mention is not a win.
- Use a baseline. Record your share of answers for 20–50 prompts for four weeks before publishing.
- Track for at least 8–12 weeks. The Ahrefs experiment showed citations coming and going; a two-week spike tells you little.
- Watch Google too. Check Search Console for the page and for the whole blog section after core updates.
- Look at business results. Did AI referral visitors from that page sign up or buy? Our guide to tracking AI traffic in GA4 shows how.
For tools that do prompt tracking, see our comparison of AI visibility tools.
A pre-publish checklist for brand-written “best” lists
Before you hit publish, run through this list. If you can’t tick most boxes, rework the page.
- The title doesn’t claim independence you don’t have.
- Ownership is disclosed near the top.
- Real, relevant competitors are included, with at least one honest reason to choose each.
- Every competitor fact has been checked and dated.
- The method of comparison is explained.
- There is first-hand evidence: screenshots, tests, real usage notes.
- No negative claims about competitors you can’t support.
- This is one of a handful of comparison pages, not one of hundreds.
- You have a plan to update it with real changes.
Questions brands ask about ranking themselves first
Will ChatGPT recommend my product if I rank it first on my own blog?
Sometimes, mainly if your category has little other coverage. In the Ahrefs test, the page was often cited while the AI recommended someone else.
Is a self-promotional listicle against Google’s rules?
Not in itself. But Google’s reviews system rewards insightful analysis and original research, and publishing many thin, templated lists can look like scaled content abuse.
How long do AI citations from these pages last?
Often not long. About a quarter of citations in the Ahrefs experiment appeared once and never returned.
Should I delete the self-ranking lists I already have?
Not automatically. Review them. Merge thin ones, rewrite the useful ones as disclosed comparisons and remove the ones that add nothing.
Is it better to get into other people’s lists?
For established brands, yes. In the Ahrefs data, 94% of new Ahrefs tool mentions came from third-party content, not the brand’s own pages.
Where to put your effort if you sell in a crowded category
To sum it up: a self-ranking list can help a new or niche brand fill a gap in AI answers, for a while. For an established brand in a busy market, independent reviews and third-party lists do most of the work. And on Google, a large pile of “best” lists with your name on top is a real risk.
My suggestion is simple. Write a few honest, disclosed comparisons, then spend most of your energy earning mentions on sites buyers already trust. If you’d like help with both, our content writing team builds comparison pages that hold up, and our SEO services cover the off-site work. If you want to see where this fits in a wider plan, read our AI search strategy framework. Get a free quote and tell us about your category.
Further reading on self-promotional listicles
- Content Engineering: What It Is and How to Start
- Does Content Length Matter for AI Citations?
- Reddit Threads in Google Results: How to Show Up Where They Rank
Sources for this self-promotional listicles guide
- Ahrefs – Self-Promotional Content Works—Until It Backfires (AI SEO Experiment) (Mateusz Makosiewicz, July 6, 2026)
- Ahrefs – “Best” lists research (750 ChatGPT prompts) (Glen Allsopp, December 4, 2025)
- Lily Ray – Is Google Finally Cracking Down on Self-Promotional Listicles? (February 3, 2026)
- Search Engine Land – Google cracking down on self-promotional “best of” listicles (February 4, 2026)
- Google Search Central – Reviews system (updated December 10, 2025)
- Google Search Central – Spam policies for Google web search (updated August 28, 2026)
- Ahrefs – AI Search Strategy (July 24, 2026)
