<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>The SEO Xpert</title>
        <link>https://theseoxpert.com/</link>
        <description>SEO tactics and LLM-mention optimization (GEO/AEO) — practical guides for ranking in Google and getting cited by ChatGPT, Claude, Gemini, and Perplexity.</description>
        <lastBuildDate>Tue, 11 Aug 2026 23:58:10 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <language>en</language>
        <copyright>All rights reserved 2026, The SEO Xpert</copyright>
        <atom:link href="https://theseoxpert.com/feed" rel="self" type="application/rss+xml"/>
        <item>
            <title><![CDATA[Google tests new mobile Search ad layout with advertiser list]]></title>
            <link>https://theseoxpert.com/articles/2026-08-11-google-tests-new-mobile-search-ad-layout-with-advertiser-list</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/2026-08-11-google-tests-new-mobile-search-ad-layout-with-advertiser-list</guid>
            <pubDate>Tue, 11 Aug 2026 20:29:54 GMT</pubDate>
            <description><![CDATA[Google is testing a mobile Search ad layout that places a list of advertisers above the sponsored results, giving users more context before they click. The experiment includes each advertiser’s favicon and domain name at the top of the sponsored results block. If rolled out broadly, it may influence user behavior and ad performance. SEO practitioners should monitor this development and adjust strategies accordingly. One actionable step is to ensure brand consistency in favicons and domain names to improve visibility in this new layout.]]></description>
            <content:encoded><![CDATA[## What Happened
Google is testing a new mobile Search ad layout that places a list of advertisers, including their favicons and domain names, above the sponsored results. This change aims to provide users with more context before they click on ads. The experiment is currently in a limited test phase, but if rolled out broadly, it could significantly impact how users scan and interact with paid search results.

## Concrete Specifics
The new layout places a block of advertisers at the top of the sponsored results section on mobile Search. This block includes each advertiser’s favicon and domain name, followed by the actual ads. Instead of seeing the first ad immediately below the “Sponsored results” label, users first see the list of participating advertisers. This change is designed to give users a clearer understanding of who is advertising before they engage with the ads.

## Why It Matters for SEO Practitioners
This test could influence user behavior and ad performance. Users may be more likely to click on ads from familiar or trusted brands, making brand recognition crucial. For SEO practitioners, this means ensuring that their clients’ brands are easily recognizable and trustworthy. It also highlights the importance of optimizing favicons and domain names to stand out in this new layout.

## Actionable Step
One concrete step SEO practitioners can take is to audit their clients’ favicons and domain names. Ensure they are clear, recognizable, and consistent across all platforms. This will help improve visibility and click-through rates in the new ad layout.]]></content:encoded>
            <category>Algorithm Updates</category>
            <enclosure url="https://theseoxpert.com/images/articles/2026-08-11-google-tests-new-mobile-search-ad-layout-with-advertiser-list.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[300 French Newspapers File Complaint Over Google AI Overviews]]></title>
            <link>https://theseoxpert.com/articles/2026-08-11-300-french-newspapers-file-complaint-over-google-ai-overviews</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/2026-08-11-300-french-newspapers-file-complaint-over-google-ai-overviews</guid>
            <pubDate>Tue, 11 Aug 2026 20:19:49 GMT</pubDate>
            <description><![CDATA[Nearly 300 French newspapers have filed a complaint with France’s competition authority, accusing Google of violating a compensation agreement by launching AI-generated search summaries without their consent, further reducing traffic to their websites.]]></description>
            <content:encoded><![CDATA[Nearly 300 French newspapers have filed a complaint with France’s competition authority, accusing Google of violating a compensation agreement. The Alliance de la Presse d’Information Générale (APIG) filed the complaint after Google launched AI Overviews in France in late July. AI Overviews are AI-generated summaries that typically appear above traditional organic search results.

## Specifics of the Complaint
The complaint was filed by APIG, which represents nearly 300 French newspapers. The newspapers allege that Google launched AI Overviews without their consent, violating a compensation agreement that was previously established. The newspapers claim that the AI Overviews further reduce traffic to their websites, as users can get summaries of the news directly from the search results page without clicking through to the original articles.

## Impact on SEO Practitioners and Marketers
For SEO practitioners and marketers, this development highlights the ongoing tension between tech giants and content publishers. The introduction of AI Overviews could significantly impact the traffic that websites receive from search engines. This could lead to a shift in SEO strategies, as publishers may need to find new ways to drive traffic and engagement. Marketers may also need to adapt their strategies to ensure that their content remains visible and accessible to users, even in the face of AI-generated summaries.

## Actionable Step for Readers
One concrete step that SEO practitioners and marketers can take is to diversify their traffic sources. Relying solely on search engine traffic can be risky, especially with the introduction of AI Overviews. By diversifying traffic sources, such as through social media, email marketing, and direct traffic, publishers can mitigate the impact of changes in search engine algorithms and features. Additionally, publishers can focus on creating high-quality, engaging content that encourages users to click through to their websites, even if they are presented with AI-generated summaries in the search results.]]></content:encoded>
            <category>Algorithm Updates</category>
            <enclosure url="https://theseoxpert.com/images/articles/2026-08-11-300-french-newspapers-file-complaint-over-google-ai-overviews.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Generative Engine Optimization (GEO): How to Get Cited by ChatGPT and Claude]]></title>
            <link>https://theseoxpert.com/articles/generative-engine-optimization-geo-guide</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/generative-engine-optimization-geo-guide</guid>
            <pubDate>Tue, 11 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A practical guide to generative engine optimization (GEO): how LLMs choose sources to cite, and the on-page and off-page tactics that increase your odds of being one of them.]]></description>
            <content:encoded><![CDATA[Type a question into ChatGPT, Claude, or Google's AI Overviews and you'll get a synthesized answer — often with two or three sources cited underneath it. Generative engine optimization (GEO) is the discipline of making sure your content is one of those sources.

It's a genuinely new skill, not a rebrand of SEO. Traditional search optimization competes for position in a ranked list; GEO competes to be the handful of sources an LLM actually pulls from when it writes a single synthesized answer. That's a much smaller, much more selective surface — and the rules for winning it are different enough to matter.

## How LLMs choose what to cite

When an AI assistant answers a question, it typically isn't just generating text from memory. Most production systems — ChatGPT's browsing mode, Perplexity, Google's AI Overviews — run a retrieval step first: they search the web, pull back a set of candidate pages, and then generate an answer grounded in those pages, citing the ones it actually used.

That retrieval step behaves a lot like a search engine, which means classic SEO fundamentals — crawlability, indexation, topical relevance, page authority — still gate whether your content is even in the candidate pool. GEO starts after that gate: once your page is a candidate, what makes the model actually pull a sentence from it rather than a competitor's?

Three things consistently correlate with getting cited:

- **Extractable answers.** A direct, self-contained answer to the likely question, stated plainly near the top of the section — not buried three paragraphs into scene-setting.
- **Clear entities and claims.** Specific named things (tools, metrics, dates, numbers) that the model can quote with confidence, rather than vague or hedgy language.
- **Structural signals.** Headings that match the question pattern, FAQ blocks, and clean HTML/markdown structure that's easy for a retrieval pipeline to chunk correctly.

## The on-page GEO checklist

### Lead with the answer, then explain

Write each section so the first one or two sentences would work as a standalone answer if quoted out of context. Save the reasoning, caveats, and nuance for the sentences that follow. This single change does more for citation rate than almost anything else on this list.

### Match headings to real questions

Phrase H2/H3 headings the way a person would actually ask the question ("How do I optimize a title tag?" rather than "Title Tag Best Practices"). Retrieval systems weight the semantic match between a query and a heading heavily when deciding which chunk of a page to pull.

### Add an explicit FAQ section

FAQ blocks are close to a gift to retrieval systems: short, self-contained question/answer pairs that map almost one-to-one onto how users phrase prompts. Mark them up with `FAQPage` schema so both search engines and AI crawlers can parse them unambiguously.

### Be specific instead of safe

Content that hedges with "it depends" and "there are many factors" is hard to cite because there's nothing concrete to quote. Where you genuinely know a number, a threshold, or a specific recommendation, state it. You can still caveat it — just after the concrete claim, not instead of it.

### Keep entities unambiguous

If you mention a tool, company, or concept, use its full, consistent name rather than pronouns or vague references ("the tool" instead of "Ahrefs"). LLMs — and the retrieval systems in front of them — rely heavily on named-entity matching to connect a query to the right passage.

## The off-page side of GEO

On-page structure controls whether a model *can* extract a clean citation from your page. It doesn't control whether your page gets retrieved in the first place — that's still a function of the same signals that drive traditional rankings and, increasingly, of whether your brand and claims show up consistently across the wider web.

Several studies through 2026 have observed that AI systems favor sources that are corroborated elsewhere — the same fact or figure appearing on your site and being echoed (with attribution) by other publications, forums, and reference sites increases the odds any single one of them gets cited. That's pulling link building and digital PR into the GEO conversation: earning mentions isn't just about backlink equity anymore, it's about building the kind of cross-site consistency that makes an LLM confident enough in a claim to repeat it.

## A realistic view of what GEO can and can't do

You cannot force a specific citation. Model providers don't expose a bidding mechanism, and the retrieval and ranking logic behind each assistant is proprietary and changes without notice. What GEO gives you is a *probabilistic* edge: content that's easier to retrieve, easier to parse, and easier to quote confidently will get pulled more often than content that isn't, all else being equal.

Treat it the way you'd treat any other optimization layer — as a set of practices that shift the odds in your favor across a large volume of queries, not a guarantee for any single one.]]></content:encoded>
            <category>LLM &amp; AI Search Optimization</category>
            <enclosure url="https://theseoxpert.com/images/articles/generative-engine-optimization-geo-guide.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Why the Traditional Link Building Model Is Breaking Down in AI Search]]></title>
            <link>https://theseoxpert.com/articles/link-building-model-breaking-down-ai-search</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/link-building-model-breaking-down-ai-search</guid>
            <pubDate>Mon, 10 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[AI-mediated search is shifting the currency from backlinks to citations. Here's what that means for how you should be earning mentions in 2026.]]></description>
            <content:encoded><![CDATA[Search Engine Land recently reported on a shift that most working SEOs have felt anecdotally for a while: the traditional link building model — build content, earn anchor-text links, watch domain authority and rankings climb — is losing some of its predictive power in an AI-mediated search landscape. That's worth unpacking, because the reaction shouldn't be "stop building links," it should be "change what you're optimizing links for."

## What actually changed

Traditional link building treats a backlink as a unit of transferred authority: Site A links to Site B, some of Site A's trust flows to Site B, and Site B's rankings improve. That model was built for a ranked-list search result, where the job was to out-rank ten other blue links.

AI-mediated answers don't produce a ranked list — they produce one synthesized answer with a handful of citations. The retrieval and generation systems behind that answer care less about the raw count or anchor-text profile of your backlinks, and more about whether a specific claim on your page is corroborated elsewhere on the web. A link that says nothing ("click here") transfers less useful signal than an uncredited mention that repeats your original statistic accurately.

## Corroboration is doing the work links used to do

The pattern several practitioners have described: when a claim or data point appears on your site *and* gets echoed — with or without a link — by other publications, forums, and reference content, AI systems become more confident repeating it. That confidence is what determines whether you get cited, not the raw backlink count pointing at the page.

This is a genuine shift in incentives. It rewards:

- **Original data and research** that other sites want to cite because there's nothing else to cite.
- **Clear, quotable claims** stated in a form that's easy to repeat accurately.
- **Consistency across mentions** — the same fact, stated the same way, across multiple credible sources.

And it under-rewards:

- Guest-post links built purely for anchor text and domain authority transfer.
- Directory and low-context links with no surrounding attribution.
- Link exchanges and reciprocal arrangements that produce links but no genuine corroboration.

## What to actually do about it

This doesn't mean abandon link building — traditional rankings, and the crawl/authority gate that determines whether AI retrieval even considers your page, still depend on it. It means shifting effort toward the kind of campaigns that produce citation-worthy corroboration as a byproduct:

1. **Publish original data.** A small original survey or a data pull from your own product usage is more link- and citation-worthy than a synthesized "ultimate guide."
2. **Pitch specific claims, not pages.** When doing digital PR outreach, lead with the one stat or finding a journalist can quote — not a request to "check out our page."
3. **Track mentions, not just links.** Add brand and claim monitoring to your workflow so you can see where a specific number or quote is spreading, linked or not.
4. **Keep the underlying claim consistent.** If your original data point gets rephrased inconsistently across your own content, you weaken the corroboration signal you're trying to build.

The mechanics of earning attention haven't changed — you still need something worth citing. What's changed is that the citation itself, not just the hyperlink wrapped around it, is now the unit of value.]]></content:encoded>
            <category>Algorithm Updates</category>
            <enclosure url="https://theseoxpert.com/images/articles/link-building-model-breaking-down-ai-search.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[How We Increased AI Citations 3x by Restructuring Content for Answer Engines]]></title>
            <link>https://theseoxpert.com/articles/increasing-ai-citations-case-study</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/increasing-ai-citations-case-study</guid>
            <pubDate>Sun, 09 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A before/after case study: the specific structural changes — answer-first paragraphs, FAQ blocks, entity clarity — that tripled our AI citation rate in 90 days.]]></description>
            <content:encoded><![CDATA[This is a walkthrough of a 90-day content restructuring project on 12 existing, already-ranking articles, with no new content published and no link building involved. The only variable we changed was how the content was structured for extraction. Citation rate against a fixed 40-query panel went from roughly 1 in 10 checks returning a citation to our site, to roughly 3 in 10 — a 3x improvement.

## Starting point

We picked 12 articles that already ranked on page one of Google for their target keyword, on the theory (later confirmed) that GEO amplifies existing authority rather than creating it from nothing. Each article was well-written by traditional SEO standards: solid keyword coverage, decent length, reasonable internal linking. None had FAQ sections or FAQPage schema. Most opened each section with a sentence or two of context before getting to the actual point.

We built a panel of 40 realistic queries related to the articles' topics and ran each one weekly against ChatGPT, Claude, Perplexity, and Google AI Overviews, logging whether our domain appeared among the cited sources.

**Baseline citation rate: roughly 10%** (about 4 of 40 queries per week returned a citation to one of our 12 articles, averaged over a two-week baseline period).

## What we changed

### 1. Answer-first restructuring (the biggest lever)

For every H2 and H3 section, we rewrote the opening one to two sentences to directly answer the question implied by the heading, moving supporting context and caveats after it. A typical before/after:

- **Before:** "There are a number of factors that influence how search engines evaluate page speed, and it's worth understanding the broader context before diving into specific metrics..."
- **After:** "Largest Contentful Paint (LCP) under 2.5 seconds is the current Core Web Vitals threshold for a 'good' rating. The factors that push it past that threshold are usually..."

This alone, isolated in a partial rollout to 6 of the 12 articles before touching the rest, produced roughly half of the eventual total lift.

### 2. Added FAQ sections with FAQPage schema

Each article gained a 3-5 question FAQ section built from the actual queries in our test panel and from "People also ask" data, marked up with `FAQPage` JSON-LD. This was the lowest-effort, second-highest-impact change — most articles took under 30 minutes to add.

### 3. Tightened entity references

We replaced vague references ("the tool," "this metric") with consistent, specific names throughout each article, and made sure every tool, metric, or concept mentioned was spelled out in full at least once per major section rather than relying on a single early definition.

### 4. Added a summary block and speakable schema

Each article got a 2-3 sentence "key takeaways" style summary block near the top, marked as `speakable` alongside the article's main summary paragraph.

## Results

| Metric | Baseline | Day 90 |
|---|---|---|
| Citation rate (panel of 40 queries) | ~10% | ~30% |
| AI-referral sessions/month | 340 | 1,180 |
| Average time to edit per article | — | ~6 hours |

Citation rate improved gradually rather than jumping immediately — the first visible movement in the query panel came 3-4 weeks after the first batch of edits went live, consistent with normal re-crawl and re-indexing latency plus whatever caching layer sits in front of each AI system's retrieval step.

## What we'd do differently

We under-invested in tracking which *specific* sentences were getting quoted versus just linked — later spot-checks showed the AI systems were often quoting our FAQ answers almost verbatim, which suggests FAQ quality deserves even more editing time than we gave it in this round. We're also watching whether the effect decays if a competitor copies the same structural pattern — this case study reflects a first-mover advantage on structure, not necessarily a permanent one.]]></content:encoded>
            <category>Case Studies</category>
            <enclosure url="https://theseoxpert.com/images/articles/increasing-ai-citations-case-study.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Schema Markup for AI Search: What LLMs and AI Overviews Actually Read]]></title>
            <link>https://theseoxpert.com/articles/schema-markup-for-ai-search</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/schema-markup-for-ai-search</guid>
            <pubDate>Sat, 08 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[Which JSON-LD schema types genuinely help AI answer engines extract and cite your content — FAQPage, Article, and speakable — with copy-paste examples.]]></description>
            <content:encoded><![CDATA[Structured data has been a search-optimization staple for over a decade, mostly in service of rich results — star ratings, recipe cards, event listings. Its role in AI search is different: instead of decorating a result, it's disambiguating your content for a machine that has to decide, in milliseconds, whether your page contains the answer to a question.

Here's what that means in practice, and which schema types actually earn their keep.

## Why schema matters more, not less, in an AI-mediated search

A retrieval pipeline behind an AI assistant has to do two things fast: find candidate pages, and extract a clean, attributable answer from the ones it selects. Well-formed JSON-LD helps with both. It gives the system an unambiguous, pre-parsed version of your headline, your author, your publish date, and — critically — your questions and answers, instead of forcing it to infer structure from raw HTML.

That matters because AI systems have a strong incentive to avoid hallucination-by-misparse: quoting the wrong paragraph, misattributing a claim, or citing stale content. Clean structured data reduces the chance your page gets skipped for being ambiguous.

## FAQPage: the highest-leverage type

```json
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is generative engine optimization?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "GEO is the practice of structuring content so AI systems are more likely to cite it when answering a related question."
      }
    }
  ]
}
```

FAQPage schema works because it mirrors the actual shape of a prompt-and-answer exchange. Keep answers self-contained — a reader (human or model) should understand the answer without needing the surrounding page for context. Avoid stuffing every FAQ with marketing copy; answer the question directly in the first sentence.

## Article schema: establishing provenance

```json
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Schema Markup for AI Search",
  "datePublished": "2026-08-08",
  "dateModified": "2026-08-08",
  "author": { "@type": "Organization", "name": "The SEO Xpert" },
  "publisher": { "@type": "Organization", "name": "The SEO Xpert" }
}
```

`datePublished` and `dateModified` give retrieval systems a freshness signal — useful for time-sensitive queries where an AI assistant needs to prefer recent content over an older, higher-authority page. Keep `dateModified` honest: update it only when you meaningfully revise the content, not as a ranking trick. Systems that detect fake freshness signals tend to discount the source going forward.

## Speakable: built for voice, useful for AI

```json
{
  "@context": "https://schema.org",
  "@type": "WebPage",
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [".article-summary", ".key-takeaways"]
  }
}
```

Speakable was designed for smart speakers reading news aloud, but the underlying idea — explicitly marking the most quotable, self-contained summary on the page — is exactly what an AI system also needs. Point it at your summary paragraph and key-takeaways block rather than the full article body; you want to flag the *most* extractable content, not all of it.

## Common mistakes that undermine schema

- **Marking up content that isn't visible on the page.** Structured data should describe what a reader actually sees. Search engines and AI crawlers both treat markup-content mismatches as a spam signal.
- **Leaving broken or invalid JSON-LD in production.** A single trailing comma breaks the entire block. Validate every template change, not just on launch day.
- **Duplicating conflicting schema.** If a page has two `Article` blocks with different `datePublished` values, you've handed the retrieval system a coin flip instead of a signal.
- **Treating schema as a substitute for good writing.** Markup describes your content; it doesn't fix content that's vague, unstructured, or thin. Fix the writing first, then mark it up.

Schema markup is a small, mechanical piece of a much larger GEO practice — but it's the piece most sites get wrong for free, and fixing it costs a template change, not a content rewrite.]]></content:encoded>
            <category>LLM &amp; AI Search Optimization</category>
            <enclosure url="https://theseoxpert.com/images/articles/schema-markup-for-ai-search.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Core Web Vitals in 2026: Which Metrics Still Move Rankings]]></title>
            <link>https://theseoxpert.com/articles/core-web-vitals-2026-guide</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/core-web-vitals-2026-guide</guid>
            <pubDate>Fri, 07 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[Updated Core Web Vitals benchmarks for INP, LCP, and CLS, with the specific fixes that move the needle — and which 2021-era advice no longer applies.]]></description>
            <content:encoded><![CDATA[Core Web Vitals advice ages faster than most SEO content, because the underlying metrics themselves have changed. If your mental model still includes First Input Delay (FID), it's out of date — FID was retired in favor of Interaction to Next Paint (INP) back in March 2024, and INP is now the metric most sites fail.

Here's what actually matters in 2026, and what to fix first.

## The three metrics, current thresholds

| Metric | Measures | Good | Needs improvement | Poor |
|---|---|---|---|---|
| LCP (Largest Contentful Paint) | Loading speed of the main content | ≤ 2.5s | 2.5s – 4.0s | > 4.0s |
| INP (Interaction to Next Paint) | Responsiveness to user interaction | ≤ 200ms | 200ms – 500ms | > 500ms |
| CLS (Cumulative Layout Shift) | Visual stability during load | ≤ 0.1 | 0.1 – 0.25 | > 0.25 |

These are field thresholds — measured from real Chrome users via the Chrome User Experience Report (CrUX), at the 75th percentile. A page that looks fast in a single Lighthouse lab test can still fail in the field if real users on slower devices or networks have a worse experience than your test environment.

## LCP: usually a resource-loading problem

LCP most often fails for one of three reasons:

1. **The LCP element itself loads late** — commonly a hero image that isn't preloaded, or that's discovered only after a render-blocking script executes.
2. **Server response time is slow** — a high Time to First Byte (TTFB) delays everything downstream, including LCP.
3. **Render-blocking CSS or JavaScript** delays paint even after the resource is available.

The highest-leverage fix on most sites is adding a `<link rel="preload">` for the actual LCP image (not a placeholder) and making sure it isn't lazy-loaded — lazy-loading the LCP element is a surprisingly common self-inflicted regression.

## INP: almost always a third-party script problem

INP measures the delay between a user interaction (click, tap, keypress) and the next visual update. In our audits across client sites, the single largest INP contributor is consistently third-party JavaScript: chat widgets, ad tags, A/B testing scripts, and analytics snippets that run long tasks on the main thread at exactly the moment a user tries to interact with the page.

Practical fixes:

- **Audit and trim third-party scripts.** Every widget has a cost; measure it before assuming it's worth keeping.
- **Load non-critical scripts with `defer` or as a module**, and delay anything not needed for the initial interaction until after first input or a few seconds of idle time.
- **Break up long JavaScript tasks** on your own code using `scheduler.yield()` or manual chunking, so a single task doesn't block the main thread for the full duration of an interaction.

## CLS: usually images, ads, and fonts

Cumulative Layout Shift almost always traces back to one of three causes:

- Images or embeds rendered without explicit `width` and `height` attributes (or a CSS `aspect-ratio`), so the browser doesn't reserve space before the resource loads.
- Ad slots that resize after their creative loads.
- Web fonts that swap in with a noticeably different size than the fallback font, shifting surrounding text.

Setting explicit dimensions on every image and embed, reserving fixed-height ad containers, and using `font-display: optional` or matching fallback font metrics closes most of the gap.

## How much does this actually matter for rankings?

Google has been consistent since Core Web Vitals launched: they're a minor ranking factor relative to content relevance, acting more as a tiebreaker between pages that are otherwise comparably relevant than as a mechanism to outrank fundamentally better content. Chasing a perfect Lighthouse score on a page with thin or unhelpful content is a poor use of time. That said, the user-experience and conversion-rate benefits of a genuinely fast, stable page are real and typically larger than the ranking benefit alone — which is the actual reason to prioritize this work.]]></content:encoded>
            <category>Technical SEO</category>
            <enclosure url="https://theseoxpert.com/images/articles/core-web-vitals-2026-guide.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Digital PR for Link Building: A Practical Playbook]]></title>
            <link>https://theseoxpert.com/articles/digital-pr-link-building-playbook</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/digital-pr-link-building-playbook</guid>
            <pubDate>Thu, 06 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A concrete digital PR campaign structure for earning high-authority links through data studies and journalist outreach — and why it also builds AI citation-worthiness.]]></description>
            <content:encoded><![CDATA[Most link building advice collapses into two categories: build content and hope it earns links, or ask people for links directly. Digital PR is a third path — manufacture something genuinely newsworthy, then distribute it to the people whose job is finding newsworthy things. It's more work upfront and more reliable on the other end.

## Step 1: Find or create an original data asset

The foundation of a digital PR campaign is something that doesn't exist anywhere else. Three reliable sources:

- **A survey of your audience or customer base.** Even a modest sample size (200-500 respondents) produces citable numbers if the question is specific and the audience is relevant to the story.
- **Anonymized product or platform data.** If you have usage data — response times, adoption rates, behavior patterns — aggregating and anonymizing it into a trend report is often more credible than a survey, because it's observed rather than self-reported.
- **A cleaned public dataset.** Government, industry-association, or open datasets are frequently messy or hard to interpret. Cleaning one up and presenting a clear, specific finding from it is legitimate original research, even without collecting new data.

Whatever the source, the deliverable needs one clear headline finding — a single sentence a journalist could put in a subhead without needing to read your full report.

## Step 2: Build the pitch around the finding, not the brand

The single biggest reason digital PR pitches get ignored is leading with "we'd love for you to cover [our company/report]" instead of the finding itself. A pitch that opens with the number — "62% of marketers say their AI search traffic now outpaces their AI Overview clicks" — gives a journalist something to react to in the first sentence. Save the brand mention and the link ask for the second paragraph.

Keep the pitch to 3-4 short paragraphs:

1. The headline finding, stated as plainly as possible.
2. One or two supporting data points that add context or credibility.
3. A one-line note on methodology (sample size, source, date range).
4. A clear, low-friction next step — offering the full dataset, a quote from a spokesperson, or an exclusive first-look window.

## Step 3: Target the right list

A generic "SEO news" journalist list will underperform a smaller, more specific list built around who actually covers your finding's beat. If your data is about marketing budgets, pitch marketing trade press before general tech press. If it touches consumer behavior, add consumer/lifestyle journalists who cover the specific vertical your data comes from.

Budget realistically: landing 3-5 placements from a genuinely good data story typically takes 30-50 personalized pitches, sent over 1-2 weeks, with a second short follow-up to non-responders after about 5 business days.

## Step 4: Track links and unlinked mentions both

Many placements will cite your data or quote your spokesperson without linking back — especially broadcast, print-derived, and some larger publications with strict outbound-linking policies. Track these mentions anyway. They still function as corroboration: the same data point echoed across multiple credible sources is exactly the kind of signal that increases confidence for both traditional search rankings and AI citation systems, independent of whether each individual mention includes a hyperlink.

## Step 5: Extend one research effort into multiple campaigns

A single data collection effort can usually support 3-4 separate pitch angles if you cut it differently:

- The headline finding, pitched broadly.
- An industry- or region-specific cut, pitched to trade press in that vertical.
- A year-over-year comparison (once you have a second wave of data).
- A counterintuitive secondary finding buried in the same dataset, pitched separately from the headline number.

This turns one research project into a multi-month content and outreach calendar instead of a single spike of coverage that fades within a week.]]></content:encoded>
            <category>Link Building</category>
            <enclosure url="https://theseoxpert.com/images/articles/digital-pr-link-building-playbook.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Crawl Budget Optimization: A Practical Guide for Large Sites]]></title>
            <link>https://theseoxpert.com/articles/crawl-budget-optimization-guide</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/crawl-budget-optimization-guide</guid>
            <pubDate>Wed, 05 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A framework for diagnosing and fixing crawl waste on large sites — log file analysis, faceted navigation traps, and robots directives that actually work.]]></description>
            <content:encoded><![CDATA[Crawl budget optimization gets recommended far more often than it's actually needed. If your site has fewer than 10,000 URLs and a healthy server response time, you almost certainly don't have a crawl budget problem — Google has said as much directly. If you're running a large ecommerce catalog, marketplace, or publisher site with URL counts in the hundreds of thousands, this guide is for you.

## Step 1: Confirm you actually have a problem

Before optimizing anything, check Search Console's Crawl Stats report (Settings → Crawl Stats) for two signals:

- **Total crawl requests trending flat or down** while your site's URL count is growing — a sign Google isn't keeping pace with your content.
- **A high proportion of 4xx/5xx or redirect responses** in the crawl breakdown — a sign of wasted crawl activity that could be spent on real content instead.

If neither shows up, stop here — you likely don't have a crawl budget issue, and the time is better spent on content or on-page work.

## Step 2: Analyze server logs, not just Search Console

Search Console's crawl data is sampled and aggregated. Server log files are the ground truth — every request, every status code, every user agent, timestamped. Pull at least 30 days of logs and filter to Googlebot's verified IP ranges (verify via reverse DNS lookup, don't trust the user-agent string alone).

What to look for:

- **Which URL patterns consume the most requests.** Group by path pattern (e.g., `/products/*`, `/search?*`, `/category/*/page/*`) rather than looking at individual URLs.
- **Crawl frequency versus page value.** Cross-reference against analytics or revenue data — if Googlebot is spending 20% of its budget on pages that generate 0.1% of your organic traffic, that's your target.
- **Status code distribution.** A high volume of 301/302 redirects or soft-404s eats budget without adding indexable pages.

## Step 3: Fix the most common sources of waste

### Faceted navigation and filter combinations

Ecommerce and marketplace sites are the most common offenders. Filter and sort parameters (`?color=blue&size=M&sort=price`) can generate a combinatorial explosion of crawlable URLs, most of which are near-duplicates of a canonical category page. Fix with a combination of:

- `rel=canonical` pointing filtered/sorted variants back to the base category URL.
- Disallowing parameter patterns in robots.txt once you've confirmed they add no unique value (do this only after canonicals are in place and working — don't block before you've resolved the indexing side).
- Using the URL Parameters tool (where still available) or consistent internal linking that avoids generating those combinations in the first place.

### Internal search result pages

Internal site search results getting indexed and crawled is a near-universal issue on larger sites. These pages should almost always be `noindex` and, once confirmed clean, disallowed in robots.txt.

### Redirect chains

Every hop in a redirect chain costs a crawl request without adding a new indexable page. Audit and flatten multi-hop redirects (A → B → C should become A → C) periodically, especially after site migrations or URL restructuring projects.

### Orphaned or low-value pages still linked internally

Old pagination pages, discontinued product pages left live instead of properly redirected or removed, and thin tag/archive pages all continue consuming crawl budget as long as they're internally linked. A periodic internal-link audit against a list of pages you actually want indexed catches most of these.

## Step 4: Improve crawl rate limit, not just crawl demand

Crawl budget is a function of two things: how much Google *wants* to crawl (demand, driven by perceived value and freshness) and how much your server can *handle* without degrading (rate limit). Improving server response time directly increases the rate limit side — a server that responds in 200ms rather than 2s gets more requests in the same crawl window, independent of any content or linking changes.

## What not to do

Don't disallow parameter URLs or thin pages in robots.txt as a first step if they're already indexed — this blocks the crawler from seeing a noindex directive on those pages and can freeze bad URLs in the index indefinitely. Clean up indexing first (noindex, canonical, proper redirects), confirm it's taken effect, then restrict crawling.]]></content:encoded>
            <category>Technical SEO</category>
            <enclosure url="https://theseoxpert.com/images/articles/crawl-budget-optimization-guide.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[Title Tag Optimization: A Framework That Balances CTR and Rankings]]></title>
            <link>https://theseoxpert.com/articles/title-tag-optimization-framework</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/title-tag-optimization-framework</guid>
            <pubDate>Tue, 04 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A repeatable formula for writing title tags that satisfy both algorithmic relevance signals and human click-through psychology.]]></description>
            <content:encoded><![CDATA[A title tag has to do two incompatible-sounding jobs at once: read as an unambiguous relevance signal to a ranking algorithm, and read as a compelling reason to click to a human scanning ten other results. Optimizing for only one usually costs you the other. Here's a framework that handles both.

## The structure

A reliable default formula, in order:

**[Primary keyword or specific promise] — [Differentiator or specificity] | [Brand]**

For example: *"Core Web Vitals in 2026: Which Metrics Still Move Rankings | The SEO Xpert"* leads with the keyword-bearing topic, adds a specific angle ("which metrics still move rankings" — implying some don't, which is inherently more interesting than a generic "guide"), and closes with the brand.

This isn't the only valid structure, but it's a strong default because it satisfies both audiences in order: the keyword and specificity load first, for both algorithmic parsing and the first few words a scanning eye actually reads; the brand loads last, where it costs nothing if truncated.

## Keep it under ~60 characters, and understand why

Google doesn't enforce a hard character limit, but titles significantly longer than about 60 characters get truncated in the search results display, and — more importantly — Google has stated it's more likely to generate its own replacement title when the original is too long, too generic, or keyword-stuffed. A rewritten title is one you didn't choose, optimized by an algorithm that may not preserve your intended framing or CTA.

Check actual rewrite rates for your site in Search Console by comparing your `<title>` source against what's actually displayed for top queries — a surprisingly high percentage of sites discover 20-30% of their titles are being rewritten without realizing it.

## Front-load the keyword, but don't force it

Words earlier in a title carry more weight — both for algorithmic relevance matching and for a human eye that reads left to right and often doesn't finish scanning a long title before moving to the next result. Where it reads naturally, put your primary keyword or its core concept within the first 3-4 words.

The caveat: don't force an awkward phrase order just to front-load a keyword. "SEO Title Tag Optimization: A Complete 2026 Guide" front-loads awkwardly; "Title Tag Optimization: A Practical 2026 Framework" reads naturally and still leads with the topic.

## Specificity beats cleverness

Across most CTR studies and our own testing, titles that promise something concrete and specific — a number, a timeframe, an outcome, a named comparison — consistently outperform generic or purely descriptive titles, and they usually outperform attempts at cleverness or wordplay too. "7 Crawl Budget Fixes That Cut Wasted Requests 40%" beats both "Crawl Budget Guide" (too generic) and "Don't Let Googlebot Waste Its Time" (clever but vague about what's actually inside).

## Brand placement: usually last, sometimes first

For most content pages, put your brand name at the end of the title, after the value proposition — it's the part users care about least when scanning results, and the part that costs the least if truncated. Exceptions:

- **Homepages**, where the brand often *is* the primary search intent.
- **Strongly brand-driven queries**, where users are specifically looking for your product or service by name.
- **High-trust categories** (finance, health, legal) where leading with an established, recognizable brand name can measurably improve click-through by signaling credibility before the user reads further.

## Audit for duplicates

Duplicate or near-duplicate title tags across a site are one of the most common, and most mechanically simple to fix, technical SEO issues. They typically show up from templated pages (paginated listings, faceted category pages, thin location pages) that share a title pattern without enough per-page differentiation. Pull a full site crawl, group by title, and look for any group with more than a handful of URLs sharing an identical or near-identical title — that's usually a templating fix, not a page-by-page rewrite.]]></content:encoded>
            <category>On-Page SEO</category>
            <enclosure url="https://theseoxpert.com/images/articles/title-tag-optimization-framework.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[How to Build a Topic Cluster That Actually Ranks]]></title>
            <link>https://theseoxpert.com/articles/topic-cluster-content-strategy</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/topic-cluster-content-strategy</guid>
            <pubDate>Mon, 03 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[Step-by-step pillar-and-cluster content planning with internal linking rules, illustrated with a worked example you can copy.]]></description>
            <content:encoded><![CDATA[Most sites that attempt topic clusters get the concept right and the execution wrong — usually by treating internal linking as an afterthought rather than the mechanism that actually makes the strategy work. A pile of related articles isn't a topic cluster; a pile of related articles with a consistent, bidirectional linking structure is.

## What a topic cluster actually is

A topic cluster has two parts:

- **One pillar page** covering a broad topic comprehensively — think "Technical SEO" or "Link Building" — typically 2,000-4,000 words, structured to touch on every major subtopic without going deep into any single one.
- **Multiple cluster pages**, each covering one subtopic from the pillar in real depth — "Crawl Budget Optimization," "Digital PR for Link Building" — typically 800-1,800 words, targeting a narrower, more specific keyword.

The pillar links to every cluster page; every cluster page links back to the pillar. That consistent, bidirectional structure is the actual mechanism — it tells search engines these pages are a coherent topical unit, and it distributes authority and relevance signals across the whole cluster rather than concentrating them on one URL.

## Step 1: Choose pillar topics correctly

Good pillar candidates share three traits:

- **High search volume, but too broad to target with one page.** "SEO" is too broad to rank for or usefully write about in one document. "Technical SEO" is broad but has a coherent, boundable scope.
- **Room for at least 6-8 genuinely distinct subtopics.** If you can't list that many without stretching, the topic may be too narrow to be a pillar — it might just be a cluster page under something broader.
- **Commercial or strategic relevance to your business**, not just search volume. A pillar with high volume but no path to your actual audience isn't worth the investment.

## Step 2: Map the cluster before writing anything

Before writing a single article, list every cluster page you intend to publish under the pillar, with its target keyword. This matters because the internal linking plan depends on knowing what will eventually exist — you want the pillar page's outline to already anticipate every cluster topic, even the ones you haven't written yet, so you can build placeholder anchor text and update links incrementally.

**Worked example — pillar: "LLM & AI Search Optimization":**

| Cluster page | Target keyword |
|---|---|
| Generative Engine Optimization (GEO): How to Get Cited by ChatGPT and Claude | generative engine optimization |
| Schema Markup for AI Search | schema markup for AI search |
| How We Increased AI Citations 3x (case study) | increase AI citations |
| Why Traditional Link Building Is Breaking Down in AI Search | link building AI search |

Each of those already exists as an article on this site, interlinked back to a broader LLM optimization overview — the same structure applies whether your pillar has 4 cluster pages at launch or 12.

## Step 3: Build the linking rules, and actually follow them

Three rules, non-negotiable if you want the structure to work:

1. **Every cluster page links to the pillar**, using consistent (not necessarily identical, but topically clear) anchor text, ideally within the first third of the article rather than buried in a footer list.
2. **The pillar links to every cluster page**, organized under the subtopic heading it belongs to, so the link's surrounding context reinforces relevance.
3. **Cluster pages can link to each other** where genuinely relevant, but shouldn't substitute for the pillar link — lateral links support the structure, they don't replace the hub-and-spoke backbone.

## Step 4: Publish the pillar first, even incomplete

It's tempting to perfect the pillar page before publishing, especially since it's meant to be comprehensive. Don't wait. Publish a solid, if not yet fully expanded, pillar page first so every cluster article you publish afterward has somewhere real to link to from day one. Expand the pillar's own written sections over time as you have more cluster pages to reference and more expertise to add — but the URL and basic structure should exist before the first cluster page goes live.

## Step 5: Revisit the cluster quarterly

Topic clusters aren't a one-time project. Revisit each cluster every quarter to check for: new subtopics worth adding as cluster pages, outdated cluster content that needs a refresh (which should also bump the pillar's "last updated" reasoning if it references that content), and any cluster pages that have quietly stopped ranking and need either a content refresh or a redirect into a stronger sibling page.]]></content:encoded>
            <category>Content Strategy</category>
            <enclosure url="https://theseoxpert.com/images/articles/topic-cluster-content-strategy.webp" length="0" type="image/webp"/>
        </item>
        <item>
            <title><![CDATA[SEO Glossary: Essential Terms for the AI Search Era]]></title>
            <link>https://theseoxpert.com/articles/seo-glossary-ai-search-era</link>
            <guid isPermaLink="false">https://theseoxpert.com/articles/seo-glossary-ai-search-era</guid>
            <pubDate>Sat, 01 Aug 2026 09:00:00 GMT</pubDate>
            <description><![CDATA[A reference glossary covering classic SEO terms alongside the new GEO/AEO vocabulary — RAG, citation rate, answer engine, entity SEO, and more.]]></description>
            <content:encoded><![CDATA[Terminology in AI search is moving fast enough that even practitioners mix up terms week to week. This glossary covers the classic SEO vocabulary alongside the newer GEO/AEO terms, grouped by theme rather than strict alphabetical order so related concepts sit near each other.

## Core SEO terms

**Crawl budget** — The number of URLs a search engine's crawler is willing and able to crawl on your site within a given period. See our [full guide to crawl budget optimization](/articles/crawl-budget-optimization-guide).

**Core Web Vitals (CWV)** — Google's set of page-experience metrics: Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS).

**Canonical tag** — An HTML element (`rel="canonical"`) that tells search engines which URL is the authoritative version among duplicate or near-duplicate pages.

**Topical authority** — The degree to which a site is considered a comprehensive, trustworthy source on a given subject, typically built through topic clusters and consistent, in-depth coverage over time.

**Domain authority** — A third-party metric (not a Google ranking factor itself) estimating a site's overall link-based authority, often used as a proxy for how competitive it will be to rank.

## AI search and GEO terms

**Generative engine optimization (GEO)** — The practice of structuring content so generative AI systems are more likely to cite or reference it in their answers. See our [full GEO guide](/articles/generative-engine-optimization-geo-guide).

**Answer engine optimization (AEO)** — Used largely interchangeably with GEO, sometimes applied more narrowly to systems producing a single direct answer rather than a conversational response.

**Retrieval-augmented generation (RAG)** — The technique where an AI system retrieves relevant documents before generating an answer, then grounds its response in — and cites — those documents. This is the mechanism behind most AI Overviews and AI-assistant search features.

**Citation rate** — An informally defined but increasingly tracked metric: the percentage of relevant test queries where an AI assistant cites your content as a source. See our [citation rate case study](/articles/increasing-ai-citations-case-study) for one way to measure it.

**AI Overview** — Google's AI-generated summary shown above traditional search results for many queries, synthesizing an answer from multiple retrieved sources with citations.

**Entity SEO** — Building clear, consistent, unambiguous associations between your brand and specific concepts, products, or topics, so search and AI systems can confidently match a query to your content. Relies heavily on consistent naming and [structured data](/articles/schema-markup-for-ai-search).

**Corroboration** — The degree to which a specific claim or data point is echoed, with or without a link, across multiple credible sources — an increasingly important signal for AI citation likelihood. See [why traditional link building is shifting](/articles/link-building-model-breaking-down-ai-search).

**Speakable schema** — A schema.org specification marking specific page content as suitable for text-to-speech, originally built for voice assistants, increasingly relevant for AI systems identifying quotable content.

**Answer-first structure** — A content-writing pattern where each section opens with a direct, self-contained answer before providing supporting context — the single highest-leverage on-page GEO tactic we've measured.

## Content and link building terms

**Pillar page** — A comprehensive overview article covering a broad topic, linking out to narrower cluster pages. See our [topic cluster guide](/articles/topic-cluster-content-strategy).

**Cluster page** — A narrower, more specific article covering one subtopic of a pillar page in depth, linking back to the pillar.

**Digital PR** — A link building and brand-visibility approach centered on creating genuinely newsworthy assets (typically original data) and pitching them to journalists, rather than direct link requests. See our [digital PR playbook](/articles/digital-pr-link-building-playbook).

**Anchor text** — The clickable, visible text of a hyperlink, historically a meaningful relevance signal for the linked page's target keyword, though its weight relative to corroboration signals is shifting in an AI search context.

This glossary will be updated as the field's vocabulary continues to shift — several of these AI-search terms are less than two years old and still settling into consistent industry usage.]]></content:encoded>
            <category>General</category>
            <enclosure url="https://theseoxpert.com/images/articles/seo-glossary-ai-search-era.webp" length="0" type="image/webp"/>
        </item>
    </channel>
</rss>