Google on Weak 'Best Of' Lists: What It Said, and How to Write an Honest One

Short answer: in a statement reported by The Verge in April 2026, Google said it knows about low-quality "best of" listicles, applies protections against manipulation in Search and in Gemini, and works to combat that kind of abuse. It didn't name an update or publish a new rule. A weak list is one that claims independence it doesn't have, usually by ranking its own publisher first on criteria it made up. An honest list dates every figure, says what the order is based on, and doesn't put itself at the top.

Where the statement comes from, and when

The statement we can trace was an emailed statement from Google spokesperson Jennifer Kutz to The Verge, published in The Verge's article on the AI SEO industry on 2026-04-06 (read 2026-09-24). Search Engine Land covered it the same day (read 2026-09-24).

Two cautions about the wording:

• The Verge paraphrased most of it. In the article, the spokesperson "noted the company is aware of the low-quality listicle content." Everything about combating abuse in Search and Gemini is The Verge's paraphrase of the email, not a quote. We haven't seen the email itself.
• The guidance was the usual guidance. The statement repeated Google's long-standing advice: make content for people, and make sure search systems can understand it. That's not a new rule.

We looked for a more recent official Google statement specifically about "best of" lists and didn't find one. If you've seen claims about a newer statement, check them against a source you can open.

What makes a "best of" list weak

The Verge article walks through examples of a pattern anyone researching software has run into: a company publishes "the best tools for X," compares a dozen products, and names its own product the winner. Search Engine Land's summary of the pattern (read 2026-09-24) is a good checklist:

• the publisher includes competitors it hasn't tested;
• it uses a subjective or invented scoring system;
• it ranks itself first;
• the page implies independence or hands-on evaluation that didn't happen.

Add a few more we'd flag from reading these pages ourselves:

• Undated figures. A competitor's price with no read date looks equally current whether it was checked this morning or last year. Pricing pages change without notice.
• "Updated for [year]" with nothing updated. A new year in the title and the same body underneath.
• No stated basis for the order. If a reader can't tell why tool three sits above tool four, the order is an opinion presented as a finding.
• Descriptions no vendor page supports. "Best-in-class support" and "most accurate data" are claims. If the author didn't measure them and the vendor doesn't state them, they're filler.

None of these need a regulator or an algorithm to recognize. A careful reader spots them in a minute, and if AI engines are going to cite these pages, a careful reader is the right standard.

How to write an honest list

This is how we wrote our own AI visibility tools comparison. We're using it as a worked example because it's the one whose choices we can explain, not because it's the only way to do this.

1. Put a date on every external figure

Every price on that page carries the date it was read, and the page opens by saying so. When a vendor's pricing page wouldn't show a price on the day, the older reading stays in with its older date, labelled as older. The date is what lets a reader judge whether a figure is still likely to hold.

Do this: next to each price or feature claim, write where it came from and when you read it. When you re-read, update the date. Don't backdate.

2. State what the order is based on

That comparison isn't ranked. It says so in its second paragraph and gives the reason: we didn't run the tools against each other, so we have no evidence that would support a quality ranking. The order is each vendor's lowest published monthly price, ascending. That's a fact about their pages, not a judgement about their products.

Do this: if you ran a real test, describe the test and let the order follow the results. If you didn't, pick an order you can defend as a fact (price, alphabetical, date founded) and say which one you used.

3. Don't put yourself first

We're one of the tools in that category, so TriRank sits in its own section at the end, outside the order, with a plain disclosure at the top of the page. Whether our product would come out "best" isn't a question that page can answer, and pretending otherwise would undo everything else on it.

Do this: disclose the relationship before the list starts. Put your own product in a separate section, or order it by the same fact-based rule as everyone else, without an exception in your favour.

4. Write "not stated" instead of guessing

Where a vendor's own pages don't say whether their tool does something, that comparison says not stated, and defines that as "we couldn't find it on their pages," not "the product can't do it."

Do this: only describe what you can point to. An empty cell with an honest label is more useful than a confident guess.

5. Link the source

Every row in that comparison links the vendor page it was read from. A reader who cares about a number can check it in one click, and that's what makes the date worth anything.

Why this matters for AI answers too

The statement named Gemini alongside Search. More generally, lists like these get cited: an assistant answering "what are the best tools for X" often pulls from exactly these pages. A misleading list does more damage when an assistant repeats it as a neutral answer.

If you publish comparisons, it's worth checking which of your pages AI engines actually cite. Our free AI visibility checker shows, per engine, whether your domain is among the cited sources. To see which pages are cited instead, ask the assistant the same question and read its source list. If a competitor's self-ranked list is being cited for your category, you probably can't get it removed. What you can do is publish a page that deserves the citation more.

FAQ

Did Google announce a penalty for 'best of' lists?

Not in the statement we could trace. In April 2026, in an emailed statement reported by The Verge, a Google spokesperson said the company is aware of low-quality listicle content, applies protections against manipulation in Search and Gemini, and works to combat that kind of abuse. That is a stated position, not a named update or a published rule, and we haven't found an official announcement of a specific ranking change.

Is it wrong to include my own product in a comparison?

No. A comparison that includes your own product can be useful to readers. What makes a list weak is presenting it as independent when it isn't, ranking yourself first on a scoring system you invented, and describing competitors you never used. Disclose the relationship, say what the order is based on, and keep your own product out of the ranking if you have no fair way to rank it.

Does this matter for AI answers or only for Google rankings?

Both, according to the statement: it named Search and Gemini. More broadly, a list that AI engines cite becomes part of the answer people read, so a misleading list can mislead through an assistant as well as through a results page. Writing the list honestly is the part you control either way.