As a content strategist, your pages might be highly relevant and authoritative. But does that automatically mean their content will be cited in AI-generated answers?

Not necessarily. In our experience, generative search engines cite specific sections of content, rather than pages as a whole. So, even when everything else looks right, your content may not be cited if the sections themselves aren’t safe to quote.

In this article, we’ll explore what makes a section citable and walk through five practical checks to help you identify which parts of your content are more likely to be cited.

Short on time? Here are the key takeaways: 

Why assessing individual sections matters

Not every section of a page has the same potential to be cited. Some are well suited to citation, others may need more clarity or context, and many are unlikely to be cited at all.

The difficulty is knowing which sections fall into which categories. Without that visibility, content teams risk spending time optimizing content without knowing which sections need attention.

Understanding what makes a section citable is the first step toward identifying where improvements may be needed.

What “citable” actually means

A citable section needs to stand on its own. Ideally, this means it should be possible for a generative search engine to:

• Retrieve it independently, without relying on the rest of the page.
• Understand it in isolation, without needing surrounding context.
• Quote it safely, without introducing risk.

Content teams don’t have to make every section long, comprehensive or SEO-optimized in order for it to be citable. Even well-written content may not be suitable for citation if it relies too heavily on context elsewhere on the page. 

Ultimately, citability comes down to whether a section can be safely used as a source, rather than how well it’s written.

The section test: five checks for your content

To assess how citable your content is, choose a section of your page and work through the following five checks.

In our experience, sections that fail any of these checks are less likely to be cited in AI-generated answers.

1. Does the section answer one question?

Start by asking yourself: What question does this section answer?

If your answer is “a few things,” “it depends,” or “it sets context,” the section may not be focused enough to be cited independently.

Citable sections should have one clear purpose.

2. Can the section be read in isolation?

Imagine someone reading a section on its own, without the rest of the page for context.

Ask yourself:

• Are there pronouns like this, it or they without a clear reference?
• Does the section rely on a definition introduced earlier?
• Does it use phrases like “as mentioned above”?

If so, the section may not be sufficiently self-contained to be cited independently.

From what we’ve observed, citation systems tend to favor explicit repetition over elegant writing that relies on surrounding context.

3. Are the facts inside the section?

A citable section should contain the information needed to understand and use it independently. This includes:

• Entities: Names, models and standards.
• Units: Measurements, percentages and versions.
• Constraints: When, where or under what conditions something applies.

For example, consider a section that says:

“The torque value is listed in the table below.”

If the actual value isn’t included in the section, it may be less likely to be cited independently.

4. Does the section stay on topic?

Sections can sometimes drift away from their main topic, making them less focused and harder to retrieve.

Watch for:

• Background information mixed with instructions
• Opinions mixed with facts
• Explanations mixed with exceptions

This kind of topical drift could reduce embedding precision, retrieval confidence, and the likelihood of a section being cited.

Citable sections should stay focused on one clear topic.

5. Is the section the right size?

The length of a section can also affect how suitable it is for citation.

If a section is too short, it may lack important constraints or context, making its meaning ambiguous. If it’s too long, it may cover multiple ideas, diluting the focus of the information.

In our experience, the most citable sections tend to fall somewhere in the middle. They focus on one clear idea, provide enough detail to be understood independently and avoid unnecessary filler.

While there’s no simple way to measure the ideal section length, generative systems can still assess whether a section contains the information needed to support an answer.

TL;DR: the citable section checklist

When reviewing your content, use this quick checklist to assess whether a section is suitable for citation:

• One clear question: Does it answer one specific question?
• Self-contained: Can it be understood without surrounding context?
• Necessary facts: Does it include the information needed to support an answer?
• Topical focus: Does it stay focused on one topic?
• Appropriate length: Does it provide enough detail without losing focus?

If a section falls short on any of these points, it may need further work to improve its chances of being cited.

How chunkability affects section-level citability

The five checks help you assess individual sections of your content. But how does this look across an entire page?

In practice, sections of the same page can have very different citation potential. For example:

Section	Citation potential
Overview	Less likely to be cited
Background	Less likely to be cited
Step-by-step process	More likely to be cited
Specifications	More likely to be cited
History	Less likely to be cited

The important distinction here is structural. A section’s citation potential depends on whether it can stand on its own and provide information that can be retrieved and used independently, rather than simply on how well the content is written.

Learn more about chunkability and how it affects citation

Why good writing isn’t always enough

Even well-written sections can struggle to earn citations in AI-generated answers.

This can happen when content prioritizes flow over precision, assumes the reader has surrounding context, avoids repetition, or takes too long to introduce key facts.

Instead, generative systems may tend to favor sections that provide:

• Early clarity: Key information is introduced upfront.
• Explicit restatement: Important details are repeated where needed so the section makes sense independently.
• Dense facts: Relevant information is included within the section itself.
• Predictable structure: Information is organized clearly and consistently.

These characteristics can make a section easier for generative systems to retrieve, understand and cite—even when the original writing is already strong.

Why traditional SEO tools can’t assess section-level citability

Traditional SEO tools typically focus on metrics such as page performance, keywords, and links. But they don’t tell content teams whether individual sections of a page are suitable for citation.

Assessing section-level citability requires looking at different factors, including:

• Section cohesion: Whether the content stays focused on one topic.
• Boundary clarity: Whether the section has a clear beginning and end.
• Self-contained meaning: Whether it can be understood without surrounding context.
• Citation safety: Whether its information can be quoted without introducing risk.

Without this level of insight, content teams are often left guessing which sections need improvement. And that makes it difficult to optimize content effectively at scale.

What comes next?

To summarize, your content may already contain highly citable sections. But without visibility into which sections are suitable for citation, it’s hard to know where improvements are needed. These five checks aim to help shine a light on where the issues might lie.

In the next article in our GEO series, we’ll explore whether authority is holding your content back or whether the way it’s structured is.

Understanding the difference can help you avoid spending time optimizing the wrong things.

The post Which Sections of Your Content Are Actually Citable? appeared first on Lumar.

Your page ranks well in traditional search. But does that automatically mean it’s eligible for AI search? 

Not necessarily. In our experience, eligibility describes whether a page is considered as a potential source for a particular query. Ranking position alone won’t tell us whether it is.

In this article, we’ll explore what eligibility means in generative search and walk through six practical checks to help you assess what may affect your pages’ chances of being considered.

Only got 30 seconds? Here are the key takeaways:

Eligibility is not the same as ranking

A page’s ranking position, traffic potential, link authority, or “AI-friendly” formatting doesn’t guarantee that it will be considered as a source for a generative query.

From what we’ve seen, eligibility is closer to a binary decision: whether a page makes it into the pool of potential sources for a particular query. In that sense, it’s either considered, or it isn’t.

Unlike traditional search rankings, generative search doesn’t give you a visible position that tells you how close your page is to being used as a source.

Why eligibility matters

Generative search systems have practical constraints. Retrieving and processing information is expensive and takes resources, and context is limited. They also have high trust requirements. 

From what we’ve observed, eligibility helps systems reduce the search space and manage risk by narrowing the set of potential sources considered for a query.

Considering every available page and passage for every query would be impractical. From what we’ve observed, eligibility helps systems reduce the search space and manage risk by narrowing the set of potential sources considered for a query.

This is why eligibility is a useful concept when investigating generative search visibility. It helps us consider whether a page is likely to be among the potential sources for a particular query.

The eligibility test: run these six checks on your page

We can’t see exactly how every generative search system selects its potential sources. But based on what we’ve seen, we can examine several factors that may affect whether a page is considered for a particular query.

The following six checks look at topical clarity, query alignment, page focus, credibility and the query itself to help identify possible weaknesses:

1. Can you state the page’s main topic in one sentence?

Start by asking yourself what the page is about. Can you clearly state its main topic in one sentence?

If your answer includes:

• “And”
• Multiple concepts
• Several audiences

take a closer look at whether the page’s topical focus is clear.

This doesn’t mean a page can only cover one topic. In fact, Google says its systems can understand the nuance of multiple topics on a page. A page can cover several related concepts or audiences while still having a clear overarching subject. The point of this check is whether that overarching subject is easy to identify.

2. Does the title clearly reflect the query you’re targeting?

Next, compare your title with the language and intent of the query. 

Check whether:

• The main topic is explicit.
• The title accurately describes what the page delivers.
• The wording reflects how people might search for the information.

You don’t need to phrase every title as a question or match a query word for word. But if the title is clever or vague at the expense of clarity, it may make the page’s relevance to the query less immediately clear.

3. Is the core topic introduced early?

Look at the title, H1 and opening section. Is the main concept clearly introduced there, or is it buried under context and storytelling?

If readers have to work through a long introduction or extensive background before reaching the main subject, consider bringing the core topic forward.

This aligns withGoogle’s guidance for generative AI search, which recommends making content easy to follow and organizing it into clear paragraphs, sections and headings. 

Google also says its generative features retrieve relevant pages before reviewing specific information from those pages to help generate a response.

4. Does the page have a clear primary purpose?

Once you’ve established the page’s main topic, consider the page as a whole. Does the content support one clear primary job?

Look for:

• Unrelated FAQs that take the page in a different direction.
• Background sections that are longer than the answer readers came for.
• Multiple “primary” headings or sections that serve substantially different purposes.

Comprehensive content isn’t the problem. A page can answer several related questions, but those elements should work together to support its primary purpose. This is consistent withGoogle’s advice to focus on what users actually want, rather than creating content around every possible variation of how people might search.

5. Would this page be considered safe and credible to cite?

From what we’ve observed, source authority can play a stronger role in whether content is considered for queries where getting the answer wrong carries greater risk.

Ask:

• Is the query particularly sensitive or safety-critical?
• Are authoritative sources, such as manufacturers or institutions, consistently appearing?
• Is the page clearly credible for this particular topic?

If authority expectations are high, a less authoritative source may find it harder to be considered, even when its content is relevant.

6. Have you assessed eligibility against specific queries?

From what we’ve observed, eligibility appears to be query-dependent rather than a permanent status attached to a page. A page may therefore be considered for one query but not another.

For example, an introductory guide to oil changes might address “how oil changes work” without providing the vehicle-specific information needed for “Honda Civic oil drain torque.”

Compare each target query with what the page actually covers. Does it provide the information needed to answer that particular query, at the right level of detail?

Assessing pages against specific queries gives you a more useful starting point than treating eligibility as a permanent, site-wide status.

Signs that eligibility is worth investigating

You might have pages that rank well in traditional search but rarely appear as cited sources in generative search. Or you may see the same page appear for some relevant queries but not others.

These patterns don’t prove that eligibility is the problem. But they can give you a reason to investigate whether the page is likely to be considered as a potential source for the queries you’re targeting.

Why traditional SEO metrics don’t tell us this

Traditional SEO tools can tell us a lot about how a page performs in search, including whether it’s indexed, where it ranks and the organic visibility it receives.

What they can’t tell us is whether that page was considered as a potential source for a particular generative query. A strong ranking therefore doesn’t necessarily explain why one page is cited and another isn’t.

That leaves us with an important measurement gap. We can use the checks above to investigate whether eligibility may be an issue, even though generative search systems don’t directly expose their candidate selection decisions.

Without that visibility, the aim is to make more informed decisions about what to investigate and optimize, rather than assuming rankings or citations tell us the whole story.

What comes next

Eligibility describes one early part of the process: whether a page is considered as a potential source for a query. But being considered doesn’t guarantee that its content will be retrieved or ultimately used in a response.

Next, we’ll look more closely at semantic relevance, including the different roles it can play in whether content is identified, retrieved and ultimately used in a generative response.

The post Is Your Page Even Eligible for Generative Search? appeared first on Lumar.

A lack of visibility can have several possible causes, and ranking position alone won’t necessarily tell you which one you’re dealing with.

In this article, we outline a practical GEO diagnostic checklist to help you investigate what may be limiting your content’s visibility in generative search.

Short on time? Here are the key takeaways:

A GEO Diagnostic Checklist:

1. Is your page likely to be eligible?

A useful place to start is candidate page selection. Here, we want to understand whether a page is likely to be considered for a particular query.

For Google’s generative AI features, there are some basic eligibility requirements. According to Google’s guidance, a page must be indexed and eligible to appear in Google Search with a snippet, although meeting those requirements doesn’t guarantee that Google will serve the content.

Beyond those basic requirements, we can also consider how clearly the page aligns with the query and topic you’re targeting.

Check:

□    Can you clearly describe the primary topic of the page in one sentence?

□    Does the title clearly reflect the topic and align with the query intent?

□    Does the page stay focused on that topic, or does it cover several loosely related subjects?

If those questions are difficult to answer, the page may be a weaker candidate for the query you’re targeting. A page can rank well in traditional search without necessarily being a strong candidate for a particular generative query.

2. Does the page signal clear aboutness?

Candidate eligibility is one part of the diagnosis. We also want to consider how clearly the page communicates what it’s actually about and how closely its overall topic aligns with the query. 

Check:

□    Is the core concept introduced early enough to establish what the page is about?

□    Does the opening content make sense without relying on information much further down the page?

□    Do the secondary sections support the main topic, or do they start to pull the page in different directions?

Long introductions, extensive background information, or loosely related sections can make the page’s main topic less clear. If the content drifts from its primary focus, its page-level semantic signals may become less precise.

3. Are your chunks actually retrievable?

A page being a potential candidate for a query doesn’t necessarily mean its individual passages will be retrieved. That makes the organization of information within the page worth examining.

Google has confirmed that its generative AI features use retrieval techniques to find relevant content from its Search index and review specific information from retrieved pages to help generate a response.

That makes it useful to look at how clearly relevant information is organized within the page, and whether it forms a coherent passage that could be useful for the query.

Check:

□    Does each section cover a clear, focused topic?

□    Are related steps, facts, or specifications grouped together?

□    Is important information contained within a coherent section, rather than spread across several parts of the page?

When an explanation is split across multiple sections, it may be harder for a retrieval system to identify a single passage that contains the relevant information.

This doesn’t mean creating artificially small pieces of content for AI systems. Google says there’s no requirement to break content into tiny pieces for AI to understand it, while still recommending content that is well organized into clear paragraphs and sections.

4. Can a chunk stand on its own?

Once you’ve looked at whether information is organized into retrievable passages, the next check is whether those passages contain enough context to make sense on their own.

Check:

□    Can the passage be understood without heavily relying on the surrounding content?

□    Are references such as “this” or “it” clear when the passage is read independently?

□   Do passages include the relevant entities, units, constraints, or other context needed to understand the information?

Try reading a passage in isolation. If its meaning becomes unclear without the surrounding content, it may be less suitable for use as supporting evidence in a generative response.

5. Could authority be limiting your visibility?

If the content looks strong across the areas we’ve checked so far and you’re still not showing up in AI-generated answers, authority may also be worth investigating. As we explored earlier in the series, its influence can vary depending on the query and the level of risk involved.

Check:

□   Is the query particularly sensitive or safety-critical?

□   Are authoritative sources, such as manufacturers or institutions, consistently being cited?

□   How does your source’s authority compare with the sources appearing for the same query?

If more authoritative sources consistently appear, authority could be limiting your ability to compete for that query. But it’s worth checking the content-level factors above before assuming authority is the cause.

6. Are you addressing the right problem?

Once you have an idea of what may be limiting your content’s visibility, consider whether the changes you’re making actually address it.

Improving one area of a page won’t necessarily resolve an issue elsewhere. For example, building authority may do little to help a passage that lacks enough context to make sense on its own. Similarly, restructuring a section may improve its chunkability without addressing weak alignment between the page and the query.

Check:

□   What specific problem are you trying to address?

□   Does the change you’re making relate directly to the problem?

□   Are you relying on general “AI-friendly” changes without a clear reason for making them?

The aim is to connect each optimization to the issue you’ve identified. If candidate eligibility looks weak, investigate the signals that may affect candidate selection. If the problem appears to be at a passage level, look more closely at semantic relevance and chunkability, including whether the passage contains enough context to stand on its own.

The GEO diagnostic summary

A lack of visibility in generative search can have several possible causes. The checks we’ve worked through let you investigate systematically and narrow down the areas that deserve a closer look.

Taken together, this checklist asks:

□   Is the page likely to be a candidate for the query?

□   Does the page clearly communicate what it’s about?

□   Is important information organized into coherent passages that could be retrieved for the query?

□   Do those passages contain enough context to make sense on their own?

□   Could authority be limiting the page’s visibility for this particular query?

□   Do the changes you’re making address the problem you’ve identified?

While these questions won’t necessarily reveal exactly why a generative system did or didn’t use a particular source, they can give you a clearer basis for investigating visibility and deciding what to optimize next.

What comes next

We’ve touched on candidate page selection throughout this checklist. The next article explores what may influence whether a page is considered for a particular generative query.

The post Why Your Content Isn’t Getting Cited: A GEO Diagnostic Checklist appeared first on Lumar.

Over the course of this series, we’ve looked at the different factors that can influence whether content appears in generative search.

Now, we need to think about how we measure them. How can we understand whether a page is being considered for a query, or whether its content is likely to be retrieved and cited?

In this article, we’ll bring those factors together and look at what a GEO analytics stack could actually look like.

Only got 30 seconds? Here are the key takeaways:

The problem with traditional SEO analytics

Traditional SEO analytics largely focus on rankings, assuming an indexed page is eligible to compete and that its position explains its visibility.

In our experience, generative search complicates this. A page can be indexed and highly relevant to a query without its content appearing in a generated response. That’s because other points in the retrieval process can affect whether it is considered or used.

Simply put, the reason a page doesn’t appear may not be something a ranking metric can show us. If we only measure the outcome, we have little insight into the factors that may have affected its visibility.

A GEO-aligned mental model

A GEO analytics stack, therefore, needs to account for the different factors that can affect whether content appears in a generative response. That starts with whether content is available and accessible, alongside the four areas we’ve explored so far in this series.

Some of these apply at particular points in retrieval, while others can influence more than one part of the process. For analytics purposes, we can organize them into five layers: 

1. Availability and indexability
2. Candidate page selection
3. Semantic relevance
4. Chunkability and retrieval
5. Authority and trust

For each layer, we’ve included an importance score based on how much we believe it can influence GEO performance, alongside our confidence in the role it plays within this model.

The GEO analytics stack (layer by layer)

1. Availability & Indexability layer

Before looking at the other layers, we need to know whether the content is accessible in the first place.

That means looking at crawlability and indexability, along with whether the content can be successfully rendered and extracted from the page.

Problems this early on can prevent a system from reliably accessing the content it needs, making the other factors in the stack largely irrelevant until they’re resolved.

• Importance: 9/10
• Confidence: 9/10

2. Candidate page selection layer

At the candidate page selection layer, we want to understand whether a page is likely to be considered for a particular query.

Here, we’d want to look at: 

• Lexical alignment in the title and H1,
• Page-level semantic alignment with the query, and
• How clearly the page is focused on a single topic.

We could also look at whether authority may affect the page’s eligibility for that query.

Together, these could feed into a Candidate Eligibility Score for each query, giving us a way to assess how likely a page is to be considered a candidate.

• Importance: 9/10
• Confidence: 9/10

Learn about candidate page selection and why traditional SEO tools don’t measure it

3. Semantic relevance

Semantic relevance needs to be measured in context because it can play different roles across candidate selection, retrieval, and citation.

For candidate selection, that means measuring page-level aboutness and how closely the overall page aligns with the query, whereas retrieval looks at the similarity of individual chunks. Citation focuses on how precisely the retrieved evidence supports the cited information.

A relevance score therefore needs context. Page-level aboutness, chunk-level similarity and citation-level precision are measuring different things at different points in the process.

• Importance: 9/10
• Confidence: 8/10

Discover why semantic relevance serves multiple purposes across retrieval stages

4. Chunkability and retrieval

At the chunkability and retrieval layer, we want to understand how suitable individual passages are for retrieval and citation. That means looking at:

• Topical cohesion
• Boundary clarity
• Answer density
• Whether a chunk contains enough context to make sense on its own.

These factors could feed into a ‘Chunk Citability Score’, designed to indicate how suitable a passage is for use as supporting evidence in a generative response.

• Importance: 10/10
• Confidence: 9/10

Learn about why some content gets cited, and how to spot whether your content is failing

5. Authority and trust

The authority and trust layer looks at whether the source itself may affect its eligibility for a particular query. As we covered in the previous article, the influence of authority can vary depending on the query and the level of risk involved.

To measure this, we could look at:

• Domain provenance 
• How well the source’s level of authority aligns with the risk of the query
• The authority gap between the source and others competing for the same query 

This gives us a way to identify where authority may be limiting a source’s ability to compete, even when its content performs well in other areas of the stack.

• Importance: 8.5/10
• Confidence: 8/10

Learn more about authority in GEO and why it’s a filter, not a score

What a GEO dashboard should look like

Bringing these measurements into one dashboard would give GEO teams a clearer view of where content may be succeeding or struggling. Alongside ranking position, we could start asking more specific questions:

• Which pages are likely to be eligible for which queries?
• Where could semantic relevance be falling short?
• Which chunks are most suitable for citation?
• Where could authority be limiting eligibility?

Together, these give us a more detailed view of GEO performance than the final visibility outcome alone.

Making GEO performance actionable

Separating GEO analytics into layers helps because different visibility problems can point to different causes. For example, a page that appears to have an eligibility problem calls for a different response than content that appears eligible but lacks the precision needed for retrieval or citation.

Being able to make that distinction gives teams a clearer idea of where to focus their efforts. Without it, we risk optimizing for the outcome without understanding what may have caused it.

Bringing the framework together

Taken together, these layers give you a more complete way to measure GEO performance than citations alone.

They also give us a natural next question: what happens when your content still isn’t being cited?

Next, we’ll turn this framework into a practical diagnostic checklist to identify where the problem might lie.

The post The GEO Analytics Stack: What We Should Be Measuring (But Aren’t) appeared first on Lumar.

So far, we’ve looked at page eligibility, the different roles semantic relevance can play, and what makes individual passages easier to retrieve and cite.

Next, we’re covering authority.

We tend to misunderstand authority in generative search. But in our experience, getting it right matters, with important implications for whether and where your content appears in AI search.

In this article, we’ll explore how authority appears to work in generative search, where our confidence in that understanding comes from, and what it means for GEO.

Only got 30 seconds? Here are the key takeaways:

What we mean by authority in GEO

In SEO, authority is often treated as something that can be measured and optimized over time to help a page or domain compete better in search results. The focus, therefore, tends to be on how authoritative one source is compared with another.

In generative search, our experience suggests that comparison is less important. Instead, what matters is whether the source is considered suitable and trustworthy for the query. 

That’s why we describe authority as a filter in GEO. 

Where authority can influence retrieval

From what we’ve seen, a source’s authority and trustworthiness can come into play early in retrieval and may affect whether its content is considered for a query. In some cases, these signals may even contribute to a source being excluded from consideration.

This could help explain why highly relevant pages don’t always surface, while particular domains appear consistently across certain types of queries. A better explanation can’t compete if its source is excluded before the passage itself is considered.

Why authority depends on the query

How strongly authority influences that process appears to depend on the query itself. From what we’ve observed, the greater the potential harm of getting an answer wrong, the more important it becomes for the information to come from a trusted or qualified source.

We see a similar principle in Google Search. Google says its systems give more weight to content that aligns with strong Experience, Expertise, Authoritativeness, and Trustworthiness (E-E-A-T) for topics that could significantly affect people’s health, financial stability, safety, or the welfare of society.

Authority by Query Class (Observed Behavior)

In practice, this can range from a hard gate for high-risk queries to having minimal influence over more subjective or creative queries:

What signals may contribute to authority

We don’t know the exact signals or weighting generative systems use to assess source authority, but we can be more confident about some signals than others.

High confidence

Source provenance is one of the strongest signals we look at. For queries where trust and accuracy matter, this can include official documentation, regulatory bodies, institutional publishers, and first-party sources such as original equipment manufacturers (OEMs).

From what we’ve observed in generative search, these sources are particularly prominent for high-risk and YMYL queries, where less authoritative sources may struggle to be considered.

Medium confidence

Beyond source provenance, we believe other signals may contribute to authority, although the evidence for their role is less conclusive. These include consistency with trusted sources, citation overlap, topic ownership, and historical trust.

Rather than determining eligibility on their own, these signals appear more likely to strengthen the case for a source being considered alongside other indicators of authority.

Lower confidence 

Confidence is lower for signals such as author credentials, structured data like author and reviewedBy, and engagement metrics. While each could plausibly contribute to how authority is assessed, their role isn’t clearly confirmed.

If they do contribute, we expect them to play a secondary role and be unlikely to override stronger, domain-level trust signals.

Confidence summary

How authority in GEO relates to E-E-A-T

The signals we’ve covered have clear parallels with E-E-A-T, but E-E-A-T is most useful here as a way of thinking about authority rather than a formula for how generative systems assess it.

In our model, the strongest parallels are around trust, provenance, and reducing the risk of citing unreliable information. Factors such as brand popularity or link volume alone tell us much less about whether a source is suitable for a particular query.

Ultimately, authority in GEO comes down to whether a system has enough reason to trust a source and use its information in a response.

Why smaller sites can still break through

Authority can strongly influence which sources are considered, but that doesn’t mean smaller sites are always excluded. When authoritative sources don’t adequately cover a query, other sources may have more opportunity to compete.

This is particularly relevant for niche or operational queries, where a smaller site may provide a more precise answer than established sources.

In these cases, the factors we’ve covered earlier in this series become especially important. Semantic relevance, chunkability, and a clear, specific answer can give a smaller source a chance to surface where the authoritative corpus has a gap.

That opportunity still depends on the gap being there. When a query is already well served by highly authoritative sources, these other strengths may not be enough to overcome the authority gap.

What authority means for your GEO strategy

If authority can act as a filter, it changes where GEO efforts are most likely to have an impact. For YMYL and other high-risk queries, you may not be able to optimize your way past a significant authority gap, however relevant or well-structured your content is.

That makes choosing where to compete particularly important. Instead of focusing on highly competitive head terms already well served by authoritative sources, look for gaps where existing sources don’t fully answer the query. As we’ve seen, niche and operational queries can create more room for precision to matter. 

For non-YMYL queries, authority can still influence which sources are considered, but clarity, relevance, and chunkability have more influence on which passages are ultimately retrieved and cited.

In that sense, authority can set the ceiling for where a source is able to compete, while relevance and chunkability influence what happens once it gets there.

What comes next

With authority added to the picture, we now have the four parts of the GEO framework we’ve explored throughout this series:

1. Eligibility — Candidate selection
2. Precision — Semantic relevance
3. Citability — Chunkability
4. Safety — Authority

Together, these help explain the different stages and considerations that can influence whether content is considered, retrieved, and ultimately used in a generative response.

In the next article, we’ll bring these pieces together and look at what a GEO analytics stack should actually measure!

The post Authority in GEO: It’s A Filter, Not a Score appeared first on Lumar.

Your content can be highly relevant to a query and still never make it into an AI-generated response.

In the previous article, we touched on one reason why: a relevant passage may still be unsuitable for citation if it relies on the rest of the page to make sense.

This is where chunkability comes in. 

In this article, we’ll explore what chunkability means, why it matters for retrieval and citation, and how to improve it without rewriting entire pages.

Only got 30 seconds? Here are the key takeaways…

What chunkability actually means

Chunkability describes how well a section of content can function independently when it’s separated from the rest of the page. It isn’t determined by word count, readability scores or formatting alone.

Importantly, this doesn’t mean breaking long-form content into smaller pages. In our experience, it means ensuring individual passages contain enough information to be retrieved, understood independently, and used safely in an AI-generated response.

A highly chunkable passage is atomic, meaning it contains a complete idea or answer within a clearly defined section, without depending on information elsewhere on the page to make sense.

Why chunkability matters

Chunkability is most relevant when we look at individual passages and how suitable they are for retrieval and citation. At this level, the structure and context within a passage become particularly important.

In our experience, chunkability can influence chunk retrieval precision, citation selection and answer confidence.

This helps explain why valuable information within a long, relevant page may still go uncited, while shorter, more precise passages can be easier to retrieve and use in an AI-generated response.

The four pillars of chunkability

In our experience, there are four structural characteristics we look at when evaluating chunkability:

• Topical cohesion
• Boundary clarity
• Answer density
• Self-contained meaning

Together, these provide a practical way to assess chunkability, identify where a passage may struggle and pinpoint what can be improved.

1. Topical cohesion

Topical cohesion means keeping a chunk focused on one clear topic or answer. When a paragraph contains multiple ideas, an explanation starts to drift, or the context changes midway through a section, that focus can become diluted.

Keeping the topic cohesive can make the passage’s meaning more focused, which may help it align more closely with a relevant query during retrieval.

2. Boundary clarity

A focused passage also needs clear boundaries. It should start with a clear premise and end when that idea has been fully addressed, keeping related information together within the same chunk.

When those boundaries aren’t clear, concepts can become split across chunks, leaving explanations without the context they need or creating ambiguous references between passages.

3. Answer density

Answer density is about how much useful, specific information a passage contains. Details such as entities, units, constraints and clearly defined steps can make a chunk more useful when answering a specific query.

Extra information that doesn’t contribute to the answer can make those details harder to isolate. Keeping a passage concise and focused helps surface the information needed for retrieval and citation, without adding length for its own sake.

4. Self-contained meaning

Self-contained meaning is about whether a passage can be understood without relying on information elsewhere on the page. References to previous sections, pronouns without clear referents or phrases such as “as mentioned above” can make that harder.

The aim is to give each passage enough context to make sense on its own. If important information sits outside the chunk, the passage may become more difficult to interpret and, as a result, less suitable for citation.

Why chunkability matters at citation

A passage can be relevant, on-topic and semantically similar to the query, but still be unsuitable for citation.

When it comes to citation, passages that can be understood with minimal ambiguity or risk of misinterpretation may be more suitable as supporting evidence. Passages that are clearly bounded, specific, and self-contained depend less on missing context, reducing the risk of incorrect use or contribution to an inaccurate response.

This is where chunkability becomes particularly important. The same qualities that make a passage easier to understand independently can also make it more reliable to use in an AI-generated response.

What low chunkability looks like in practice

Low chunkability isn’t necessarily a sign of poor writing. Content can read well as a complete page but lose important context when individual passages are separated from it.

Common examples include:

• Long introductions that delay the main information
• Multiple FAQ questions or topics grouped into one section
• Steps or instructions split across several headings
• Tables that depend on explanations elsewhere on the page

In each case, the information may rely on context outside the passage itself. This makes it harder to retrieve, understand or use independently.

How to improve chunkability without rewriting everything

Improving chunkability doesn’t necessarily require rewriting an entire page. Often, it means looking at how information is structured within individual sections and making targeted changes.

This could include:

• Splitting multi-topic sections into more focused passages
• Moving key facts into the passage where they’re needed
• Structuring each section around one clear topic or question
• Replacing ambiguous pronouns with the entities they refer to

These changes help individual passages carry more of the information and context they need to work independently.

to work independently.

What comes next

So far, we’ve looked at how pages become eligible for retrieval, how semantic relevance works at different stages, and what makes individual passages easier to retrieve and cite.

The next article turns to authority and the role it can play in filtering which sources are considered for use in AI-generated responses. We’ll explore why authority can act as a filter, particularly for queries where trust and accuracy carry greater weight.

The post Chunkability: Why Some Content Gets Cited, and Most Doesn’t appeared first on Lumar.

In the previous article, we explored candidate page selection and how AI systems determine which pages are eligible for retrieval.

Semantic relevance can help inform that decision, but its role extends beyond candidate selection. It can serve different purposes throughout retrieval, which makes the advice to simply “improve semantic relevance” less straightforward than it sounds.

In this article, we’ll look at those different roles and why measuring semantic relevance in one place can give an incomplete picture of AI search visibility.

Only got 30 seconds? Here are the key takeaways…

Semantic relevance isn’t a single score

Semantic relevance is often treated as though it can be reduced to a single embedding, a single cosine similarity score or a single judgment of quality.

That simplification may be useful for illustrating the concept, but it doesn’t reflect how relevance appears to be applied throughout a production retrieval system. Where relevance applies matters just as much as how much semantic similarity exists.

Semantic relevance for candidate page selection

At the candidate page selection level, semantic relevance can help answer a relatively broad question:

“Is this page about the query?”

Here, the focus is less on identifying the single best answer and more on establishing whether a page is clearly about the topic being searched for. 

In information retrieval, this is sometimes described as “aboutness”. A page can become eligible for retrieval without containing the strongest answer, as long as it can demonstrate clear topical relevance.

That’s why pages may become weaker candidates when they:

• Cover multiple unrelated subjects.
• Use lengthy introductions that delay the main topic.
• Rely on vague titles and headings that obscure the page’s primary focus.

Semantic relevance for chunk retrieval

Once a page becomes eligible for retrieval, the focus shifts from the page as a whole to the individual passages within it. The question now becomes:

“Which passage best answers this question?”

Here, semantic relevance may become much more fine-grained, with AI systems comparing individual passages rather than the page as a whole. As a result, precision and specificity matter far more than they did during candidate page selection.

This is where details such as entities, units, exact phrasing and clearly defined steps start to matter. They help AI systems identify the passage that most directly answers the user’s query. 

Without that level of precision, a page may pass candidate selection but still fail to surface a relevant passage during chunk retrieval.

Semantic relevance for citation

When it comes to citation, relevance alone may not be enough. A useful passage also needs to provide suitable evidence for the generated response. This means considering:

“Is this chunk safe and precise enough to cite?”

A passage may be semantically similar to the query and contain the right answer, yet still not be used in the final response.

One reason is that the passage can’t stand on its own. Ambiguous wording, missing units, references that depend on surrounding paragraphs, or instructions spread across multiple sections can all make a chunk less suitable for citation.

For a passage to be cited, it needs to be complete enough to be understood in isolation, without relying on additional context elsewhere on the page.

Why a single similarity score can be misleading

Some GEO tools summarize semantic relevance using a single page-level similarity score. While that can be useful as a high-level indicator, it can blur the boundaries between different stages of retrieval.

A page may demonstrate strong topical relevance during candidate selection, yet fail to retrieve the most relevant passage, or retrieve a strong passage that isn’t suitable for citation. A single score makes it difficult to see where in that process content succeeds or falls short.

Reducing those stages to a single score makes it easy to optimize the wrong thing. This helps explain why some pages appear highly relevant yet rarely contribute to AI-generated responses.

A better way to think about semantic relevance

The three roles can be summarized as follows:

Practical implications for GEO

These differences have practical implications for how content is structured for GEO. 

Start by making the page’s primary topic clear. Within it, keep sections narrowly focused so individual passages can closely match specific queries.

Those passages also need to make sense independently, without relying on surrounding content to carry essential context.

Each of these serves a different purpose: 

• Topical clarity supports candidate selection.
• Focused answers support chunk retrieval.
• Self-contained passages support citation.

What comes next

Semantic relevance can play different roles depending on whether we’re looking at candidate pages, individual passages, or their suitability for citation. The next article looks more closely at how those passages are structured for use in AI-generated responses.

We’ll explore chunkability, what makes a passage suitable for citation, and how content structure can influence whether information is ultimately used in a generated response.

The post Semantic Relevance Isn’t One Metric — It Has Three Jobs appeared first on Lumar.

In the first article, we explored why rankings alone no longer explain visibility in AI-powered search. This time, we take a closer look at candidate page selection—a concept we use at Lumar to help explain why some pages become eligible for retrieval while others don’t.

We’ll explore why candidate page selection matters for GEO, how it differs from traditional rankings, and what influences whether a page becomes eligible for retrieval.

Short on time? Here are the key takeaways…

Candidate selection isn’t the same as ranking

Traditional SEO is built around rankings. Once a page is indexed, the goal is to improve its position in search results. Rankings fluctuate over time, but every eligible page is competing for visibility.

Candidate page selection is different. Instead of determining where a page appears, it determines whether that page is considered a potential source for a particular query.

From what we’ve seen, this stage has traditionally been largely invisible because standard SEO tools don’t measure it. Rankings, keyword positions and traffic show how a page performs in search results, but they can’t tell you whether it was ever considered for an AI-generated response.

That’s why strong rankings don’t always translate into AI visibility. A page may rank highly in traditional search yet never contribute to an AI-generated response if it isn’t selected as a candidate.

Why candidate selection exists

If candidate selection determines which pages are eligible for retrieval, the next question is why AI systems need this extra stage at all.

To generate a coherent response, AI systems need an efficient way to identify useful source material. Evaluating every indexed page or every possible passage for every query isn’t practical, so one way to think about candidate selection is as narrowing the search space and reducing the number of potential sources considered for a query.

That also explains why page-level signals matter so much. Before individual passages can contribute to a response, the page itself needs to establish that it’s relevant to the query and appropriate to retrieve from.

What influences candidate page selection?

Candidate page selection appears to be influenced by a combination of lexical, semantic, authority and freshness signals. Their relative importance changes depending on the query. A breaking news search, for example, is likely to place greater emphasis on freshness, while a medical query may depend more heavily on authority and trust.

Lexical signals

Despite advances in semantic search, traditional keyword signals still play an important role. Exact and near-exact matches in titles, H1s, headings and anchor text provide strong topical cues, helping search systems identify what a page is primarily about.

Semantic relevance

AI systems also appear to evaluate whether a page is genuinely about the topic being searched for, rather than simply containing matching keywords. Pages that explore a subject in depth may provide stronger topical signals than pages that only mention it in passing.

At this stage, the focus is on establishing broad topical relevance before identifying the strongest individual answer.

Learn more about semantic relevance for GEO/AEO

Authority and trust

Authority also appears to be important during candidate selection, particularly for YMYL (Your Money or Your Life) topics where inaccurate information could have real-world consequences, such as medical, financial or legal advice.

In our experience, official documentation, government guidance, recognized organizations, and established publishers are more likely to be considered for these types of queries. For broader informational searches, authority is balanced alongside topical relevance and content quality.

Freshness

Freshness doesn’t influence candidate page selection equally for every query. Evergreen content may remain relevant for years, while news, product updates and other rapidly changing topics often favor more recent sources.

Candidate selection depends on the query

Whether a page becomes a candidate depends on the query. A page may be considered a strong source for one search, but never enter the candidate pool for another.

For example, a general automotive blog might be a suitable source for a query such as “how oil changes work.” A much more specific search, such as “Honda Civic oil drain torque,” is more likely to favor manufacturer documentation or specialist technical resources. This is because where accuracy and precision carry greater weight.

Why traditional SEO tools don’t measure candidate selection

Traditional SEO tools are built around the assumption that indexed pages are eligible to compete for visibility. From there, they measure performance through rankings, keyword positions, backlinks and traffic.

Candidate selection introduces an earlier stage that sits outside those metrics. A page may be fully indexed and perform well in traditional search, yet never become a candidate for a particular AI-generated response.

As a result, existing SEO tools can’t tell you whether a page entered the candidate pool, why it may have been excluded, or which signals influenced that decision. Understanding AI visibility will require ways of measuring performance beyond rankings alone.

What this means for GEO

Candidate selection starts at the page level, so optimization should too. A page with a clear, well-defined topic is easier for AI systems to interpret than one trying to cover multiple unrelated subjects. Aligning titles, H1s and headings with the queries people actually use to search helps reinforce that topical focus.

Once the page’s intent is clear, credibility becomes just as important. For queries where trust carries greater weight, the quality and authority of the source may influence whether the page is considered for retrieval at all.

Only then does it make sense to optimize individual passages. If the page never becomes a candidate, even the strongest content is unlikely to contribute to an AI-generated response.

What comes next

Candidate selection gives us one way to think about which pages may be eligible for retrieval. From there, we can start to think about which parts of those pages are most relevant to a particular query.

In the next article, we’ll look at: 

• How AI systems evaluate semantic relevance once a page becomes a candidate
• Why the same content can be interpreted differently at different stages of retrieval, and 
• How chunk-level retrieval ultimately determines which passages contribute to an AI-generated response.

The post Candidate Page Selection: The Stage SEO Tools Don’t Measure appeared first on Lumar.

This is the first article in our Generative Engine Optimization (GEO) series, where we explore how AI-powered search is changing SEO—and how SEO teams can successfully adapt. 

We’ll first look at why rankings are no longer the best way to understand search visibility. Then, the rest of the series takes a closer look at the concepts shaping AI search, from semantic relevance to authority. 

For the past 20 years, search optimization has revolved around one question:

“Where do I rank?”

This made sense when search engines returned lists of blue links, and users chose which result to click. It makes far less sense when the interface generates a single answer instead.

Generative AI hasn’t killed search, but it has changed how information is selected and presented. Yet many SEO teams still rely heavily on traditional ranking metrics, even though rankings alone no longer explain visibility.

This article explores why the traditional SEO mental model no longer fits AI-powered search, and what that means for search visibility.

Only got 30 seconds? Here are our key takeaways:

Why rankings alone no longer guarantee visibility

Traditional search engines were designed to present users with a ranked list of results. In most cases, the interface looked like this: 

• Multiple results were returned.
• Results were ordered by relative relevance.
• Users decided which result to click.

SEO’s focus on rankings emerged because that’s how users navigated search. Generative search changes the interface—and with it, the way visibility is earned.

Rather than presenting users with a page of ranked links, AI can now:

• Collapse the Search Engine Results Page (SERP) into a single answer.
• Synthesize information from multiple sources. 
• Prioritize relevance and trustworthiness.

Content also needs to be recognized as a source that can inform the generated response.

Generative engines don’t just “rank” — they cite

One way to think about the process is that AI systems need to identify relevant sources to generate a response.

For GEO, that means creating content that AI systems can recognize as relevant, trustworthy, and useful for the query being answered.

In practice, that means producing content that is:

• Precise, answering the query clearly and directly.
• Semantically rich, with enough semantic recall to cover the topic comprehensively.
• Unique, offering information or perspectives that add value beyond competing content.
• Designed to answer real user questions, rather than simply targeting keywords.
• Fresh, particularly where information changes over time.

Content quality, however, is only part of the picture. Brand presence, authority signals and clear content structure also influence how AI systems interpret and use information.

Read Tech GEO: Technical Website Fixes to Optimize for AI Inclusion

The GEO retrieval pipeline (simplified)

In our experience, it can be helpful to think about AI-powered search as a series of stages that determine whether content is ultimately used in a generated response. 

A simplified version looks like this:

1. Availability and indexability
   Can the content be discovered, accessed, and processed?
2. Candidate page selection
   Which pages appear relevant to the topic and credible enough to be considered as potential source material?
3. Chunk retrieval
   Which sections of those pages best address the user’s question?
4. Answer assembly and attribution
   Which passages provide clear, reliable information that can support the generated response?

In our experience, candidate page selection can be an important point of failure. It isn’t a ranking; it’s closer to a yes/no. A page that never enters the candidate pool can contain the perfect answer and still be invisible, because no passage inside it is ever retrieved. That’s the stage traditional analytics can’t show you.

Why page context still matters

AI-generated responses often reference specific passages rather than entire pages. Those passages, however, don’t exist in isolation.

Google has recently clarified that “chunking” content into smaller pages isn’t a GEO strategy. Its systems can understand long-form content and identify the most relevant sections without requiring pages to be artificially split up.

The broader context of the page helps AI systems interpret individual passages. A well-written paragraph is far more useful when it’s supported by comprehensive, topically relevant content.

This is why GEO extends beyond optimizing individual paragraphs. The surrounding context helps establish the topical focus and credibility that AI systems use to interpret individual passages.

Different queries prioritize different signals

The type of question influences which sources are most useful for an AI-generated response. Queries where accuracy is critical aren’t evaluated in quite the same way as those that are open-ended or exploratory.

Let’s look at two examples:

Example 1: “Honda Civic oil change”

Someone looking for instructions on maintaining their car expects accurate, reliable guidance. 

For safety-critical, YMYL-adjacent queries like this, AI-generated responses are generally more likely to reference sources that demonstrate strong expertise and trust, such as vehicle manufacturers, recognized repair manuals and established automotive publications.

This doesn’t mean smaller publishers can’t be cited, but demonstrating credibility becomes especially important when accuracy has real-world consequences.

Example 2: “Best productivity hacks for students”

This is a non-YMYL, exploratory query with no single definitive answer. A wider range of sources may be useful, as long as they address the topic clearly and comprehensively.

In this case, topical relevance and content quality may carry more weight than institutional authority alone.

Put simply, the relative importance of authority, topical relevance, and content quality depends on the information the user is looking for.

Semantic relevance is doing different jobs

Semantic relevance is often discussed as if it were a single signal, but it contributes to several aspects of how content is understood.

At the page level, it establishes the content’s overall topic and scope. Within individual passages, it connects specific questions to the information that answers them.

Strong semantic recall comes from covering a topic in sufficient depth and making relationships between concepts clear. Pages with broader topical coverage provide more context for both readers and AI systems.

What’s next

This is the first article in our GEO series. The rest of this series explores that framework in more detail, including:

• Candidate selection and how AI systems identify potential source material
• Semantic relevance across different stages of retrieval
• Chunkability and its role in AI citations
• Authority and trust signals
• Measuring GEO performance beyond traditional SEO metrics

The post From SEO to GEO: Why Rankings Alone Won’t Get You Visibility in AI Search appeared first on Lumar.

Each month, Lumar’s in-house tech SEO experts hand-pick some of the SEO industry’s top news stories from across the web to keep you up-to-date on all things SEO and website optimization.

For our July 2026 SEO & GEO/AEO news roundup, the top headlines include:  

Dive into the details behind the SEO and GEO/AEO headlines this month below…

Google’s 2026 spam update completes its rollout in just two days

It took Google just two days to finish rolling out its second spam update of  the year. . Unlike  March’s spam update, this one didn’t include  any noteworthy changes to its spam policies, nor were there any announcements about specific behaviour being penalised. For sites affected, Google advises reviewing its spam policy documentation in detail. Site owners should also be prepared that recovery from spam-related penalties typically takes months rather than days.

(Source: Google Search Status Dashboard)

Data shows continued decline in CTR for 2026

New data from Advanced Web Ranking indicates that CTR for top organic positions continues to decline year over year. , particularly on SERPs  featuring AI Overviews and other rich features. The data highlights a widening  gap between ‘clean’ SERPs with 10 organic results and feature-heavy results pages, where AI summaries, ads and other rich results compete for attention. Rankings  still matter, but visibility is increasingly  decoupled from clicks as zero-click behaviour rises.

(Source: Advanced Web Ranking)

Free eBook

Read our guide to GEO/AEO

Want a more in-depth guide to AI search optimization strategies? Our 80-page GEO / AEO playbook has you covered.

Google clarifies how impressions work in AI Overviews and AI Mode

After some initial confusion among SEOs, John Mueller has confirmed that only links — not brand mentions — count as impressions in Google’s AI search features. Links must also be visible to the user to be counted. If a user needs to expand or interact with an AI feature before a link is displayed, the impression is only recorded once that link becomes visible..

(Source: John Mueller via BlueSky)

Google doubles-down on its ‘Good SEO is Good GEO’ messaging

A new Google Think article was published last month by Brendon Kraham, VP of Search & Commerce Global Ads Solutions at Google. Kraham explains how AI-led features like AI Overviews and AI Mode are designed to surface helpful, high-quality content from the existing search index. He reiterates that success in this environment still depends on strong technical SEO, useful content and a focus on user needs, rather than AI-specific optimisation tactics.

The article also positions AI Search as an extension of traditional search systems, with existing ranking and quality signals still central to visibility in generative search.

(Source: Think With Google)

Study finds just 3% of llms.txt files were accessed by AI systems in May

A new study from Ahrefs analysed 137,000 domains that were live in May 2026. 28% of those had published an llms.txt file yet 97% of those files received zero requests..

Of the small subset that did get traffic, AI-related crawlers accounted for a minority share. Most traffic instead came from SEO tools, research bots and general crawlers. The study further raises questions about the practical value of llms.txt files for AI visibility,  particularly if maintaining one comes as the expense of  higher-priority tasks.

(Source: Ahrefs)

Microsoft Clarity adds robots.txt violation tracking to Bot Analytics

Microsoft has added a new feature in Clarity’s Bot Analytics dashboard that  flags when bots request URLs blocked by a site’s robots.txt file. The update allows site owners to monitor violation rates, track trends over time and break down activity by bot, operator, and content type.

The feature also highlights which sections of a site are being accessed despite restrictions, offering a clearer view of how AI crawlers and automated systems interact with published content and crawl directives.

(Source: Microsoft)

Google downplays the benefits of markdown for AI understanding… and then surfaces markdown in AI overviews

In the latest Search Off the Record episode, Martin Splitt and John Mueller discuss whether websites should convert content to Markdown to improve AI visibility. These tactics have been gaining a lot of traction recently, especially via the launch of features like Cloudflare’s “Markdown for AI Agents”. Google’s argument  is that modern crawlers and LLMs are already well-equipped to process standard HTML, including complex page structures. Therefore, they  don’t require simplified markdown formats.

 That advice, however,  appeared to be contradicted when one user spotted actual Markdown appearing in Google’s AI Overview snippets. At the very least, this   confirms that Google’s own AI systems are accessing Markdown and seemingly using it as part of the response generation process.

(Source: Google Search Off the Record Podcast)

Bing adds more new AI visibility reporting features to Webmaster Tools

Microsoft has introduced four new metrics in Bing Webmaster Tools: Intents, Topics, Citation Share, and Compare. The update builds on its existing AI Performance report, giving site owners more  detailed insight into how content is surfaced and cited in AI-generated answers.

Intents and Topics group queries into broader semantic categories, while Citation Share measures how often a site is cited relative to competitors. The changes  help formalise AI search optimisation as a measurable layer of SEO.

(Source: Bing)

Headline formats have minimal impact on Google Discover performance

Analysis of over 3.4 million articles indicates that common assumptions around headline structure—whether they’re formatted as questions, statements or quotes—do not reliably drive Discover visibility alone. Instead, performance metrics are more strongly tied to broader factors like content quality and topical relevance. The findings challenge the growing popularity of  ‘headline hacks’ as an optimisation strategy.

(Source: Search Engine Land)

Did you know you can now optimize for AI search and LLM mentions with Lumar?

Lumar’s GEO tools for AI search optimization provide data-driven insights to help you optimize your website for AI-driven search visibility, earn more AI citations, and improve brand visibility across LLMs.

Explore Lumar GEO Platform Features

Get started with Lumar

Get Lumar pricing

The post AI Search & SEO Industry News – July 2026 appeared first on Lumar.