The worst SEO strategies in 2026

Where this list comes from

This article is based on an episode of Luis M. Villanueva’s SEO podcast with Alex Novoa, published in January 2026. They put together the six worst SEO strategies you can follow this year, based on habits they kept seeing during 2025: websites and brands still doing things that made sense three, four or five years ago and don’t anymore.

The reason behind almost every point is the same. Search has changed: it’s no longer enough to rank in Google. You also have to appear and be visible inside the answer an AI gives, whether that’s Google’s AI Overviews and AI Mode, Gemini or ChatGPT. If you want the basics of that, I wrote a guide on how to increase your visibility in ChatGPT, Gemini and other AI.

The hosts agreed on five of the six points. The sixth (updating only the date) ended in a friendly disagreement, so I present it the way they did: as a controversial one you can test yourself. I’ve added my own comments along the way.

1. Link building based on vanity metrics (DR and DA)

The first bad strategy is the link building many people still do the way it was done years ago: buying links in bulk from any outlet, choosing them by their DR (Domain Rating, from Ahrefs) or DA (Domain Authority, from Moz), without looking at whether the outlet covers your topic, your market or your communication strategy. Those are third-party metrics. They estimate how “strong” a site’s links are, but Google doesn’t rank you by them.

Thinking in DR and DA only makes sense if the only thing you care about is the link and the “strength” it passes. The hosts never chose links that way, even back when it seemed to work. And Google is clear about the extreme version: its spam policies define link spam as creating links “primarily for the purpose of manipulating search rankings”, and list “buying or selling links for ranking purposes” as an example.

What to do instead: digital PR

Instead of asking “where do I put my link?”, ask what your brand’s message is, where you need to be, and what has to be said there. The goal is that when someone asks an AI or a search engine about the problem you solve, not only your website appears, but also other websites that mention you as the solution.

• Choose outlets by topic and audience, not by a number on a tool.
• Align every placement with your brand message: the same story, told in the places your customers read.
• Mention a real expert from your company by name, for example “Ana López, payments expert at YourBrand”. It builds that person’s reputation and gives AI answers a clear source to cite.

The debate: who should write the article?

Here the hosts disagreed. One argued it matters more and more that you write the article the outlet is going to publish, instead of sending four or five guidelines and letting them write it. The other prefers to explain exactly who you are and what you want to communicate, hand over a draft, and let the outlet write it in its own words and tone, the way a tech reviewer talks about a product in their own voice. They ended up agreeing that the second option is ideal, but that most outlets don’t work like that.

My take: if the outlet has a real editor with a recognizable voice, let them write it. If it’s a site that publishes whatever you send, at least make sure what you send is good, specific and useful, because it will carry your brand’s name.

2. Ignoring searches for your brand

Many SEO strategies focus only on informational or transactional searches close to the business, and forget the searches for the brand itself. That’s a mistake, because more and more people talk to an AI, get a list of options, and then go to Google to search for the brand they were recommended: “YourBrand reviews”, “YourBrand + product”, and so on.

That results page is your brand SERP, and it’s the last step before someone trusts you and buys. If somebody is speaking badly about you there, or a mistake is showing up, you’re losing sales you never see.

A real case: the competitor inside your own search

One of the hosts told the story of a client with a strong brand: when you searched its name, the second result was a competitor’s article comparing the two. Those comparison pieces work very well, and that’s exactly the problem. Nobody wants a competitor sitting right under their own homepage.

What I like about this point is that it flips the usual approach. Most SEO strategies start by trying to capture new traffic. But people searching your brand already convert much better, and fixing what appears for that search usually costs less than winning new traffic from zero.

How to check your brand SERP

1. Search your brand name on Google and look at every result on the page, not just the first one.
2. Search your brand + “reviews” and see what people find.
3. Search your brand + each type of product or service you sell.
4. Note any negative content, any competitor comparison and any outdated profile.
5. Create or claim profiles on third-party review sites such as Trustpilot. The hosts warned they can be gamed, but Google looks at them and uses them to enrich your knowledge panel.
6. Ask ChatGPT and Gemini the same questions and compare what they say about you.

Don’t forget your own pages are part of that SERP: the title and description Google shows for your brand come from your website. If they’re missing or say “Home”, that’s the first thing to fix. I explain why in the SEO mistakes I see most often on business websites.

3. 3,000-word “ultimate guides” full of filler

For years the formula was to write a giant guide to cover as many searches as possible. The result: endless articles where the answer to the reader’s question is in the fourth or fifth paragraph, after a long introduction nobody asked for. In the episode they say it plainly: this doesn’t work anymore, and if it still works for you, you don’t have much time left.

Google itself asks content creators whether they are “writing to a particular word count” because they’ve heard Google has a preferred one. It doesn’t.

The fix: content that’s easy to split into chunks

AI systems don’t read your article from top to bottom like a person. They take pieces of it (chunks) and use the ones that answer the question best. So the most effective change, and the one the hosts liked most, is to work on the guides you already have:

1. Remove the filler. If a paragraph doesn’t help the reader, it goes.
2. Move the answer to the top. The first paragraph should solve the biggest problem the user had when they typed that search.
3. Structure it as problem → solution, again and again, with clear headings.
4. Use lists and comparison tables wherever they make the content easier to scan.
5. Write for a real person: their pain points and their context, not a list of keywords.

One of the hosts said the most significant change he made on a project was simply asking: “when the user lands here, which paragraph solves their problem?”, and moving that paragraph to the top.

Practical beats theoretical

Their example: instead of writing “the complete guide to robots.txt”, publish how to check your robots.txt with this tester, and include the tester. ChatGPT can explain what a robots.txt is in seconds, but people will come to your page to use the tool. That’s what makes the content unique.

It’s the same idea behind our basic SEO recommendations: a checklist you can tick and use to track your progress, not just something to read.

4. Hiding the author

Signing articles as “admin” or leaving them anonymous. Both hosts admit you can’t measure how much an author byline affects rankings. But they agree on what matters: if someone needs to trust your brand before buying from you, “admin” gives them zero reasons to.

Google is direct about it: it “strongly encourages adding accurate authorship information, such as bylines to content where readers might expect it”, and among the questions it asks about content is “Do pages carry a byline, where one might be expected?”.

Author pages: rank for the expert’s name

One of the hosts is skeptical that an author’s expertise weighs in the algorithm, but still recommends giving each author their own page on your site, especially if that person has authority, because that page can start ranking for their name. His exaggerated example: if Messi collaborates with you, give him an author page, and when people search “Leo Messi” your author page has a chance to appear.

• Give every author a page with a photo, a real bio and links to their profiles.
• Link each article to its author’s page.
• If you have a writing team and don’t want individual names, signing as “Editorial team” is better than “admin”.
• Please don’t sign articles as “AI”.

That’s what we do here: every article links to its author’s profile, with a photo, a bio and links to social profiles.

5. Optimizing for informational traffic that AI already answers

There are still sites publishing “what is X”, “what is Y”, “what is Z”. If any AI can write it, it’s a commodity: available to everyone, so it doesn’t set you apart. Neither do articles where you only add information and your point of view is nowhere to be seen.

What does work, according to the episode:

• Practical content, like the robots.txt tester example.
• Unique data that nobody else has: your own measurements, your own cases.
• Comparisons, reviews and opinion pieces with a well-defined point of view, even polarizing ones.
• Topics that require human judgment, where people don’t trust a generic answer. Their example: how to recover after a hair transplant. A definition of “mitochondria” is something else.

The question to ask before writing anything: why would someone visit my page instead of settling for ChatGPT’s answer? If you don’t have an answer, don’t write it.

Is the traditional blog over?

One host threw out a more provocative idea: in 2026 he would put more of the writing effort into product pages and category listings than into the blog. A product page is transactional; a blog post is usually informational. The other host replied that the blog still has room, especially for content that needs human judgment, or when you don’t have enough authority yet to rank for anything transactional. They met halfway: there’s still room in informational content, but “working on content”, which used to mean blog posts, will mean product pages more and more.

6. (Controversial) Updating only the publication date

It’s been said a lot over the past year that when AI tools need up-to-date information they search the web and look at the date of each result. From there came a tempting shortcut: change the date of an old post to make it look recent, without touching the content.

The argument against

Changing a post you haven’t touched in three years from 2024 to December 2025 won’t do anything on its own. Besides, for evergreen topics (their example: how to do a SWOT analysis) ChatGPT doesn’t even need to search the web, so the date doesn’t come into play.

The argument for

Google Search may not care much about dates (it can keep an article from 2019 in its top 5), but when an AI does search for queries that need fresh information, it might get, say, 20 results from a search engine. If three of them have more recent dates, it may favor those. The host defending this cited cases in recent months where simply updating an article seemed to make ChatGPT give it more weight.

The verdict: test it

Neither of them can prove causality, and in SEO with AI that’s very hard today. So they changed the premise of the episode on the fly: five strategies you shouldn’t follow, and one you can test. The pro side’s logic: if it costs almost nothing and isn’t going to be a catastrophe, try it. The risks they pointed out: a system could compare cached versions and see nothing changed, and if outdated content ranks better because of a new date, a bad user experience can end up sinking the article.

Keep in mind that Google lists this exact practice among the questions it uses to judge content: “Are you changing the date of pages to make them seem fresh when the content has not substantially changed?”. My recommendation is the one they also closed with: if the content needs updating, update it for real, use the update to prepare it for AI answers (everything in point 3), and then change the date. On this site, the “updated” date only changes when the content does.

Summary: what to do instead

The hosts closed with an idea I agree with: SEO in 2026 is more sophisticated, and that raises the barrier to entry. The shortcuts that used to work are the ones that stop working first.

Before you work on any of this, make sure the basics are in place: titles, meta descriptions, H1s, alt text and no broken links. basicseo.net crawls your site for free and shows you page by page what’s missing. Request free access.

FAQ

Is link building dead in 2026?

Buying links in bulk based on DR or DA no longer makes sense. Getting mentioned on relevant outlets still does, as part of a digital PR strategy: the right outlets for your topic, a clear brand message and real experts from your company.

What is a brand SERP?

It’s the results page people see when they search your brand’s name, with or without words like “reviews”. It’s often the last step before someone buys, so you should check what appears there regularly.

Does an author byline improve rankings?

Nobody can measure its direct effect. But Google recommends accurate bylines, an author page can rank for the author’s name, and “admin” gives readers no reason to trust you.

Should I change the date of my old articles?

Only if you really update the content. Changing the date of an untouched article is the one point the podcast hosts didn’t agree on, and Google lists it among the practices it looks out for.

Sources

• Las 6 peores estrategias SEO para 2026 (podcast, in Spanish) (Luis M. Villanueva and Alex Novoa, YouTube)
• Creating helpful, reliable, people-first content (Google Search Central)
• Spam policies for Google web search (Google Search Central)