Posts

3.5 Flash-Lite Rolling Out In Google Search

Google has released three new AI models today Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber and one of those models is rolling out now for Google Search. 3.5 Flash-Lite is Google’s “fastest, most cost-effective 3.5-class model, delivering 350 output tokens per second according to the Artificial Analysis Index, also significantly outperforming prior Flash-Lite generations in agentic workflows,” the company announced.

Google said 3.5 Flash-Lite is rolling out now for Google Search. Where in Google Search? Google did mention in the blog post that it is used for agentic search. But it might be used for Google AI Overviews and Google AI Mode as well.

Agentic Search. Google announced at Google I/O back in May about Google Search’s information agents and improved agentic experiences. “We’re entering the era of Search agents, where you can easily create, customize and manage multiple Al agents for your many tasks, right in Search,” Liz Reid, the head of Google Search said a couple of months back.

Rollout. 3.5 Flash-Lite is rolling out for everyone in the Gemini app and also within Google Search, the company announced.

Why we care. Google will continue to improve its AI models; the latest improvements are on its lower-end models but also faster models. Those faster models are often uses for Google Search and you soon may see those changes in Google AI Overviews, AI Mode and the agentic experiences within Google Search.

Read more at Read More

How to build curiosity into high-performing social ads

How to build curiosity into high-performing social ads

Capturing attention is no longer enough to create high-performing social ads. You also need to create curiosity that keeps people watching.

Stopping the scroll and hooking viewers in the first three seconds still matter. But as video-heavy platforms like Meta and TikTok continue investing in AI-powered delivery systems that optimize for engagement, watch time, and downstream conversions, what happens next matters, too.

Attention is the moment someone notices your ad. Curiosity is the reason they stick around instead of scrolling. Understanding that difference can help you build creative that earns both attention and consideration.

Attention is only the beginning

Many creative briefs I’ve seen begin with questions like “what’s our hook?” or “how do we stop the scroll?” While these questions matter, they focus almost entirely on the first few seconds of the ad, not how to maintain that fleeting attention in the seconds that follow.

Every piece of content on social media is competing for attention, and platforms are becoming increasingly better at predicting what users are most likely to engage with next.

A flashy visual, a bold headline, or an unexpected opening can earn a pause, but if the rest of your creative immediately answers every question or transitions into a predictable sales pitch, people will scroll onward.

High-performing creative doesn’t just interrupt the scroll. It creates a gap in information that people want to close. That’s curiosity.

Dig deeper: How to measure paid social’s impact on paid search performance

See exactly how your competitors win.

Uncover the keywords, ads, landing pages, and strategies driving your competitors’ paid search success—and find your next opportunity to outperform them.

Analyze your competitors

Curiosity keeps the algorithms working for you

As user behavior changes and algorithms evolve, platforms increasingly reward signals that indicate genuine user interest. These signals include:

  • Longer watch times.
  • Meaningful engagement (e.g., saves and sharing videos via DMs).
  • Higher video completion rates.
  • Clicks that occur after viewing the content.
  • Replays.

Platforms use these signals to identify content that people choose to spend their time with. Human curiosity naturally produces these signals because it encourages viewers to keep watching rather than scroll until something else grabs their attention.

The highest-performing creator ads often don’t feel rushed. They gradually reveal information through conversation, yapper-style content, demonstrations, storytelling, and personal experiences.

While the goal may be to secure a conversion quickly, sometimes slow and steady wins the race. A thoughtfully curated approach invites viewers to discover something new rather than asking them to purchase right away.

The rise of curiosity-driven creators

One recent creative shift is the growing popularity of longer-form, conversational creator ads called “yapper ads.” These ads are longer, less polished, more authentic, often unscripted, and sometimes don’t even mention the product immediately.

They seem to violate nearly every traditional best practice, but they perform well because they stand out, feel like conversations rather than commercials, and create curiosity throughout the video. They often address what happened next, why something solved a problem, and whether something would work well for the viewer.

Creators who do this well reward the attention they’ve earned with more interesting information throughout the content, and brands that boost these ads are seeing impressive results.

A screenshot of a Meta video ad from Gratsi Wine and creator, Kayce Smith
Source: Gratsi Meta Ads

Dig deeper: Why ugly ads outperform polished creative and how to test them

Don’t improvise curiosity, build it

Curiosity should be intentionally designed because most creators can’t naturally create it on the fly. The creative strategy behind these successful ads is to delay the obvious and gradually reveal information.

To do this, create open loops with a surprising statement, an unexpected result, or a relatable problem that encourages viewers to stay for the outcome. Demonstrate something rather than just describing it, or show a transformation rather than simply listing features.

While this type of content is strategic, it can feel more human and natural by telling stories, answering questions along the way, and perhaps even jumping between thoughts. That conversational rhythm creates authenticity, trust, and ultimately, the momentum you’re aiming for.

Curiosity builds trust while clickbait breaks it

There’s an important distinction between curiosity and manipulation. Clickbait typically withholds information without delivering meaningful value, whereas curiosity rewards attention by eventually answering questions in a way that satisfies the viewer.

Every promise should be fulfilled, and every story should reach a worthwhile conclusion. Don’t trick people into watching longer and create an unpleasant experience. Make every second genuinely worth the viewer’s time because prospective customers will recognize your intent, and their behavior will send signals to the platforms.

Dig deeper: Why search ROAS depends on paid social more than you think

Optimize beyond the first three seconds

The advice has long been to obsess over the first three seconds of every video, and while that advice isn’t wrong, it may be incomplete. Instead of asking only, “Will someone stop scrolling?” creative teams should also ask, “Why would someone keep watching?”

Considering both can improve how creative is written, filmed, edited, and evaluated. Every transition should introduce another point of interest, and every scene should answer one question while creating another without overwhelming the viewer. It should also deliver on the promise to close earlier loops. Don’t just earn attention. Sustain it.

Creative is what you still control

AI has made audience targeting on social media advertising platforms smarter and easier than ever. Campaign setup continues to become more automated and streamlined, while machine learning handles more optimization.

Creative strategy and development remain in your hands. Algorithms can decide who sees an ad, but you can decide whether it’s worth watching. You have the ability to create compelling creative and develop novel ideas that tap into people’s natural curiosity.

That doesn’t mean you should stop trying to earn attention. You also need to sustain it because curiosity is what keeps the conversation going.

Dig deeper: How to structure paid social creative testing for better performance

Read more at Read More

The new SEO rules for bloggers in 2026: Why clarity matters in AI search

The new SEO rules for bloggers in 2026: Why clarity matters in AI search

For years, I told bloggers through audits, podcasts, and live trainings to write their content for “toddlers and drunk adults.”

That was simple advice, and it worked. If a 5-year-old and someone half-paying attention could follow the page content and still find what they needed, the content was probably clear enough for users and Google alike.

But in 2026, that rule needs an update.

Bloggers are no longer writing just for readers skimming on phones, distracted parents trying to make dinner, or travelers planning from an airport terminal.

They’re also writing for large language models, AI Overviews, AI Mode, and search systems that are scanning, summarizing, and deciding whether their content is clear enough to retrieve, cite, or ignore.

That’s the real shift. And as such, my advice to bloggers has evolved to “write for toddlers, drunk adults, and LLMs.”

Google is no longer just asking, “What is the best page to rank for this query?”

Increasingly, it’s asking, “What answer can we construct, what sources support that answer, and what might the user want to do next?”

That changes the job for bloggers.

The old SEO playbook was built around keywords, titles, H2s, and rankings. Those things still matter, but they’re no longer enough by themselves.

The new SEO playbook is built around clarity.

  • Can Google understand who you are?
  • Can users understand what your site is about?
  • Can AI systems understand how your articles connect?
  • Can your content answer the question quickly without making readers dig through clutter, filler, popups, and five nearly identical photos of the same finished dish?

For food, lifestyle, and travel bloggers, this isn’t a small adjustment. It’s a business survival issue.

The bloggers who continue to win won’t be the ones chasing every new acronym, every plugin score, or every AI optimization gimmick.

They’ll be the ones building clearer sites, stronger brands, better internal connections, and more useful content experiences for real people.

In 2026, clarity isn’t just good UX. It’s your SEO strategy.

Be the brand AI recommends.

See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.

See your AI visibility

Why search changed (and why it matters)

For years, most bloggers understood SEO through a pretty simple model.

A user searched for something. Google returned a list of blue links. The blogger’s job was to create the best page for that query and earn the click.

That model still exists. But it’s no longer the only model that matters.

Traditional search mostly asked: “What are the best documents to rank for this query?”

AI search increasingly asks: “What answer can we construct, what sources support that answer, what else might the user need, and what can we help them do next?”

That’s a very different environment.

In classic search, a food blogger trying to rank for “easy chicken enchiladas” mostly had to think about the visible query, the competing pages, the title, the headings, the recipe card, the photos, the intro, the internal links, and the overall usefulness of the page.

Again, those things still matter. But AI search adds another layer.

Google’s AI experiences can use what it calls query fan-out, where a single user query may trigger multiple related searches across subtopics and data sources to build a more complete response.

A search for “easy chicken enchiladas” may also involve questions around cook time, substitutions, whether to use corn or flour tortillas, how to avoid soggy enchiladas, what sauce works best, what to serve with them, how to make them ahead, and whether they freeze well.

In other words, Google may not just be matching one page to one keyword. It may be trying to understand an entire task. That’s the shift bloggers need to internalize.

The old SEO model looked something like this: Keyword → title → H2s → rank.

The modern SEO model looks more like this: Topic clarity → answer clarity → entity clarity → internal links → technical crawlability → structured data accuracy → firsthand experience → brand trust → direct audience.

That’s not as catchy. But it’s a lot closer to how search works now.

For bloggers, this means Google needs to understand more than the fact that you published a recipe, tutorial, itinerary, or guide. It needs to understand:

  • Who you are and what your site is about.
  • What topics, cuisines, destinations, or lifestyle categories you specialize in.
  • Whether your content is based on real experience.
  • How your articles connect to one another.
  • Whether the page is easy to crawl, read, summarize, and trust.
  • Why your version deserves to be surfaced instead of the dozens, hundreds, or thousands of similar options already online.

This is where many bloggers get uncomfortable. For a long time, blogging SEO rewarded a certain kind of mechanical execution.

  • Find a keyword.
  • Write a long post.
  • Put the keyword in the right places.
  • Add the recipe card.
  • Insert the photos.
  • Answer a few FAQs.
  • Hit publish.

That approach was never as foolproof as people wanted it to be, but it worked often enough that bloggers kept doing it.

AI search makes that weaker because keywords alone are a poor substitute for clarity.

A food blogger who publishes another generic chocolate chip cookie recipe with the same tips, photos, substitutions, and “why you’ll love this” section as everyone else is making Google work too hard.

The same goes for a travel article without firsthand details, original recommendations, strong structure, or clear audience focus, or broad lifestyle advice that could have been produced by any website, freelancer, or AI tool.

That kind of content was already vulnerable. Now it actively works against you for algorithmic consideration.

Google’s own guidance around AI search continues to point bloggers back to the fundamentals: helpful content, crawlability, internal links, page experience, visible textual content, high-quality images, and structured data that matches what users can actually see on the page.

There’s no secret AI SEO template, magic schema type, or plugin setting that suddenly makes unclear content useful.

AI search doesn’t remove the need for good SEO. It raises the cost of bad SEO.

If your content is confusing, your site structure is messy, your internal links are generic, or your experience and author signals are weak, you’re giving Google fewer reasons to understand and trust your content.

And if the only reason a page exists is because a keyword tool said it had search volume, that’s not a strategy. That’s content roulette.

For food, lifestyle, and travel bloggers, the opportunity in 2026 is to become easier to understand – for users, Google, and AI systems.

In an AI-shaped search environment, the winning content isn’t necessarily the longest, loudest, or most aggressively optimized. It’s the content that makes the answer obvious, the expertise clear, and the next step easy.

That’s the foundation for the new SEO rules that actually move the needle.

Dig deeper: AI isn’t the enemy: How bloggers can thrive in a generative search world

The new SEO rules that actually move the needle

When search changes quickly, bad advice gets louder.

Every week, bloggers are being told they need a new AI plugin, a new content template, a new schema trick, a new prompt strategy, a new content score, or a new way to “optimize for AI.” Most of it is noise.

Just last week, a blogger told me, “I was told not to add notes to a recipe card because it’s duplicate content.” And I literally fell out of my chair laughing.

First, that’s not how duplicate content works. Second, “notes” are incredibly important in recipe cards because users often print the card and never return to the full article.

Removing “notes” removes a valuable user-first feature that ensures your users can actually make your recipe perfectly the first time by implementing your “best” advice.

Instead of giving nonsense like this validation, bloggers should focus on making real progress by embracing the boring, obvious, high-impact things better than everyone else.

That starts with three priorities.

1. Improve high-value existing content

Most bloggers don’t need to publish more random content. They need to improve the content that already has signals.

This is one of the biggest differences between mature SEO strategy and “I saw someone say this in a Facebook group” SEO strategy. A mature strategy starts with data.

Go into Google Search Console and identify articles with traffic history, strong impressions, revenue potential, seasonal upside, rankings sitting just outside meaningful visibility, clear signs of decay, or content that hasn’t been meaningfully improved in 12 to 24 months.

That’s where you start.

Not with a random list of articles you personally feel like updating. Not with every recipe that has an old date. Not with whatever your SEO tool says is missing 14 semantically related phrases.

Use Search Console. Look for opportunity. Then make the page better.

Tip: Sign up for an account with SEO Gets. It has a fantastic Content Decay Reporting option, which makes finding and updating your weakest content within GSC a snap.

A real update should improve the experience for the user. It should make the recipe, guide, tutorial, or article clearer, more accurate, more complete, or more useful.

For a food blogger, that may mean adding better testing notes, clearer process shots, more useful substitution guidance, storage information, make-ahead tips, or troubleshooting details.

For a travel blogger, it may mean updating prices, hours, logistics, neighborhood recommendations, transportation details, seasonal considerations, or firsthand observations.

For a lifestyle blogger, it may mean tightening the structure, removing generic filler, adding examples, clarifying steps, or improving the connection between related articles.

What it shouldn’t mean is changing the date and pretending the content is fresh.

A good content update answers one question: “Is this page now more useful than it was before?”

If the answer is no, you didn’t improve the content. You just touched it.

2. Build direct audience channels

Google traffic is rented land, and bloggers need to start treating it that way.

That doesn’t mean Google traffic is bad. For many publishers, organic search is still the largest traffic source, the biggest revenue driver, and the reason their businesses exist in the first place.

But you don’t own that traffic. A ranking can disappear. A featured snippet can vanish. An AI Overview can satisfy more of the query directly in the search results. A SERP layout can shift. A core update can rewrite the competitive landscape overnight.

None of that means bloggers should abandon SEO, but it does mean they need to stop pretending Google is a stable business model by itself. It’s not.

The better approach is to use Google as one discovery channel among many, not the entire foundation of the business.

Search should introduce new people to your brand, but once those people arrive, the goal should be to give them a reason to come back without needing Google as the middleman every time.

That means bloggers need to get more serious about email, branded search, repeat visitors, Pinterest, YouTube, short-form video, Facebook groups, communities, direct partnerships, products, courses, cookbooks, memberships, or whatever audience channel actually makes sense for their niche and bandwidth.

Not every blogger needs to do all of those things. In fact, most shouldn’t.

But every blogger should be asking the same basic question: “If Google sent me fewer clicks tomorrow, how would my audience still find me?”

That question is uncomfortable, but it’s also necessary.

An email list isn’t perfect. Deliverability is annoying. Open rates fluctuate. People unsubscribe. Platforms change. But an email subscriber is still closer to an owned audience than a ranking you rent from Google.

The best question for bloggers in 2026 isn’t just, “How do I get more Google traffic?”

It’s, “How do I make sure the people who find me once can find me again?”

That mindset changes how you write, how you structure your site, how you promote content, and how you think about brand.

A blogger with a recognizable point of view, repeat readers, and a direct audience is in a much stronger position than a blogger whose entire business depends on Google continuing to send traffic at the same rate it did three years ago.

SEO still matters. But it should feed the business. It shouldn’t be the entire business.

3. Make content easier for humans and machines to understand

This is where clarity becomes the strategy.

Google’s guidance around AI search continues to emphasize many of the same fundamentals it has emphasized for years: helpful content, crawlability, internal links, page experience, visible text, images, video, and structured data that matches what users can actually see on the page.

That should tell bloggers something important. There’s no separate magic version of SEO for AI search.

There’s no special “AI search” schema that suddenly makes a mediocre article useful. There’s no plugin setting that turns unclear content into trusted content. There’s just less tolerance for bloated, confusing, poorly structured pages.

For bloggers, making content easier to understand usually means improving the basics that should have mattered all along.

A recipe article should tell readers what they’re making, why the recipe works, what ingredients matter, what substitutions are realistic, what can go wrong, how to store it, and why they should trust your version.

A travel article should make it clear who the itinerary is for, when the advice applies, what has been personally tested, what to book in advance, and what mistakes readers can avoid.

A lifestyle article should be organized enough that a reader can quickly understand the problem, the recommendation, and the next step.

This isn’t about dumbing down content. It’s about removing friction.

A tired parent trying to make dinner doesn’t want to scroll through clutter to find the answer. A traveler standing in an airport doesn’t want to dig through vague storytelling to find the train information. A reader looking for a home project doesn’t want 1,500 words of generic filler before the actual instructions.

And an AI system trying to understand your content doesn’t benefit from bloated intros, weak headings, generic anchor text, mismatched schema, or buried answers either.

This is why I keep telling bloggers to write for toddlers, drunk adults, and LLMs.

The toddler needs the page to be simple enough to follow. The drunk adult needs it to be forgiving enough to skim. The LLM needs structure, context, entities, and relationships that are clear enough to interpret accurately.

If your content works for all three, you’re probably moving in the right direction.

That means using clear top-of-article summaries when they help, headings that guide the reader instead of stuffing keywords, internal links with descriptive anchor text, schema that accurately reflects visible content, and FAQs that answer real questions.

Add to that stronger about me and author signals, original images or process shots where they add value, and a page experience that doesn’t actively fight the reader. None of that is flashy. But it’s effective.

The bloggers who win in 2026 won’t be the ones who publish the most content, install the most AI tools, or chase the newest acronym. They’ll be the ones who make their content easier to understand, easier to trust, and easier to remember.

That’s what moves the needle now.

Dig deeper: Blogging, AI, and the SEO road ahead: Why clarity now decides who survives

Get the newsletter search marketers rely on.


The outdated advice bloggers should ignore

Part of succeeding in 2026 is also knowing what to stop doing.

That may be harder than it sounds because a lot of outdated SEO advice did work at some point. Or at least, it appeared to work well enough that bloggers kept repeating it.

The problem is that many of those tactics were built for an older version of search. They were built for a world where ranking often felt like a checklist: find the keyword, put it in the title, repeat it in the headings, write more words than the competitor, add a few FAQs, hit publish, and hope Google approved.

That kind of thinking isn’t completely useless, but it’s painfully incomplete. And in an AI-shaped search environment, incomplete strategies become expensive.

The first thing bloggers need to stop doing is chasing word count.

A 3,000-word recipe article isn’t automatically better than a 900-word recipe article. A 5,000-word travel guide isn’t automatically more helpful than a tighter, better-organized 2,000-word guide. Length isn’t the goal. Usefulness is the goal.

Sometimes a topic needs depth. Sometimes a reader needs context, troubleshooting, photos, variations, and a complete explanation. But sometimes the reader just needs to know whether they can use Greek yogurt instead of sour cream or how long to bake chicken thighs at 400 degrees.

Adding more words doesn’t make a page more helpful if those words don’t solve the user’s problem.

The same goes for stuffing awkward keyword variations into H2s. Bloggers don’t need headings like “Best Easy Moist Homemade Banana Bread Recipe With Ripe Bananas” because a tool suggested combining every phrase into one linguistic crime scene.

Headings should guide the reader. They should make the page easier to scan, easier to understand, and easier to navigate. If a heading sounds like it was written for a spreadsheet instead of a human being, it probably needs to be rewritten.

Bloggers also need to be more skeptical of publishing generic content just because a keyword tool says there is search volume. This is especially important in crowded niches like recipes, travel, and lifestyle, where the web already has more than enough generic versions of almost everything.

Another “chocolate chip cookie recipe” isn’t a strategy unless there is a reason your version deserves to exist.

Another “best things to do in Paris” article isn’t a strategy unless it reflects real experience, clear audience fit, original insight, or a better way of helping the reader make decisions.

Another “how to organize your pantry” article isn’t a strategy if it says the same thing every other article says with different stock photos.

Search volume tells you there is demand. It doesn’t tell you that you deserve to rank. That part is still on you.

Food bloggers especially need to reconsider the way they structure recipe articles visually. Somewhere along the way, many articles turned into an obstacle course of nearly identical finished-dish photos before the reader gets to the actual recipe.

One hero shot is great. A useful process shot is great. A labeled ingredients photo, awesome. Photos that show texture, doneness, technique, or troubleshooting can all absolutely improve the content.

But five beauty shots of the same casserole from slightly different angles before the recipe card aren’t helping a tired person make dinner.

This doesn’t mean bloggers should strip personality, visuals, or storytelling from their sites. On the contrary, it means every element on the page should earn its place.

A photo should help, inspire, or clarify. A paragraph should answer a question, build trust, or move the reader forward. A section should exist because it improves the experience, not because everyone else’s template includes it.

Bloggers also need to stop hiding the answer.

This is one of the simplest and most common problems I see in daily audits. The user arrives with a specific need, and the page makes them work too hard to satisfy that need.

The answer is buried under a long intro, vague personal story, oversized ad units, popups, duplicate photos, or a “jump to recipe” button doing the emotional labor for the entire page.

If the page is about whether you can freeze banana pudding, say so clearly.

If the page is about the best time to visit Sedona, answer it clearly.

If the page is about how to keep meatballs from falling apart, don’t make the reader scroll through 11 paragraphs before you explain the binder.

This isn’t just an SEO issue. It’s a respect-the-reader issue.

FAQs are another place where bloggers have been led astray. A helpful FAQ can be great when it answers real questions users actually have. FAQs also allow the average blogger to optimize for People Also Ask and Things to Know search accordions.

But too many FAQ sections are still being treated as a dumping ground for keyword variations, thin answers, or filler content added because a tool, course, or template said every article needs them.

Every FAQ should pass a basic test: Would a real reader ask this, and does the answer help them do something? If not, cut it.

Updating dates without meaningful updates is another tactic that needs to die.

Changing “2024” to “2026” in the title, or adjusting the publish date and calling it refreshed, doesn’t make the content better. It just makes it newer-looking.

A real update improves accuracy, usefulness, structure, media, internal links, testing notes, examples, recommendations, or user experience.

A fake update changes the timestamp and hopes nobody notices. Readers notice. Google notices.

Bloggers also need to stop writing for SEO scores instead of users.

I understand why this happens. Tools give clear feedback. Green lights feel good. Scores feel objective. It’s comforting to believe that if a tool says the page is optimized, the page must be good.

But tools aren’t users.

Tools don’t make dinner. Tools don’t plan vacations. Tools don’t troubleshoot a broken frosting, choose a hotel near a train station, or decide whether a slow cooker recipe is safe to leave on all day.

Tools can support judgment. They can’t replace it. The same goes for panic-based SEO.

Every Google announcement doesn’t require a sitewide overhaul. Every AI update doesn’t require installing a new plugin. Every ranking drop doesn’t mean your site is broken. Every Facebook thread isn’t an emergency.

One of the worst things bloggers can do right now is make extreme technical or content changes based on fear, screenshots, rumors, or one person’s traffic chart.

The answer to uncertainty isn’t panic. It’s better analysis.

It’s Search Console. It’s content audits. It’s user behavior. It’s competitive review. It’s better internal linking. It’s clearer content. It’s understanding what actually changed before deciding what to change next.

The blunt version is this: Stop producing content that looks like it was made to satisfy a plugin, a keyword tool, or an ad network.

Start producing content that makes a tired person at 5:30 p.m. say, “Thank God, this actually helps.” That’s the standard. 

Not word count. Not tool scores. Not awkward keyword repetition. Not fake freshness. Not more filler.

The standard is whether the content helps a real person solve a real problem faster, better, or with more confidence than the alternatives.

That’s the advice bloggers should carry into 2026. Ignore anything that pulls you further away from that.

Dig deeper: The implosion of the blogging-for-dollars revenue model

Why clarity beats complexity in an AI world

When search gets more complicated, the natural reaction is to make the strategy more complicated as well.

That’s how bloggers end up chasing every new newsletter strategy, every new tool, every new AI feature, every new schema recommendation, and every new “future of SEO” thread that shows up in a Facebook group.

I understand the impulse. When traffic is volatile and Google keeps changing the interface, doing something feels better than doing nothing. The problem is that a lot of the “something” being recommended right now is either unnecessary, unproven, or a distraction from the work that would actually help.

If a blogger asked me what to focus on in 2026, my answer wouldn’t be to make their site more complicated. It would be to make their site easier to understand. That’s the through-line.

AI search doesn’t reward confusion. Users don’t reward confusion. Google doesn’t need more clutter to crawl, interpret, and summarize.

What everyone needs is clarity. That starts with the page itself.

When someone lands on a recipe article, they should quickly understand what the recipe is, why it works, what makes it different, how long it takes, what ingredients matter, and whether it solves the problem they came with.

When someone lands on a travel article, they should quickly understand who the advice is for, when the information applies, whether the writer has firsthand experience, and what decisions the reader needs to make next.

When someone lands on a lifestyle article, they should quickly understand the problem, the recommendation, the steps involved, and why the advice is trustworthy.

That doesn’t mean every article has to be short. It means every article has to be organized.

There’s a big difference between depth and bloat.

  • Depth helps the reader make a better decision. Bloat makes them scroll.
  • Depth answers follow-up questions. Bloat repeats the same point three different ways.
  • Depth shows experience, testing, judgment, and specificity. Bloat fills space because a tool, template, or competitor word count made the blogger nervous.

In an AI world, that difference matters even more.

Large language models, AI Overviews, and AI-assisted search experiences are trying to understand relationships. They’re looking for entities, context, steps, attributes, comparisons, caveats, and source quality signals.

If your content is buried under vague headings, generic filler, weak internal links, and unclear structure, you’re making that job harder.

And making Google work harder is rarely a winning SEO strategy.

This is why “clarity beats complexity” isn’t just a nice phrase. It’s a practical publishing standard.

  • Before adding another section, ask whether it makes the page easier to use.
  • Before adding another FAQ, ask whether a real reader would ask that question.
  • Before adding another image, ask whether it clarifies a step, shows texture, demonstrates technique, or builds confidence.
  • Before adding another internal link, ask whether it helps the reader move naturally to a related answer.
  • Before adding another AI tool, ask whether it improves the user experience or just makes the blogger feel like they’re keeping up.

That last one is important. AI tools aren’t automatically bad. AI summaries, AI buttons, chat features, personalization tools, and recipe modification features can all be worth testing in the right context.

But they should be treated like UX experiments, not magic ranking levers.

If an AI button helps a reader scale a recipe, find a substitution, adjust serving sizes, or understand a process faster, great. Test it. Measure it. See how users respond.

But don’t install something because everyone else is panicking. Don’t assume “AI” in the feature name means it improves SEO. And definitely don’t let the tool become more important than the content.

For most bloggers, the highest-impact AI strategy is still not very sexy: make the page easier to read, easier to crawl, easier to summarize, and easier to trust.

That means clear introductions, useful summaries, descriptive headings, original experience, accurate schema, strong internal links, and visible text that actually says something.

It also means removing the things that get in the way.

  • Over-aggressive ads get in the way.
  • Intrusive popups get in the way.
  • Repetitive photos get in the way.
  • Generic “why you’ll love this” sections get in the way when they’re not personalized to the article.
  • Filler FAQs get in the way.
  • Weak author signals get in the way.
  • Internal links with anchor text like “click here” get in the way.
  • Redundant equipment lists added purely for affiliate reasons get in the way.

A blogger doesn’t need a 47-point AI optimization checklist to understand this.

They need to look at the page and ask a simpler question: “Is this easier to understand than what already ranks?”

Does your page help the reader faster, better, and with more confidence than the alternatives?

That’s what clarity does. It removes friction, makes expertise easier to see, helps Google and AI systems understand your content, builds trust, and makes your brand more memorable.

For food, lifestyle, and travel bloggers, clarity works across traditional organic search, AI Overviews, AI Mode, direct traffic, email, social, and conversion.

That’s why I keep coming back to the same point: clarity isn’t dumbing down your content. It’s respecting the reader’s time.

The bloggers who win in 2026 won’t be the ones who make their content more complicated to look sophisticated. They’ll be the ones who make complicated decisions feel easier for the reader.

  • What should I cook tonight?
  • Can I make this ahead?
  • Is this destination worth three days?
  • Will this work for my family?
  • Can I trust this recommendation?
  • What do I do next?

Answer those questions clearly, and you’re already doing more than most of the web.

That’s why clarity beats complexity. And that’s why clarity is no longer just a writing principle. It’s an SEO strategy.

Your 2026 blogging SEO action plan

The good news is that bloggers don’t need to rebuild their entire sites overnight.

They don’t need to chase every AI feature, rewrite every article, install every new tool, or turn their content strategy into a science experiment.

In most cases, the best next step is much simpler: 

  • Identify the pages that matter most.
  • Make them clearer.
  • Strengthen the signals around them.
  • Stop doing the things that create friction for users.

Start with your highest-value content

These are the articles that already have traffic history, strong impressions, seasonal value, affiliate or ad revenue potential, or rankings that are close enough to improve.

Don’t begin with the random article that annoys you personally or the recipe you happen to have new photos for. Start where the opportunity is visible.

Use Search Console. Look at what Google is already showing. Find the articles where better clarity, stronger internal links, improved structure, fresher details, or more useful supporting content could reasonably move the needle.

Then work through the page like a real user.

Add a clear top-of-article summary

For most bloggers, the first improvement I’d recommend is a clear top-of-article summary. This doesn’t need to be long, and it definitely shouldn’t be a generic AI-written blob pasted across the site. It should be specific to the article and useful for the reader.

  • For a recipe, that summary might explain what the dish is, why it works, how long it takes, what makes it different, which ingredients matter most, and any major make-ahead, storage, or substitution notes.
  • For a travel article, it might clarify who the guide is for, when the information applies, how much time the reader needs, and what decisions they should make first.
  • For a lifestyle article, it might define the problem, preview the solution, and make the next step obvious.

The goal isn’t to repeat the whole article at the top. It’s to reassure the reader that they’re in the right place while giving Google and AI systems a clearer understanding of the page.

Again, this isn’t magic. It’s just good communication.

Strengthen internal links

Improve internal linking with descriptive anchor text.

This is still one of the most underused SEO opportunities in blogging, especially on older sites with hundreds or thousands of articles. Too many bloggers rely on related article widgets, sidebar links, or generic anchor text and assume that’s enough.

In-content links are stronger when they’re natural, useful, and descriptive.

A reader making a slow cooker chicken dinner may also need a related side dish, a sauce, a storage guide, or another easy weeknight meal. A reader planning a weekend in Charleston may need a hotel guide, a restaurant list, a packing guide, or a nearby day trip. A reader working through a furniture troubleshooting article may need links to paint, sanding, and woodworking guides and tutorials.

The anchor text should tell the reader where they’re going.

“Click here” doesn’t help much. “Easy slow cooker chicken dinners for busy weeknights” does.

“Read more” doesn’t help much. “Beginner furniture restoration guide” does.

Clear internal links make topical relationships easier for users, Google, and AI systems to understand. They also keep readers moving through the site instead of sending them back to Google.

Make headings useful

Headings shouldn’t be keyword dumpsters. They should be signposts that help readers and search systems understand the structure of the page.

For recipes, that usually means headings like “Why this recipe works,” “Key ingredients,” “How to make X recipe,” “Recipe testing notes,” “Substitutions and variations,” “Storage and reheating,” “What to serve with X recipe,” and “X recipe FAQs.”

For travel content, it might mean “Best time to visit,” “How many days you need,” “Where to stay in X,” “How to get around X,” “What to book in advance,” “Mistakes to avoid,” and “Sample itinerary in X.”

For lifestyle content, it might mean “What you’ll need,” “How to get started with X,” “Common mistakes with X,” “What worked for me,” “When to avoid this,” and “Next steps.”

The clearer the structure, the easier the page is to use and interpret.

Review your schema

Bloggers should also review schema, but they should do so with restraint.

Structured data isn’t a place to invent things, exaggerate claims, or mark up content that users can’t see. It should accurately reflect what’s visible on the page.

If the recipe says the total time is 45 minutes, the schema shouldn’t say 30 minutes because that looks better in search results. If the author, date, images, ratings, or recipe details are inaccurate, fix them.

Schema should clarify. It shouldn’t cosplay as content quality.

Reduce content friction

Focus on reducing content friction. Ads, popups, sticky units, email captures, video players, and interstitials may all have business value, but they also create user experience costs.

If a user can’t read the recipe because the screen is covered in interruptions, or a travel guide is buried under popups and sticky elements before the reader can find the itinerary, that’s a problem.

This doesn’t mean bloggers should remove every ad or stop monetizing their work. But the page still has to be usable.

If monetization makes the content harder to trust, read, or act on, the short-term revenue win may be creating a long-term visibility problem.

Test AI features carefully

AI features, such as AI buttons, can be useful in the right context. A button that helps users scale a recipe, find substitutions, modify ingredients, summarize key steps, or troubleshoot a process may improve the user experience.

But bloggers should measure before assuming.

  • Do users click the buttons?
  • Do they stay longer?
  • Do they convert better?
  • Do they subscribe?
  • Do they view more pages?
  • Do they actually use the feature, or does it just make the site look more “AI-ready”?

AI tools should support the reader. They shouldn’t distract from the content, replace firsthand experience, or become the strategy by themselves.

Dig deeper: How to design content that AI systems prefer and promote

If AI can’t find you, customers won’t either.

Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.

See your AI visibility

Clarity is what wins in 2026

Improve the articles that already have opportunity, add clearer summaries, strengthen internal links, clean up headings, verify schema, reduce friction, build direct audience channels, and test AI features only when they clearly help users.

That’s not flashy. But most durable SEO isn’t flashy. It’s a consistent, thoughtful improvement applied to the pages that matter most.

For the average blogger, that’s the plan I’d recommend. Do fewer random things. Do more useful things. Make the best content easier to understand, easier to trust, and easier to revisit.

That’s how you build a site that can compete in traditional search, AI search, and whatever version of AI Mode Google throws at us next.

Read more at Read More

Schema for AI search: How to identify and prioritize entity gaps

After presenting a custom schema my team built to evaluate knowledge graphs for university programs, I saw firsthand how differently SEOs view the role of schema markup. Our approach used 23 existing Schema.org entities and more than 60 additional entities to assess gaps in entity coverage.

Schema connects entities (objects, people, concepts, and ideas) to build knowledge graphs that provide deeper semantic understanding. It can also serve as a framework for evaluating a site’s vector embeddings and identifying gaps in entity coverage.

Let’s look at how you can use schema and knowledge graphs to identify and prioritize those gaps.

How knowledge graphs turn entities into context

A knowledge graph stores entities as nodes and relationships as edges, so machines can understand context and meaning rather than matching keywords. It’s how a system knows that the Tulane Freeman School of Business is an organization that offers business courses taught by a person rather than a string of text.

Enterprise-level businesses use knowledge graphs to remove data silos. In the book “The Knowledge Graph Cookbook: Recipes That Work” by Andreas Blumauer and Helmut Nagy, knowledge graphs are used as the “ultimate linking engine” to create a semantic data fabric for business intelligence.

In this context, your website uses the same recipe that enterprise-level businesses use for internal data. Your website is essentially functioning as a public API through which your brand’s entities, such as organization, location, products, positioning, values, features, and key benefits, are connected to search engines and LLMs.

Dig deeper: When and how to use knowledge graphs and entities for SEO

Be the brand AI recommends.

See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.

See your AI visibility

Treat schema as the on-ramp to the graph

Schema markup declares entities in a language that search engines and LLMs already speak. JSON-LD explicitly asserts both entities and the relationships between them rather than relying on bots to infer them from website copy.

For example, an executive Master of Business Administration (MBA) degree from New York University should present a coherent body of information rather than a loose collection of unrelated facts.

The organization is NYU, which offers a Brand and Digital Strategy course taught by professor and author Scott Galloway, a faculty member in the MBA program at the NYU Stern School of Business.

Stop padding FAQ schema and optimizing for rich results. Website pages are vehicles for nodes and edges that feed the graph. This approach provides critical context about your business, helping search engines and LLMs better connect your products and services with relevant audiences.

Build the graph with markup, vectors, and agents

Our custom schema is a framework for explicitly asserting all the entities and relationships a prospective student engages with on their enrollment journey. The schema prioritizes these entities by importance. 

That declared truth is what we use to analyze the vectorized version of a partner’s website. Vector embeddings of .edu website content are measured against the schema to capture semantic proximity and context that the markup doesn’t state outright.

We use agents to compare the static custom schema markup, semantic proximity, and vector embeddings to ensure comprehensive coverage of all the entities surrounding a university program. 

The results surface entity gaps, which can inform content strategy across websites, social media properties, and earned media. The final output is a knowledge graph that highlights both covered and missing entities related to a university program.

We aim to provide robust context that helps shape and frame the value of our programs through a search or conversation. This approach is directly aligned with Google’s patents on understanding entities. 

Google’s need to understand context is driven by a desire to deliver a more relevant answer to its end user. Building your knowledge graph adds context that can help search engines and AI systems better understand your business.

Dig deeper: Google’s LLM patent suggests a new goal for SEO: Teaching AI who you are

Get the newsletter search marketers rely on.


Get honest about schema and AI visibility

In March 2025 at SMX Munich, Fabrice Canel, principal product manager at Microsoft Bing, confirmed that Copilot uses schema markup to understand content.

When we consider how other LLMs use search indexes for grounding, schema markup becomes an important factor in connecting entities back to your site. Semrush has also published guides showing correlations among cited pages.

Other research points in a different direction. A 2025 study by Search Atlas found that schema doesn’t affect LLM visibility, while a study by Mark William-Cook showed that LLMs don’t read on-page schema. However, those studies look at performance and citations rather than entity connections.

These findings may seem conflicting, but they point to a distinction between using JSON-LD as infrastructure for machine understanding and treating it as a GEO hack for increasing citations. JSON-LD provides explicit information about entities and relationships that can help machines understand your site. Visibility is a byproduct of being understood.

Use schema and vector embeddings to find entity gaps

The custom schema I discussed above is specifically designed for higher education, but you can build your own. Use Schema.org as your library to build a robust collection of entities that matter to your product, service, or business. 

Does Schema.org map to all your important entities? In our case, Schema.org covered only 23 higher education-specific entities, leaving significant gaps. We filled the gaps by creating our own entities.

Start with the ideal entity model. In a perfect world, what does that complete set of information look like? This exercise will help you determine what can be built using custom JSON-LD in addition to established entities. 

With that custom schema, you can compare it against vector embeddings to understand entity coverage. Prioritize your entity gaps by focusing on what drives the most value.

Dig deeper: How schema markup fits into AI search — without the hype

Measure entity visibility, not just rich results

As you develop a content strategy for entity coverage, it’s important to use prompt-tracking and brand sentiment tools to monitor AI visibility. Monitor your priority entities to understand how you appear across different models.

Brand sentiment tracking is especially important because it allows you to see shifts among your covered entities. Is your content matching how people are actually talking about the entity?

AI visibility and brand perception aren’t complete metrics on their own. As you see shifts in visibility and perception, compare them with conversions and performance. As citations grow, does the volume or quality of leads rise?

If AI can’t find you, customers won’t either.

Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.

See your AI visibility

Schema as infrastructure for AI understanding

Schema markup isn’t an overnight solution for driving AI citations. It’s infrastructure for helping search engines and LLMs understand your site by declaring entities and relationships, building your knowledge graph, and uncovering gaps in entity coverage.

Read more at Read More

How category framing changes which brands AI recommends

Your customers' query wording determines your AI visibility, not your brand strength

Most brands approaching AI visibility ask the wrong question: How do we get stronger as an entity so that LLMs recommend us more?

In entity SEO, we tend to say, “Build the Knowledge Graph, add schema, and get more press.” But that logic assumes the LLM is evaluating the brand and deciding whether it’s good enough to recommend for any query related to what the brand sells. The LLM evaluates the query and matches it against whatever category associations it has built for the brand from third-party content.

The difference matters enormously in practice.

As we’ve seen in multiple scenarios, recognition isn’t the same as recommendation. So being a recognized brand isn’t synonymous with being a strong brand.

What matters is whether the category your customers are using to search for you matches the category the LLM has coded you into.

What the data showed

João da Silva and I conducted a study of 12 athletic apparel brands in the U.K. over seven days, with 14,140 API runs across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews. We tested the same brands using two different category framings: athleisure and athletic footwear.

After looking at the results derived from co-mentions and putting numbers on the impact of framing on category recognition for LLMs, we took the test one step further and changed the category register in the prompt.

The results were symmetric to a degree that rules out noise:

Brand Knowledge Graph (KG) score Athleisure rate Footwear rate Δ Verdict
New Balance 64,235 1% 90% +89 Jumped (footwear-coded)
Nike 25,996 77% 90% +13 Small shift (footwear-coded with strong athleisure co-mentions)
Alo Yoga 3,062 63% 0% -63 Dropped (athleisure-coded)
lululemon 810 90% 0% -90 Dropped (athleisure-coded)
Sweaty Betty 751 9% 0% -9 Stable
Reebok 665 1% 20% +19 Small shift
Outdoor Voices 455 26% 0% -26 Small shift
Rhone Apparel 400 5% 0% -5 Stable
Varley 381 6% 0% -6 Stable
TALA 356 5% 0% -5 Stable
Gymshark 277 37% 0% -37 Dropped (athleisure-coded)
LNDR 2 0% 0% 0 Stable

Notes:

  • New Balance goes from 1% to 90%.
  • lululemon goes from 90% to 0%.

The variation is approximately 0.9 points in both directions simultaneously.

We’re not talking about correlation here, but a controlled observation: what happens when we change only one variable — the category word in the prompt.

Be the brand AI recommends.

See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.

See your AI visibility

Why this happens: Category coding

Nike, New Balance, and Reebok share the exact same Google Knowledge Graph (KG) description: “Footwear company,” so all three are recognized perfectly by every LLM we tested. From an entity standpoint (recognition), they start from an identical position. However, their behavior under different category framings isn’t identical at all.

The reason is what the paper formalizes as category coding: the combination of the KG description field and the third-party content corpus that has accumulated around a brand in a given category.

The KG description anchors a brand to a category in the model’s representation (impacts recognition).

The third-party corpus — articles, reviews, editorial comparisons, and roundups — fills in the detail of what that category association actually looks like (impacts recommendation).

Looking at the example from New Balance

New Balance’s KG description says “Footwear company,” and the third-party corpus that has accumulated around it corroborates the category by focusing on topics related to running shoes, performance footwear, and athletic training. 

When a user asks about athleisure brands, the model doesn’t find New Balance in that corpus because there is no third-party association. But it does find lululemon, Alo Yoga, and Gymshark: all brands whose corpus is built from fashion publications, lifestyle editorial, and activewear roundups.

When we changed the query to athletic footwear, the retrieval flipped: New Balance is suddenly in the right corpus, and lululemon is not.

The model itself can’t and isn’t making a judgment about brand quality or belonging. What an LLM does is pattern-match a query category against a content category. If those two things align, the brand surfaces. If they don’t, it doesn’t, regardless of how established the brand is.

So, can you just recode your KG description? 

Some brands reading this will consider the obvious shortcut: Change the KG description. If “Footwear company” is anchoring you to the wrong category, recode it to “Apparel company,” and the problem is solved.

However, the KG description is only half of what determines category coding. The other half is the third-party content corpus that has accumulated around your brand, and that doesn’t change because you updated a field in the Knowledge Graph. If your entire external content history is performance footwear, running, and athletic training, changing the description gives the model a new anchor with nothing attached to it. The corpus still says what it always said.

The corrective lever is third-party content investment in the specific category framing your customers are using: in the publications the model retrieves from, alongside the brands that already define that space. The KG description can support that work once the corpus exists.

Get the newsletter search marketers rely on.


What this means for your GEO strategy

The standard GEO advice is to strengthen your entity: a consistent name, clean schema, a strong About page, and more press coverage. That advice is correct for getting recognized and even recommended within the brand’s coded category, but it isn’t sufficient for getting recommended in adjacent category queries.

What determines recommendation in adjacent categories is whether the third-party content corpus around your brand matches the category framing your customers are actually using.

The questions worth asking about any brand are:

  • Are we visible in AI?
  • What category has the LLM coded us into?
  • Is that the category our customers are querying?

If a brand is strong in one category, but its customers are increasingly using adjacent category language to search (for example, athleisure instead of sportswear, or performance wellness instead of fitness), and the brand’s third-party corpus hasn’t kept pace with that language shift, the brand will be invisible in exactly the queries customers are using.

Nike is the study’s clearest positive case, surfacing in both athleisure (77%) and athletic footwear (90%) queries, despite being KG-coded as a footwear brand.

The reason is that Nike has accumulated enough athleisure-coded third-party content, including editorial coverage in fashion publications, inclusion in activewear roundups, and co-mentions with other athleisure brands, to register as category-eligible in both framings. It built a sub-stream in the adjacent category that New Balance didn’t.

If AI can’t find you, customers won’t either.

Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.

See your AI visibility

What is the audit question everyone should be asking?

Before investing further in entity optimization, it’s worth running a simple diagnostic: Take the five or six different ways your customers might phrase a category query for what you do, and test each one across two or three LLMs. Note which formulations surface your brand and which don’t.

For the ones that don’t, the questions to ask are:

  • Does third-party content about your brand actually use that language?
  • Are you being written about in publications that cover that category?
  • Are you appearing in editorial roundups that use that phrasing?

If the answer is no, you know where to start: getting into the external conversations that speak the language of that query.

Closing that gap means becoming a participant in the category comparison content that defines who belongs in that space.

This article is based on findings from “The recognition-recommendation gap: Empirical evidence that category coding, not knowledge-graph strength, determines brand visibility in generative AI output,” co-authored with João da Silva and published open access on Zenodo.

Read more at Read More

New in the Yoast SEO Bulk Editor: filters, focus keyphrases, and AI drafts

The Yoast SEO Bulk Editor already gave you one place to update titles and meta descriptions across your whole site. Now it does much more. This release adds focus keyphrases to the editor, new search and status filters, a dedicated view for every content type, and AI-drafted metadata with Yoast SEO Premium. In short: you can find what’s broken, fix it in bulk, and let AI handle the heavy lifting while you keep final say.

Running an online store? Jump to what this means for your shop.

A focused workspace

The Bulk Editor now shows only the fields that matter: your titles, meta descriptions, and, new in this version, your focus keyphrases. Everything is searchable, filterable, and organized per content type. As a result, posts, pages, products, and custom post types each get their own editor, which is ideal for publishers and stores alike.

Whether you manage 50 pages or 5,000, you can still move through them quickly and systematically. Just inherited a site? The Bulk Editor remains the fastest way to get it to a solid SEO baseline without starting from scratch.

New: search and status filters

You can’t fix what you can’t see all at once. The new filters get you straight to the pages that need work. For example, you can pull up every page missing a description, every draft, or every post targeting a keyphrase in seconds instead of scrolling. From there, you review, update, and move on without opening a single page.

Easily search for the page, product or post you are looking to edit

New: AI drafts it, you approve it (selected plans*)

Even when you can see what’s missing, writing hundreds of titles and descriptions by hand is a project nobody has time for. With paid plans like Yoast SEO Premium, you can now select the content that needs work and let AI draft the metadata for you. You then review each suggestion, apply what you like, and tweak or discard the rest.

Crucially, nothing saves without your permission. Per-row and batch apply or discard controls, plus unsaved-changes warnings, give you bulk editing without bulk mistakes. That means you can work through hundreds of pages with confidence, not caution.

Working with the AI powered metadata generator

*AI drafting is available with Yoast SEO Premium. WooCommerce and Yoast SEO AI+ plans extend it to products and categories.

New for stores: catalog metadata at scale

Keeping product metadata complete and unique across a large catalog is one of the most tedious jobs in e-commerce. This release helps in two ways. First, products get their own view in the Bulk Editor, so you can filter your catalog and spot the gaps fast. Second, Yoast WooCommerce SEO and Yoast SEO AI+ plans extend AI drafting to products and categories. Your store can therefore generate, review, and approve catalog metadata at scale, right in time for the next seasonal peak.

The post New in the Yoast SEO Bulk Editor: filters, focus keyphrases, and AI drafts appeared first on Yoast.

Read more at Read More

6 Digital PR Strategies to Boost AI Visibility

Digital PR is the single most important strategy to win in AI search.

A study by Muck Rack found that 84% of AI citations come from earned media, third-party sources like editorial coverage, reviews, and forums.

Where AI gets its information

In AI search, third-party validation beats self-promotion.

Publishing onsite content is still important, no doubt. But your AI visibility is mostly shaped by what happens outside your website.

The more frequently your brand name appears across the web, the more likely LLMs are to recognize, trust, and eventually recommend you.

In this guide, I’ll show you six proven digital PR strategies to earn more backlinks and brand mentions, strengthen your offsite authority, and increase your visibility in AI-generated answers.

Why Digital PR Now Shapes AI Visibility

AI systems go well beyond reading your website.

They pull information from dozens of sources across the web, then combine it into a single answer.

How AI systems discover and cite content

The brands that win in AI search are those that appear consistently across those sources.

That’s where digital PR comes in.

It’s how you get featured in the right publications that AI systems already read and reference.

The more places you show up, the more familiar and trustworthy your brand looks to those models.

Here are six digital PR strategies to start building that presence.

Strategy 1: Data-Led PR

Data-led PR revolves around publishing original research and statistical roundups, and distributing the data to relevant publications.

This is one of the most effective ways to build high-quality backlinks.

Why?

Because journalists, bloggers, and marketers are constantly looking for credible data to support their content.

How It Boosts AI Visibility

Backlinks matter a lot for SEO. And they still matter for AI visibility.

Semrush’s study of 1,000 domains found that brands with stronger backlink authority are more likely to appear in AI-generated answers.

Correlation between backlink metrics and AI visibility

Seer Interactive’s study backs this up:

The top two metrics that impact AI visibility are domain authority and high-quality backlinks (from DA 60+ sites).

Data-led content is how you earn those quality backlinks. People love citing fresh numbers.

I helped one of my clients, Resource Guru, put together the “Agency Overworking Report 2025.”

ResourceGuru – Agency overworking report

Since it was published, this report has generated 21 backlinks naturally, including coverage in Forbes.

Agency overworking report – Backlinks

It’s also consistently being cited in AI answers.

ChatGPT – Agency overworking data

How to Do It

First, publish your data-led content.

You have several options:

  • Conduct new research nobody has done before: This is an excellent choice if you’re in a growing field like GEO, where data is scarce, and any credible study will attract attention.
  • Update existing studies: Find stats that haven’t been refreshed in over two years. Outdated data is everywhere, and writers actively look for newer numbers to replace them. If you can be the source that fills that gap, you’ll earn the citation.
  • Compile existing statistics: You don’t always need to generate new data. Aggregating hard-to-find stats into one well-organized resource can be just as link-worthy.

Once your data is live, pitch it proactively.

Find articles in your niche that reference outdated statistics or cite sources that no longer exist. Reach out to the authors and offer your fresher data as a replacement.

Also, pitch your new data to existing statistical roundups in your niche. These pages are usually updated frequently, so the author is more open to new additions.

Here’s an example of a good pitch that I received for my AI SEO stat roundup:

Research outreach pitch example

Strategy 2: AI Citation Outreach

AI citation outreach is the practice of securing placements in the specific sources AI models already pull from.

In my experience, it’s the quickest way to start earning brand mentions in AI responses.

How It Boosts AI Visibility

If you can get into the content AI cites, you can influence the answers AI gives.

For example, my agency Position Digital was included in Exposure Ninja’s listicle on “The Best AI Search Optimisation Agencies in 2026.”

Exposure Ninja – Best AI search optimization agencies

And since that listicle was cited by ChatGPT, my agency ended up being recommended in one of the answers.

ChatGPT – Best AI search optimization agency

How to Do It

LLMs cite content differently. Each model has its own preferences, tendencies, and patterns.

For example, I asked AI Mode and ChatGPT to recommend the best SEO tools.

Here’s what ChatGPT said:

ChatGPT – Best SEO tools

And here’s the AI Mode answer:

Google AI Mode – Best SEO tools

Similar answers. But totally different citations.

So, you need to tailor your strategy to each AI model you want to target.

1. Research user prompts

Start by mapping out the prompts your target audience is likely to type into AI tools.

AI companies haven’t given us prompt data yet, so we’ll have to settle for educated guesses.

Here are some good places to start:

Source Tips
User questions in sales calls Note the exact language prospects use when describing their problem
Queries in Google Search Console Look for ultra-long queries in your performance report
People Also Ask in SERPs Use a free People Also Ask tool to find common questions about a topic
Keyword data in your SEO tool Filter for question keywords starting with who, what, how, why, which

Alternatively, you can use Semrush’s AI Visibility Toolkit to find out exactly what your audience is typing into LLMs.

Visibility Overview – Position Digital – Topics & Sources

2. Monitor the cited pages

Once you have your prompt list, run each prompt through each AI model and note which pages are cited.

If you don’t want to do it manually, Semrush’s AI Visibility Toolkit shows you the sources for each prompt.

You can also use ListBrew to surface “best X” listicles and comparison pages that already get cited in AI answers for your target prompts.

List Brew – AI citations

3. Find the contacts

Find the contact information for each opportunity using tools like Hunter and Apollo.

Tip: Prioritize someone who has editorial control over the content, like the author or content editor, rather than generic email addresses like hello@domain.com.


4. Send personalized pitches

Reach out to each prospect and ask to be included in the content.

Keep the pitch short and specific to their piece. To increase your chances, offer something in return, such as:

  • A free trial or demo of your product
  • A reciprocal mention in one of your own high-traffic pieces

Also, prepare a short blurb that the author can easily copy and paste into the content. It seems simple, but it makes a big difference.

Here’s an example of a pitch I sent during one of my listicle outreach campaigns:

Listicle outreach email pitch example

Strategy 3: Reactive PR

Most digital PR campaigns take weeks or even months to prepare and execute.

Reactive PR is about speed — being the first one to respond to breaking news, viral trends, and emerging topics.

How It Boosts AI Visibility

LLM training data has a cutoff date.

When someone asks about a current event, the model can’t rely on what it already knows. It has to search the web for fresh sources.

That’s your window.

When a story first breaks, there are usually only a few credible sources covering it.

If your brand is among the first to publish useful commentary, analysis, or original reporting, your content has a much higher chance of being surfaced and cited by AI systems.

Here’s a solid example:

Search Engine Journal was among the first media outlets to report on Google’s new SEO guidelines.

Search Engine Journal – Google's new guidance

The piece has since gained over 16,000 readers, 500 backlinks, and several AI citations.

ChatGPT – SEO news – Search Engine Journal citation

How to Do It

The biggest challenge with reactive PR is spotting trending stories before they become a trend.

By the time a topic is dominating headlines, you’re already too late.

Success comes from identifying emerging conversations while they’re still gaining traction.

A few ways to do that:

  • Join online communities: Many trends gain momentum on Reddit, X, and Slack before they receive mainstream coverage.
  • Follow industry leaders on LinkedIn: Experts often share breaking news, observations, and predictions here before publishing full articles or reports.
  • Follow the major players in your industry: Product launches, partnerships, acquisitions, funding rounds, and policy changes often create PR opportunities. The sooner you spot these developments, the faster you can react.
  • Track journalist requests: Platforms like Qwoted, Featured, and JournoRequests can reveal stories that journalists are actively working on before they’re published.
  • Use Google Trends and Exploding Topics: These tools help you find emerging topics that your audience is searching for right now.

LinkedIn – Kevin Indig post

A secondary layer of this process is monitoring how stories spread across publications.

Many trends don’t appear out of nowhere in major media — they first circulate through smaller blogs and niche industry sites.

Tracking this ripple effect helps you identify narratives early and position your response while the topic is still evolving.

  • Set up Google Alerts: Create alerts for important keywords, competitors, and industry topics so you get notified as soon as new stories are published.
  • Monitor industry newsletters: Curated newsletters are often one of the fastest ways to discover emerging trends and conversations.

The SEOFOMO newsletter

Once you spot an opportunity, move quickly.

Write a report or a commentary piece on your website and social media, and pitch the story to major news outlets in your field.

Strategy 4: Ego Bait

Ego bait involves creating content that features industry experts and influencers.

The goal is to stroke their ego, making them more likely to share your content, mention your brand, or link back to your site.

Expert quote example

How It Boosts AI Visibility

Featuring recognized experts can improve AI visibility in two ways.

First, it increases the credibility of your content.

An AI SEO study found that pages containing expert quotes receive 4.1 citations in ChatGPT on average, compared to 2.4 for pages without them.

And we’ve seen it firsthand.

Our article on SEO competitor analysis features insights from 20 experts, and it’s cited by both Google’s AI Overviews and AI Mode.

Google AI Mode – SEO competitor analysis checklist – Position Digital citation

Expert commentary acts as a trust signal, making the content more authoritative and cite-worthy.

Second, it creates a distribution channel.

When experts are featured in your content, many will share it with their audience, mention it on social media, link to it from their websites, or include it in newsletters.

As those mentions and backlinks accumulate, your brand becomes more visible across the web.

Stronger brand visibility = stronger LLM visibility.

How to Do It

There are three types of ego-bait content you can create:

1. Expert roundups

Use journalist outreach platforms like MentionMatch, Featured.com, and Qwoted to gather expert insights for your content.

Choose the best quotes and feature them on your blog post.

Once it’s live, tag every contributor in your LinkedIn post.

Most will reshare it, comment, or at minimum engage with it, which extends your reach well beyond your own following.

LinkedIn – Sean Begg Flint post

2. Case studies and success stories

Highlight real results from customers, partners, or collaborators.

A well-written case study flatters the subject, gives them something worth sharing, and adds a layer of credibility that generic content can’t replicate.

3. Top experts or influencers lists

Curate lists of respected people, companies, or voices in your category.

Examples include:

  • Top SEO Experts in 2026
  • Most Influential AI Search Voices
  • Leading Growth Marketers in SaaS

Being included in a curated list is often incentive enough for people to share it, especially if the selection feels credible and relevant.

This type of structured content is also highly citable by AI systems.

Google SERP – Top SEO influencers – AI Overview

Strategy 5: Thought Leadership

SEO content is built to rank.

Thought leadership content is designed to establish you as a recognizable voice — both for people and LLMs.

And when you already have an established presence, it’s much easier to market your business.

How It Boosts AI Visibility

Most content gets ignored because it fails to give people a reason to engage.

Thought-provoking content, on the other hand, is designed to challenge assumptions, introduce new perspectives, and spark discussion.

Just look at this “controversial take” from Khanh Linh Le.

This post generated an unusually high number of comments relative to its likes because it challenged conventional thinking and encouraged debate.

LinkedIn – Khanh Linh Le post

Some people agreed. Others pushed back. But everyone had something to say.

That’s the goal: to drive engagement.

Engagement drives shares. Shares drive mentions across blogs, forums, social platforms, and publications.

And when a lot of people talk about you, AI pays attention.

How to Do It

1. Build a LinkedIn presence

The first step is to start posting on LinkedIn.

Why?

Because at the time of Semrush’s study, LinkedIn was the second-most-cited domain in ChatGPT, AI Mode, and Perplexity.

Some tips:

  • Publish consistently: AI citations reward relevance and consistency more than virality. The Semrush study found that 75% of cited authors post at least one post per week.
  • Take a stance: Share opinions, challenge assumptions, and weigh in on debates happening in your industry. Posts that take a clear stance generate far more engagement than informational updates.
  • Use your personal and company pages: Some LLMs like Perplexity gravitate more towards company pages, while others like ChatGPT and AI Mode like citing personal pages.

Anatomy of LinkedIn content in AI search

2. Write guest posts for major publications

Guest blogging is a great way to distribute your ideas and share your expertise in high-authority publications.

When multiple authoritative sources cover the same topic and all point back to you as the originating voice, AI models start to recognize you as the go-to authority on that subject.

At Position Digital, I’ve developed a framework called the Guest Post (GP) Engine.

Here’s how it works:

I publish a blog post on my website.

I then publish variations of that post on other publications, ideally those with a high website authority.

As an example, I wrote about “Content Refreshes” on the Position Digital blog.

I then wrote guest posts on the same topic on Sitebulb and Surfer SEO.

All three articles are now cited by AI Overviews.

Google SERP – Content refresh guide – AI Overview

3. Appear as podcast guests

Podcasts are one of the most underutilized thought leadership channels.

A single appearance can generate:

  • Mentions on the host’s website
  • Show notes and transcripts
  • Social media clips
  • Newsletter features
  • Citations from future content creators

If you’re just getting started, don’t chase after big, established shows.

Start small. Build relationships with newer podcasts in your niche.

Paid Forward Podcast featuring Sean Begg Flint

Smaller shows are easier to get on, their hosts are often more engaged, and the content still gets indexed and cited.

As your reputation grows, the bigger opportunities follow naturally.

Strategy 6: Community Building

Community building is about establishing your brand’s presence in third-party forums, review sites, and customer rating platforms that AI models treat as trusted, independent sources.

This way, your brand is validated not just by what you say about yourself, but by what neutral platforms say about you.

Reddit – AEO – Best AI visibility tools

How It Boosts AI Visibility

AI models trust neutral, community-driven sources over brand-owned marketing.

It’s not an opinion. It’s backed by actual data.

According to another AI visibility study by Semrush, LLMs cite Reddit threads about Microsoft products much more frequently than Microsoft’s own blog.

Let that sink in.

And it’s not just forums like Reddit. Review platforms like Trustpilot and G2, where users can share authentic reviews and ratings, carry significant weight too.

Seer Interactive also did a study of 800k AI responses and found that brands with no Trustpilot profile have a median AI citation rate of just 1%.

Brands with even a minimal profile, as few as 1 to 13 reviews, jump to 53.5%.

The implication is clear:

If AI has to choose between what a brand says about itself and what hundreds of independent users say about that brand, it will often favor the latter.

How to Do It

1. Contribute insights on Q&A platforms like Reddit and Quora

These communities are scraped heavily by AI models and frequently cited in responses.

ChatGPT – Reddit – Sources

But be careful, Reddit and Quora communities have a low tolerance for overt self-promotion.

Accounts that exist purely to promote a product get flagged, downvoted, or banned.

The key is to participate as a real contributor, not as a marketer.

Be transparent about who you are and what company you represent, but focus primarily on being helpful.

A good approach is to contribute meaningfully to 3–5 relevant threads per week.

Only mention your brand when it is genuinely relevant to the question or adds value to the discussion.

You can also host an AMA (Ask Me Anything) session to share your knowledge while building brand awareness.

Reddit – Sweaty Startup – SEO AMA

2. Create optimized profiles in review sites

Depending on your niche, set up and optimize your profiles on review platforms like:

  • G2, GetApp, and Software Advice for B2B SaaS
  • Trustpilot and Clutch for agencies and service providers
  • Tripadvisor and Google Reviews for hospitality and local businesses

After that, make sure to ask your existing customers to leave positive reviews and ratings.

3. Write content on Medium

Medium is another platform that AI frequently cites, according to the Semrush study above.

Use it to republish condensed versions of your best content or share original perspectives that complement your main blog.

Publishing the same ideas on Medium that you’d put on your own site gives those ideas a better chance of being cited.

Show Up Where AI is Already Looking

The quickest path to AI citations is getting into the pages AI already trusts.

Start by using Semrush’s AI Visibility Toolkit to uncover the prompts your audience is typing into AI chatbots.

Then, review the sources each AI platform cites for those prompts and find a way to earn a placement there.

When you’re ready to go deeper into how to optimize for citations and influence AI answers, read this guide on LLM seeding.

The post 6 Digital PR Strategies to Boost AI Visibility appeared first on Backlinko.

Read more at Read More

Hosting guide: How to pick the right host for your site

Do you want to start a blog or online store? Then you’ll need a website. But did you know you’ll also need web hosting to get your website online? There are many aspects to web hosting, and you might not be sure what it entails. You have to choose a host, a hosting type, and many other things. That’s why we wrote this guide, to help you understand what it is and pick the right host for your site!

Key takeaways

  • Web hosting makes your website accessible online, requiring you to choose a host and type of hosting.
  • Decide between fully-hosted and self-hosted options based on your control and data ownership preferences.
  • When picking a host, consider server location, storage space, audience size, and additional services offered.
  • Understand the different types of hosting: shared, dedicated, and VPS hosting to find what suits your needs.
  • Specialized hosting plans exist for WordPress.org sites, which come optimized and may include support for security and updates.

What is web hosting?

Let’s start with the basics. Web hosting is a service that makes sure your website is accessible through the internet. The pages and files that your website consists of live on a server. Whenever someone visits your website, that server sends over the necessary ‘files’ to display your site to the site visitors. This server is owned by a web hosting company, and you can rent space to store your website files on a server that meets your needs. Without web hosting, your website simply can not exist.

It’s also good to know that, although it doesn’t have a direct effect, your choice of hosting impacts your SEO. Reliable hosting improves the security, speed and uptime of your website. All factors that tell search engines that your website is trustworthy and a good option to show to their users. Upgrading your hosting plan or switching to a new hosting company can therefore improve your site performance, just make sure to choose reliability and your website’s needs over low prices.

So, where do you start? By looking at web hosts or hosting companies. These companies provide servers and internet connectivity in a data center. In other words, they provide space for your website’s data, so it can exist and other people can find it. 

Fully-hosted versus self-hosted

The first choice you have to make when picking the right host for your site is whether you want to create a fully-hosted or self-hosted website. The difference is pretty self-explanatory. Fully hosted means every site on the platform is hosted on its own servers. These companies take care of all the hosting stuff for you, such as Shopify and WordPress.com

Self-hosted, on the other hand, means you’ll have to arrange hosting yourself. A good example is WordPress.org. Now, you might be thinking: why would I want to do everything myself? And while it’s true that self-hosting is extra work, it also means you own all the data yourself. Plus, you get access to the full codebase and can choose your own hosting company.

How to choose a hosting company

These days, there are thousands of companies to choose from. They all have their own benefits and hosting plans. So, how do you pick the right host for your site? What should you focus on? First, it’s important to take a couple of things into account.

Server location

As those real estate TV shows always say: location is everything. It’s not necessarily everything when it comes to hosting, but it is important. Where your host’s servers are located is an important factor in your site’s speed. Why? The smaller the distance between your audience and your server, the faster your website is. For example, if your audience is located in the US but your host is based on the other side of the world, that could slow down your site for visitors, which is something you want to avoid.

Storage space

Second, consider what kind of website you want to create. This impacts how much storage space you need to acquire. For example, a small blog needs far fewer resources than a big online shop with thousands of products. Consider how many pages your site will have and how much content it will have. If you already know you want to build a large website with many large images and videos, you’ll need to look at hosting plans with more storage space.

How big is your audience?

We know this is a hard one to answer. But you might already have an idea of the number of page views you want to reach. Take this number into account when deciding on a hosting company and plan. Because every time someone views a page on your site, it adds to the amount of bandwidth you use. In other words, all the texts and images on that page need to be downloaded onto each visitor’s computer before they can be displayed in their web browser. 

Other services offered

A final consideration is the need for other services related to your website. Many hosting companies offer additional services, such as email hosting, SSL certificates, automated website backups, and website builders. The website plan you choose should provide the services you need to get you up and running, preferably beyond basic hosting. For example, choosing a hosting plan that comes with a website builder helps you create your website right away.

screenshot of the Bluehost AI Website Builder
The Bluehost AI Website Builder

Bluehost has recently launched an AI Website Builder. You answer a few questions, and it generates your website for you. It lets you make changes through chat, and you see changes in real time, no coding or design skills needed. Just describe what you want. Hosting and a domain are included, and human support is available if needed. Ideal if you’re setting up a website for your small business and want it to go live quickly. If your website is a bit more complex, for example, when you want to create a membership portal, this might not be the best option.

Extra questions to ask

In addition to server location, storage space, and the number of visitors, there are tons of other factors that determine your ideal hosting company. That’s why we compiled a couple of additional questions you could ask yourself (and a hosting company) when picking the right host for your site. 

  • Do I need help or support setting things up? 
  • Does the company offer live chat and phone support, and do they respond quickly? 
  • What happens if my site gets hacked or goes down in the middle of the night? Will they help? 
  • What is my budget? 
  • Do I need the ability to add more websites? 
  • Does this hosting company have good reviews? 
  • And, for higher-end packages, what type of hardware, like SSD storage, CPUs, and memory chips, do I need for my site? 

Different types of hosting

Finally, let’s look at the different types of hosting. In this guide, we’ll focus on the three most common ones: shared hosting, dedicated hosting, and VPS hosting. But there are lots of other options out there.

Shared hosting

Shared hosting means you share the same server resources with multiple customers. These server resources determine how much space, bandwidth, mailboxes, etc., you have available. It also means there will be resource limits on each hosting package. Don’t worry, sharing resources doesn’t mean other customers can see or access your data. Shared hosting is probably the cheapest option, but it might not be the best choice if you expect to need a lot of server space and bandwidth.

Dedicated hosting

Dedicated hosting does exactly what it says on the tin. You’ll get a dedicated, fully allocated hardware server for your use only. This means you’re not sharing the server with other customers. Instead, you’ll have your own personal server with its own processors, hard disks, and memory. This makes dedicated hosting an interesting choice for bigger sites, such as large online shops or corporate sites.

VPS hosting

VPS hosting stands for Virtual Private Server. A VPS is a virtual machine on a physical server that is divided into different sections. You get your own space on this server, with access to a certain amount of resources, like memory and processors. Unlike with shared hosting, these resources are for your use only. Additionally, you can optimize your VPS server to your own preferences. For example, you can determine which software runs on it and adjust the settings to your liking, just like with dedicated hosting.

Hosting plans for WordPress.org sites

If you’re planning to create a WordPress.org website, then we’ve got good news for you. There are tons of hosting companies that offer hosting plans specifically for WordPress.org sites. The benefit? The plans are completely optimized for WordPress-based websites. Usually, they already have WordPress installed, so you don’t have to take care of that anymore. Sometimes they also offer security and update services. This means you don’t have to worry about your site’s security or updating WordPress and the plugins you have installed. 

To help you out, here is a list of WordPress hosting companies.

Conclusion

Now you know all the basics about hosting companies, hosting plans, and the different types of hosting. Still, it’s good to consider which factors are important to you before choosing a hosting company and plan. Because your site’s hosting can impact factors like site speed and the amount of content you can host. 

If you want more information about hosting, we recommend checking out our Technical SEO course. The course dives into considerations when choosing a host, as well as practical tips for making your site easy for search engines to find. And much more!

FAQ

After reading this post, I realized that I might need a better host. Is it possible to switch hosts?

It’s possible, but it’s going to depend on your setup and where you’re moving to. Luckily, most hosts offer great support. Typically, they’ll guide you through all the processes you need to go through. These processes are, for example, moving your files, your database, and maybe some configuration info.

When you’re choosing your host, it’s worth getting in touch with their support team to see what kinds of services they offer and how much they can help you with that.

Okay, but there are many hosting options out there. Is there a tool that lets me enter my data and help me choose?

Oh, that’s tricky! There are lots of tools on the internet to help you pick a hosting provider, but it’s a bit of a minefield. We recommend checking out yoast.com/hosting.

But really, do your own research. Get a feel for what your budget is, what kind of features you might need, and talk to the support people. Many of these companies and websites also offer live chat.

I’m just starting out with my site, and I don’t know how many visits I’ll get. I’m hoping for a lot, of course, but I might get only a few. So, when picking a host, what should I look for in terms of bandwidth?

It’s good that you’re thinking about bandwidth beforehand, but don’t let it dictate your choice. Instead, research how much different hosting companies charge for their different packages, and how the prices change as you need to scale. Maybe you can start off small and upgrade later. Again, shop around, review pricing and packages, and talk to support staff.

The post Hosting guide: How to pick the right host for your site appeared first on Yoast.

Read more at Read More

Google AI Overviews now lets you create image

Google now lets you create images directly within AI Overviews in Google Search. “To help bring those unique ideas to life, we’re bringing image generation directly into AI Overviews in Search,” Google announced.

This uses Google’s latest Nano Banana AI model within AI Overviews to create these images.

Google said, “This update transforms a simple text prompt into a high-quality, custom visual made completely from scratch, seamlessly bridging the gap between imagination and reality.”

What it looks like. Here is a video of this in action:

Availability. Google will roll out this image generation feature within AI Overviews over the coming weeks in English, for all regions that currently support image creation in AI Mode.

Google also announced a redesign for Google Image Search, on its 25th anniversary of Google Image Search.

Why we care. This may have an impact on traffic to publishers, as it will add more AI-generated content (the images) to the AI Overview, potentially discouraging clicks from Google Search. Plus, if people get the image they want in the AI Overview, it might even discourage some use of Google Image Search – maybe?

In any event, it is wild to know that Google Image Search is now 25 years old.

Read more at Read More

Visual semantics: The missing piece of topical authority

Visual semantics- The missing piece of topical authority

SEO has long focused on what a page says. Increasingly, it also needs to account for how that information is presented.

As Google gets better at understanding page layout, structure, and functionality, visual semantics is becoming an important part of how search engines interpret webpages.

What is visual semantics?

Visual semantics is a meaning model for segmenting, classifying, and understanding documents by working alongside textual semantics.

Google is changing how it interprets web documents, shifting from “web text” to “web layout” to better identify real expertise, uniqueness, and originality by giving more weight to the functional components of a webpage.

Google’s Quality Rater Guidelines cite “human effort and involvement” as one of the most important quality principles, with “design effort” identified as one aspect of that evaluation.

Webpage layout has always been an important part of SEO, dating back to Google’s Page Layout algorithms. Those early algorithms focused primarily on ad placement and simple document-ranking signals, unlike today’s more sophisticated approaches to understanding webpages.

Be the brand customers find first.

Track, grow, and measure your visibility across Google, AI search, social, local, and every channel that influences buying decisions.

Start your free trial

Why Google is paying more attention to page layout

Google has introduced newer inventions and patents that highlight the importance of understanding webpage layout. Most webpages are no longer built with only prose or simple text-over-text layouts. Instead, they contain much denser information. 

Every 10 to 20 pixels can introduce a new interaction point, engagement element, clickable module, comparison unit, or dynamic component designed to help users.

That’s why some of Google’s leading engineers, including those who have worked on Gemini and AI Mode, are also associated with newer inventions such as Structured Information Cards and layout-aware multimodal document understanding.

Below is a direct citation from Google’s work on structured information cards and layout-aware multimodal document understanding. Google often finds important information within interactive card structures rather than ordinary paragraphs. 

As a result, it needs systems that can understand how different card types are structured, including product cards, hotel cards, real estate cards, trip cards, credit card cards, and other information cards.

In other words, modern search engines must understand not only the text on a page but also the layout, hierarchy, visual relationships, annotations, and functional meaning of each structured information block.

A citation from Google’s “Layout-aware Multimodal Document Understanding” patent
A citation from Google’s “Layout-aware Multimodal Document Understanding” patent

Why layout matters for search engines

Understanding structured information cards and layout-aware document interpretation requires neural networks, and possibly a new type of LLM, that can “verbalize” web documents with annotations and high-confidence citations.

Google can’t reliably rank a flight booking website, a credit card application aggregator, or similar platforms without understanding the data embedded in these documents. 

Much of that data is presented through uniquely designed card structures, comparison modules, tables, and interactive layouts rather than plain text.

Below is an early example of document layout understanding from Microsoft called ViPS, which Google has also cited.

Later, Google patented an alternative approach based on HTML-heavy segmentation.

Both approaches are closely related and rely heavily on HTML to determine which text belongs to each section, component, entity, or visual block on a page.

With the rise of embedding-based algorithms, concepts such as “chunking” have become widely discussed in the SEO industry. 

However, many discussions about text or document chunking miss a critical point: Chunking isn’t only a linguistic process. It’s also a layout-aware and structure-aware process.

If a document isn’t visually segmented and structurally understandable to search engines, the content itself becomes harder to interpret. In that case, it doesn’t matter how many entities, predicates, triples, or entity relationships you include, or how accurate they are. 

Search engines still need to understand where each piece of information belongs, how it relates to the surrounding elements, and which visual or functional component gives it meaning.

Dig deeper: Image SEO for multimodal AI

How centerpiece annotation affects rankings

In modern search, information quality alone isn’t enough. Information also needs to be presented within a layout that helps machines understand its boundaries, hierarchy, context, and purpose.

Google explained this concept through “centerpiece annotation,” describing visual annotations that help its systems better understand a document.

Martin Splitt from Google said the “centerpiece annotation” represents the “primary content” of a webpage. 

Later, documents disclosed during Google’s antitrust case showed that centerpiece annotation was also used to classify and rank news documents. 

The centerpiece annotation was primarily limited to about 400 characters, though those documents also reveal several other noteworthy details.

For example, below you can see how Google extracts the centerpiece annotation from HTML. The sentence is interrupted by unnecessary HTML elements, such as Facebook, email, Twitter (X), and Google+ share buttons.

HTML elements

In the next example from Google’s DOJ documents, proper HTML structure prevents share-button boilerplate from interrupting the centerpiece annotation, allowing Google to extract the content correctly.

What visual semantics looks like in practice

Below is a simple SEO case study. Although it involved 19 changes, the biggest ranking improvement came from one simple adjustment: moving a calculator component from the bottom of the page to the top, making it the centerpiece annotation.

The results of that change are shown below.

Metric Previous Current Increase / Change Success %
Total clicks 3.47 million 4.53 million +1.06 million clicks +30.5%
Total impressions 84.1 million 167 million +82.9M impressions +98.6%
Average CTR 4.1% 2.7% -1.4 percentage points -34.1%
Average position 8.9 8.5 Improved by 0.4 positions +4.5% improvement

This project closely connects visual semantics and textual semantics because it’s a programmatic SEO case study involving more than 100,000 pages.

At that scale, even a small sentence edit, component update, or layout adjustment is multiplied across every URL. That’s why Google re-crawled the entire website after the layout changes and why impressions and clicks increased afterward.

The project is a converter website that ranks for queries such as “2m to cm” and millions of similar numeric and metric variations. In this type of search environment, more than 10,000 competing websites provide essentially the same data and the same answer.

These websites have the same topical coverage and factual accuracy. The competitive advantage doesn’t come from providing a better answer because “1 meter to cm” has the same value everywhere.

It comes from retrieval cost, document understanding efficiency, internal PageRank distribution, and how clearly the answer is presented for Google’s initial ranking systems.

Google's Content Warehouse API leak includes similar semantic labels and annotations for webpages and PDF documents 
Google’s Content Warehouse API leak includes similar semantic labels and annotations for webpages and PDF documents 

In these types of queries, you can’t differentiate yourself by changing the answer. You differentiate yourself by changing how the answer is structured, annotated, prioritized, and visually presented.

That’s why changing the centerpiece annotation caused Google to reprocess the layout, rerank the pages, and further improve the site’s rankings.

Dig deeper: How to make products machine-readable for multimodal AI search

What is the cost of retrieval, and how does it relate to visual semantics?

“The cost of ranking a document” can’t be higher than the “cost of not ranking a document.” I introduced this concept years ago in one of my conference presentations. Google cares about search quality, but its systems also weigh quality against cost. If a website costs more to process than its quality justifies, Google will look for an alternative.

Google reduced the HTML file size limit to 2 MB and carried out large-scale deindexing following the December 2025 core update.

At the same time, it sent a clear signal to websites that scale AI-generated content without meaningful human effort. Google appears less tolerant of practices it accepted for years, and its indexing decisions are likely to become even more selective.

Retrieval costs increase when a webpage doesn’t clearly explain itself or fails to demonstrate sufficient relevance and responsiveness, especially around the “centerpiece annotation.” Google’s Content Warehouse API leak suggests the company truncates documents and predicts quality based on initial signals. If a document doesn’t meet relevance and responsiveness thresholds during those early evaluations, it won’t be considered a candidate.

During Google’s antitrust trial, Pandu Nayak, then Google’s vice president of Search, explained that Google doesn’t run its most computationally expensive algorithms on every webpage because it lacks sufficient click data. Instead, it first evaluates core topicality signals to determine whether a page is worth indexing and keeping as a candidate.

Nayak also explained that RankBrain-like algorithms are expensive to run, so Google reserves them for results that have at least one click, demonstrate strong topicality, and include annotations that justify the investment in crawling, rendering, evaluation, and further processing.

In other words, classifying documents by their layout, components, and structured information cards is a more efficient way to reduce retrieval costs while improving search quality.

Today, most large-scale content publishers rely on AI to generate more text. Far fewer invest in front-end and back-end systems that improve user engagement, interaction, and document understanding.

That distinction increasingly separates low-quality and high-quality sources. Low-quality sources primarily scale text. High-quality sources scale systems, layouts, components, structured information cards, and user interactions that help both users and search engines understand content more efficiently.

Below is Google’s concept of website representation vectors.

Google classifies websites using visual and layout-related embeddings and features to determine whether they resemble expert, apprentice, or amateur sources.

  • “For instance, the website classifications may include a first category of websites authored by experts in the knowledge domain (for example, doctors), a second category authored by apprentices (for example, medical students), and a third category authored by laypersons…”

Get the newsletter search marketers rely on.


How does Google’s helpful content system relate to visual semantics?

The helpful content system is a classifier that identifies which websites genuinely provide helpful information or meaningful engagement and which only imitate usefulness without fulfilling the searcher’s underlying intent.

Much of the SEO industry’s analysis of the helpful content system has focused on textual features. Early discussions centered on keyword stuffing, gibberish content, or adding “unique information” to improve information gain. However, many of the system’s algorithms appear to focus on the function and type of a source.

Google first classifies websites by their type rather than their content quality. That means the same content can rank differently on an affiliate website than it does on an ecommerce website. 

So how does Google distinguish among affiliate sites, aggregators, service providers, ecommerce sites, and SaaS platforms? The answer is visual semantics. What a page can do, or can’t do, is largely determined by its layout and page components.

The biggest distinction between relevance and responsiveness comes from engagement, not understanding.

Classifying search results by their page elements helps Google understand what type of document it's evaluating
Classifying search results by their page elements helps Google understand what type of document it’s evaluating

Google created systems such as neural matching to align the entity type and entity ID in a query with the most relevant documents. In simple terms, if the entity in the query doesn’t match the entity in the document, that page becomes less likely to rank. This is primarily about relevance.

Relevance alone isn’t enough. A document may rank because it’s relevant, but if it doesn’t support meaningful user actions, such as purchasing, comparing, ordering, reviewing, filtering, or watching, it isn’t responsive to the user’s actual task.

That’s why the helpful content system shouldn’t be viewed only as a system that evaluates page text. It also evaluates page function. A helpful page isn’t simply one that contains relevant words. It’s one that helps users complete the action, decision, or information-seeking task behind the query.

Google reinforced this idea by adding “misleading functionality” to its spam policies after the Helpful Content updates. A page can appear helpful by imitating a function without actually providing it.

For example, a page may suggest users can compare, filter, calculate, book, review, or purchase something even though those functions don’t genuinely exist. In those cases, the page may appear functional to both users and algorithms, but it isn’t truly responsive to the user’s task.

Google doesn’t classify websites only by page layout and design. It also appears to apply result-type constraints within the SERP. For example, a query such as “best women’s glasses” may return listicles, ecommerce category pages, product grids, videos, and commercial guides in the same results page.

To satisfy multiple search intents, Google can apply diversity constraints that limit how many ecommerce pages, listicles, videos, or other result types appear together.

Google’s DOJ documents include functions such as “max_total” and “BlogCategorizer,” which show how Twiddlers can classify results and limit the number of pages from the same cluster, category, or source type.

A similar annotation appears in the Google Content Warehouse API leak through the “WebrefFatcatCategory” module, which assigns categorical weight to a result.

In other words, Google doesn’t simply rank documents individually. It also classifies, clusters, and constrains results based on page type, source category, and categorical diversity. As a result, a page may be relevant enough to rank but still be limited by the overall composition of the SERP.

Even when a generated ranked entity list, such as a “best products” page, ranks successfully, it doesn’t rank simply because it’s a blog article. It ranks because it functions as a commercial resource. It helps users compare, evaluate, filter, review, and move closer to a decision. In that sense, Google can rank nonfunctional content when it effectively serves a functional category.

Viewed through this lens, “helpful” in the context of the helpful content system is closely aligned with “functional.”

The following case study demonstrates this principle. We moved identical content from an affiliate website to an ecommerce website, supported it with an integrated topical map, and saw rankings improve almost immediately.

The content itself didn't change. What changed was the function, context, and source type surrounding it. By placing the same information within a more functional, commercial, and task-oriented environment, Google interpreted the document as more useful for the user's search activity.
The content itself didn’t change. What changed was the function, context, and source type surrounding it. By placing the same information within a more functional, commercial, and task-oriented environment, Google interpreted the document as more useful for the user’s search activity.

How is click data used to rerank search results through visual semantics?

Google increasingly understands the purpose of a webpage through its layout, not just its text. As a result, click data is aggregated according to the type of source. Many SEOs assume that long clicks, or longer dwell times, signal quality. 

However, that’s not always true, according to Google’s research. Depending on the category, shorter dwell times can indicate a successful experience, while longer sessions may signal an “engagement trap.”

Below is Google’s reranking model, which applies different ranking and rank-modification models based on user behavior captured by its tracking components.

Another example comes from Google’s “Merging Search Engine Results” patent, alongside the “Twiddler’s anatomy” diagram revealed in the DOJ documents.

“Merging search engine results” is the name of the patent, which aligns with the “Twiddler” functionality above
Merging search engine results” is the name of the patent, which aligns with the “Twiddler” functionality above

Google also uses the concept of the “Life of a Click” to help engineers understand how search ranking algorithms interpret user behavior.

Taken together, these systems suggest that click data becomes a more meaningful classification signal when interpreted alongside a webpage’s design rather than through text alone.

Classifying documents by their visual structure can be more efficient than analyzing millions of documents, billions of word tokens, co-occurrences, named entity resolutions, attribute extractions, and value corrections.

If certain document layouts consistently generate stronger user satisfaction, Google can classify those pages as more helpful or functional. It can then use those signals to identify other documents with similar layout patterns, component structures, and interaction models.

This means topical authority doesn’t come only from a topical map that defines which topics to cover. It also comes from understanding which page layouts, component structures, information cards, comparison modules, and functional designs best match each topic, query, and search activity.

A proper topical map shouldn’t define only entities, attributes, predicates, and contextual relationships. It should also define the page type and functional layout needed to satisfy both relevance and responsiveness.

This leads to the concepts of coverage and domain-level classification. The following three examples illustrate this approach.

The first example is AudioToText.com, a sub-brand built around a single topic.

GSC Metrics of Audiototext.com. The third-party Semrush data is shown below.

Despite covering only one topic across 12 languages, or 13 pages in total, the site continues to grow in search visibility for three reasons:

  • Its exact-match domain reinforces relevance.
  • Its visual semantics improve responsiveness.
  • It earns its first clicks quickly, allowing Google to run more computationally expensive ranking systems sooner.

Click satisfaction from the other language versions may also reinforce the English version through cross-lingual information retrieval. 

Google can use webpage layout understanding and chain-of-reasoning to classify AudioToText.com as a “no-signup transcription tool” and rank it in AI Overviews. This suggests Google isn’t only reading the text. It’s also interpreting the page’s function, visual annotations, and interaction model.

In other words, Google can use agentic retrieval based on visual signals to understand what a page does and determine whether it deserves to rank for a specific query.

The Audiototext.com’s single-page topical map representation with the fundamentals are below.

The webpage was designed with minimal text while placing its primary conversion element, the content upload component, above the fold.

If that component were moved lower on the page or made smaller, rankings would likely decline, and text changes alone wouldn’t be enough to recover them.

Another example is attorneys.lexinter.net, which ranks primarily through a subdomain because its core content was moved there together with a filtering engagement component.

The primary domain didn’t meet the required thresholds, but moving the content to a subdomain with additional functional elements produced better results.

The same subdomain testing approach also worked for Pricelisto.com. Although most of the design and content remained the same, we added functions and annotations related to purchasing, comparing, examining, and reviewing.

Those functional additions made the pages behave less like passive content and more like task-completing commercial resources. As a result, the site avoided filters associated with the Helpful Content System.

The improvement didn’t come from changing the text. It came from changing how the document functioned, how users interacted with it, and how clearly Google understood the purpose of each page component.

Search engines try to reduce retrieval costs by avoiding computationally expensive algorithms whenever possible. As a result, domains affected by historical or domain-level signals may not receive a completely fresh evaluation immediately.

Testing on a subdomain can give Google a clearer reason to reprocess documents, reevaluate their layouts, and run more advanced evaluation systems. That makes it easier to determine whether improvements come from new designs, functionality, annotations, or document structures rather than from the historical state of the primary domain.

How is visual semantics related to the future of search?

Google is experimenting with fundamental changes to search results, including replacing the traditional search bar with new interfaces.

One example is its Jan. 29 patent, “AI-generated content page tailored to a specific user.” The patent describes generating a landing page that uses visual segmentation, annotations, and generative AI to satisfy a user’s query.

The patent places significant emphasis on "landing page score," using click data and explicit user feedback signals
The patent places significant emphasis on “landing page score,” using click data and explicit user feedback signals

In other words, Google can use visual semantics not only to rank web documents but also to construct new types of search results.

Dig deeper: Google patent hints it could replace your landing pages with AI versions

Google’s patent work is often complemented by its research. For example, the paper “Neural Design Network: Graphic Layout Generation with Constraints” explores how systems can understand, classify, and even generate webpage layouts to improve search performance.

This suggests that layout isn’t only a design consideration. It can also serve as a retrieval, classification, and ranking signal.

Google’s multimodal document understanding also connects to its latest announcement, Google Embedding 2, which uses generative neural networks to understand and vectorize text, images, videos, audio, and documents.

This matters because different versions of the same web document can be compared through their vector representations. Doing so makes it possible to evaluate how well Google understands layout differences, visual structure, and document-level meaning.

In other words, layout changes aren’t merely visual. They can also produce different vector representations, which may affect how a document is understood, classified, and retrieved.

Below is Google’s example of the neural network process for understanding page layouts. The centerpiece annotation that helps classify a webpage as an ecommerce category page, product page, or SaaS page comes from these types of labeling systems.

In the future, Google could apply these same principles to construct its own landing pages from multiple search results.

The patent shown below also illustrates how Google could adjust SERP features based on an entity’s primary attributes. That suggests search results aren’t simply ranked and displayed. They can also be reorganized, redesigned, and presented as dynamic interfaces based on the entity, query intent, and available document structures.

Centerpiece annotation and query processing

Google classifies and augments queries differently from how people naturally think about them. That means one of the most important parts of creating a topical map is understanding search terms the way Google’s systems do and augmenting them accordingly. This process is called query semantics. Below is an example of query augmentation from ChatGPT.

In this example, we searched for “best search engine optimization information sources,” and GPT expanded the query as follows:

  • Best SEO information sources: search engine optimization resources Google research, patents, SEO blogs

If you perform a search in ChatGPT, open the Network tab in Chrome DevTools, filter for XHR requests, and inspect the JSON file associated with the https://chatgpt.com/backend-api/conversation/6a* path. Look for search_model_queries, which shows what the system actually searches for.

Google also has a patent called query augmentation, shown below.

The patent is attributed to engineers, including Krishna Bharat and Anand Shukla. These names are significant because they also appear on patents and systems related to AI Overviews and AI Mode.

For example, the “Search with Stateful Chat” patent includes query augmentation as one of its steps, and its terminology and inventors overlap with this system.

The centerpiece annotation is the primary visual annotation that reflects a webpage’s purpose, function, and context. The context created through the augmented query needs to align with that centerpiece annotation.

The following case study shows how I classified query variations and their contexts across different document types, each with a distinct purpose, function, and visual structure, for a local service directory.

Let’s use “air conditioner” queries as an example. Each query variation should be matched with the appropriate page type, layout, and function.

  • Experience queries require a forum-style layout. For a query such as “How do I repair my AC?” the intent is experience-based. A forum structure works best because users expect real problems, answers, troubleshooting paths, and personal experiences. This content can also live on a subdomain to separate experiential content from the main commercial website.
  • Local service queries require a directory page. For “Air conditioner installation in [City],” the intent is local and service-oriented. The best page type is a local directory or listing page with providers, service areas, ratings, contact options, and conversion elements.
  • Price queries require a hybrid layout. For “air conditioner installation prices,” the intent is both informational and commercial. The page should provide an immediate answer with average prices, cost factors, and price ranges while also presenting local providers, comparisons, and quote-related elements.
  • Instructional queries require an informational layout. For “How to install an air conditioner,” the intent is instructional. The page should minimize local service elements and instead focus on a step-by-step guide, required tools, safety considerations, visuals, and practical instructions.

In short, a topical map should define not only which topics to cover but also the appropriate layout, components, and page function for each search activity. The following example shows some of the early results from this project after classifying query augmentation models for different query variations.

Early GSC results for the same brand.
Early GSC results for the same brand.

If there’s no need for a separate page for the [Local], [Service], [Forum], or [Instructional List] intent, we simply prune it. If other pages are too similar, we merge them.

As a result, the number of pages decreases along with retrieval costs, while PageRank concentration and relevance per document increase. Below are four closely connected components:

  • Mock-up design in draw.io.
  • Production design in Figma.
  • Topical map for different query types.
  • Content brief aligned with the Figma and draw.io designs.

Early on, we defined the topical authority formula as:

  • Historical data x Topical coverage

Later, we expanded it to:

  • Historical data x Topical coverage ÷ Cost of retrieval

Today, I’d extend the formula with one additional factor:

  • ((Historical data x Topical coverage) ÷ Cost of retrieval) x Right visual annotations

Even if you have the lowest retrieval cost, the highest topical relevance, the broadest topical coverage, strong accuracy, the longest duration of satisfied click data, and positive historical performance, none of it matters if the centerpiece annotation is wrong or the page isn’t functional.

Google’s ranking system largely functions as a decision tree. If the first decision-making layer rejects a website, the later evaluations, tests, and reranking processes won’t occur.

To maximize your chances of ranking from the start, visual annotations should be optimized just as carefully as the page’s text, images, and links.

Below is a conceptual model of this system.

A website consists of “letters, pixels, and bytes.” Data2Website is the process of turning a dataset that Google’s algorithms favor into a website by combining textual and visual semantics through those letters, pixels, and bytes.

The example above shows how a local law firm benefited from a topical map, semantically optimized content briefs, specific sentence structures, and visual design decisions.

The Semrush results below show the impact on the firm’s local rankings.

We previously applied the same principles to another ecommerce website.

If you examine the screenshots closely, you’ll see that the same principles carry over from an ecommerce design to a local service provider.

For every attribute within an entity-seeking query, such as “best law firm in Houston” or “birth test kit prices,” you can classify those attributes within the query network and organize them according to their importance.

Some attributes require review components, while others require directly commercial components.

Below are two design examples from the sibling websites Morethanpanel.com and StreamingMafia.com. Their above-the-fold and below-the-fold sections are structured similarly, covering different types of user engagement and functionality.

The above-the-fold area is often referred to as the macro-context because it contains the main content. Google’s Quality Rater Guidelines use the concept of main content to emphasize the importance of relevance, accuracy, and completeness in this section.

The below-the-fold area corresponds to what Google’s Quality Rater Guidelines describe as supplementary content, which we refer to as the micro-context. This section typically contains less important attributes and most internal links.

The next example shows the mock-up design and the distribution of factual content, opinionated content, structured content, and unstructured content.

Google doesn’t always prioritize factual or opinionated content, or structured versus unstructured content. Instead, it evaluates these characteristics based on how the search query is augmented. To improve language relevance, we distribute different types and formats of content using different visualization, verbalization, commercialization, and contextualization techniques.

The following example applies the same approach to the second website in the same industry, together with its topical map, content briefs, and authorship rules.

Algorithmic authorship can be explained through the research paper “Are LLMs Reliable Rankers?” It means writing content according to predefined sentence structures and rules. For example, the research shows that the “Rank anything first” framework increased rankings by 20% to 60%.

The system evaluates which words should follow one another to determine how relevance changes. It performs retrieval within a generative retrieval system and identifies the entity-attribute-value triples that best improve relevance. In the example above, “material” is selected as the attribute and “steel” as the value because they strengthen relevance within that context.

  • Factual content: Supports expertise-focused queries.
  • Opinionated content: Supports experience-focused queries.
  • Structured content: Supports attributes such as symptoms, advantages, and benefits.
  • Unstructured content: Supports concepts such as definitions, processes, and importance.
  • Visualization: Presents content using the appropriate semantic attributes.
  • Commercialization: Adds functional components that help users complete their tasks.
  • Contextualization: Maintains relevance by aligning content with the query.
  • Verbalization: Converts visually important information into text that LLMs and search engine crawlers can understand.

Depending on the query, Google may prefer opinionated and unstructured content, factual and structured content, or other combinations supported by different visualization, commercialization, contextualization, and verbalization techniques.

The following example from the online dating industry shows how different webpage components can improve relevance and responsiveness at the same time.

The next examples illustrate different ways to visualize content.

Comparing these two sections, you’ll see that one answer is highly factual, while the other, distinguished by a different background color, is more conversational and opinion-based.

We can create a Q&A component and add opinion-based content as forum-style discussions at the bottom of the page.

We can also ask users questions and let them contribute answers through voting, allowing those responses to be verbalized into content that is continuously updated.

Below is what we call the preceding question component. It reframes the original question using a semantically similar concept and gradually shifts the content from factual to more opinion-based.

The next example shows a horizontal tab component that distributes internal links to related headings, increasing contextual coverage.

The following Semrush data shows the early and later results for the URLs we modified.

The patents and research behind visual semantics

At this point, we’ve introduced the key concepts, definitions, and website examples needed to explain visual semantics.

We could explore these examples, processes, and implementation details in much greater depth, but every conceptual discussion begins with understanding where Google is heading.

Many of Google’s advances in query semantics, visual semantics, Gemini, and AI Search are driven by two influential engineers: Dr. Marc Najork and Michael Bendersky. They are among Google’s most frequently cited researchers in recent years and have played major roles in shaping the company’s AI-related direction.

They are also listed as inventors on the Layout-Aware Document Understanding and Structured Information Cards patents.

Another important contributor is Alexander Grushetsky, who identifies himself as the founder of RankLab, Google’s internal end-to-end ranking platform.

He’s worth mentioning because he’s frequently cited alongside Bendersky and Najork in foundational patents and research papers.

Grushetsky also worked with Bendersky and other Google engineers on item-ranking models based on item types, attribute sets, and attribute values. We’ll explore what RankLab represents in more detail another time.

Today’s search engines and large language models increasingly rely on visual semantics as part of their vectorization and embedding-based ranking systems.

Even the original Transformer research described extending these ideas to web documents and their layouts.

Years later, that vision became reality through WebRef, Google’s Web Page Transformer.

WebRef vectorizes webpages using not only their text but also their visual layout, page components, HTML structure, and overall document context.

Whether your rankings depend primarily on external PageRank, branded search demand, or internal signals such as semantics, a page’s visual context still carries ranking weight alongside its textual relevance.

Read more at Read More