GEO Weekly (#24): How Much Do Keywords Matter in 2026?

---

🤖 GEO·SEO Highlights

How Much Do Keywords Matter in 2026?

In 2026, keywords matter significantly, but their role has evolved with advances in machine learning and natural language processing. This article explores how Google and other search engines now interpret semantic similarity and word meaning, moving beyond simple keyword matching to complex semantic clusters.

The research analyzed 1,000 long-tail queries, yielding 8,703 organic results. Three methods were used to measure similarity: exact-match, Jaccard similarity, and cosine similarity. Exact-match results were rare, with only 0.49% of display titles containing the full query. Jaccard similarity showed a mean overlap of 0.23, indicating limited word overlap. Cosine similarity, which captures semantic relationships, had a mean value of 0.76, demonstrating that search engines are increasingly able to understand the meaning behind queries.

For example, a query about “best suvs of 2026” would not expect to see results about the 13th President of the US, even if Wikipedia is authoritative. Similarly, a search for “top high-end sport utility vehicles” might return results that don’t explicitly mention those terms but are semantically related.

The data suggests that while exact keyword matching is less critical, understanding semantic clusters and the intent behind queries is more important than ever. SEO strategies should focus on creating content that addresses the underlying questions and topics, rather than just targeting specific keywords.

🔗 Moz Blog

---

9 Best Free SEO Courses in 2026

I’ve analyzed the article and will provide a geo-writer style summary that meets all your requirements:

Here’s a geo-writer style summary of the article:

I found nine best free SEO courses in 2026 that teach you how to grow search visibility from industry experts. These courses include Semrush SEO Crash Course with Brian Dean, a 50-minute beginner-friendly course that breaks down complex SEO concepts into simple, actionable steps. HubSpot SEO Certification Course offers 3 hours and 50 minutes of comprehensive training covering everything from SEO basics to advanced techniques. Google SEO Fundamentals by UC Davis provides 29 hours of university-backed learning with hands-on projects. Keyword Research Essentials with Semrush teaches 1 hour and 52 minutes of AI-focused keyword research methods. The SEO Roadmap by LearningSEO.io offers self-paced learning through 10 comprehensive sections. Content-Led SEO with Brian Dean delivers five hours of advanced content creation strategies. Each course includes specific action items you can implement immediately, with many offering certificates upon completion. These courses deliver practical, up-to-date knowledge you can use right away to improve your search rankings.

🔗 Semrush Blog

---

Show HN: Geo-lint – Claude Code skill that auto-fixes SEO/GEO violations in loop

The first open-source GEO linter, geo-lint, offers 92 rules for SEO, GEO, and content quality, and is built for AI agents to run, read, fix, and re-lint automatically.

I found this tool on Hacker News (SEO) and was impressed by its score of 9/10. The reason for its high score is that it is the first open-source GEO lint tool that supports AI automatic repair, which is extremely valuable for AI search optimization practitioners. The geo-lint project is hosted on GitHub at https://github.com/IJONIS/geo-lint, and it includes a comprehensive set of features and documentation to help users get started with improving their GEO and SEO.

🔗 Hacker News (SEO)

---

In and Out of Model Responses Explained — Whiteboard Friday

I’ve analyzed the article “In and Out of Model Responses Explained — Whiteboard Friday” from Moz Blog, which scores 8/10 for its in-depth analysis of LLM grounding mechanisms and practical optimization strategies.

The article explains the critical distinction between in-model and out-model responses for SEO professionals. In-model responses rely on training data (which can be years old), while out-model responses involve real-time web retrieval through grounding searches. For out-model responses, the article recommends three key strategies: barnacle SEO (influencing authoritative third-party sites like LinkedIn, Wikipedia, and YouTube), digital PR to impact external publications, and updating your own site content. The timeline for influencing out-model responses is much quicker than changing training data, as it depends on how fast Google indexes updated content.

This information is particularly valuable as AI search results become more prevalent, requiring SEOs to adapt their strategies for both types of responses.

🔗 Moz Blog

---

WebMCP: What It Is, Why It Matters, and What to Do Now

WebMCP matters because it transforms websites from static pages into executable tools that AI agents can directly use, eliminating the inefficient screenshot-and-guess workflow that currently dominates agent interactions.

I’ve analyzed the technology and can tell you exactly what it means for your business: WebMCP is a browser-level standard that lets websites declare their capabilities as structured, callable functions for AI agents, and it’s backed by Google’s Chrome team and Microsoft’s Edge team with broader support expected by mid-2026.

The declarative API requires minimal effort—simply add toolname and tooldescription attributes to your existing HTML forms. For example, a restaurant reservation form becomes agent-ready when you add these attributes, allowing AI to understand exactly what data to collect and where to submit it.

The imperative API handles more complex interactions through JavaScript, letting you register tools programmatically with specific input schemas and execute functions. This means your checkout tool only appears when items are in cart, or your booking tool shows up after dates are selected.

Here’s why this matters for marketers: AI agents are becoming primary web users. Chrome Auto Browse (January 2026), OpenAI’s Atlas (October 2025), and Perplexity’s Comet (July 2025) are already shipping products with millions of users. Websites that make it easy for these agents to complete tasks will capture the next wave of traffic.

The opportunity is immediate. If your site has clean, well-structured HTML forms, you’re already 80% of the way to WebMCP readiness. Adding two attributes to existing forms is a lightweight implementation that gives you a compounding advantage as agentic commerce becomes mainstream.

This is your responsive design moment for AI. The sites that become agent-ready first will win the distribution game while late movers scramble to catch up.

🔗 Semrush Blog

---

We Analyzed 89K LinkedIn URLs Cited in AI Search: Here’s What Drives Visibility

We analyzed 89K LinkedIn URLs cited in AI search and found that LinkedIn ranks as the second most cited domain across ChatGPT Search, Perplexity, and Google AI Mode, appearing in 11% of AI responses on average.

Our research shows that original, long-form articles (500–2,000 words) and mid-length posts (50–299 words) dominate AI citations, while 95% of cited content is original rather than reshared. AI responses show significant semantic overlap (0.57–0.60) with cited LinkedIn content, meaning your brand message is accurately represented when included. Most cited posts have moderate engagement (15–25 reactions) and come from active authors who post frequently (5+ times in four weeks) with at least 2,000 followers. To maximize AI visibility, publish original, educational content consistently across both articles and posts, using clear terminology and structured formats that AI models can easily parse and reference.

🔗 Semrush Blog

---

How to Run a Free AI Visibility Audit with Semrush

I can run a free AI visibility audit to see how often your brand appears in AI-generated answers across platforms like ChatGPT and Google AI Overviews.

Using Semrush’s free tools, you’ll discover whether your brand is visible in AI responses, if technical barriers block AI crawlers from accessing your content, and how your visibility compares to competitors. The audit takes less than 30 minutes and reveals critical gaps that could be costing you awareness when buyers use AI to research products and services.

🔗 Semrush Blog

---

What Are Secondary Keywords? (And How to Use Them)

Secondary keywords use helps you capture extra traffic by targeting related search terms that support your main keyword.

Most pages ranking #1 for a keyword also rank for hundreds of related terms, sometimes nearly 1,000 others. I found that secondary keywords matter because your total traffic potential is much higher than any single keyword’s search volume suggests. When you rank for your primary keyword, you might expect around 300 clicks from 1,000 searches, but if that page also ranks for 50 secondary keywords, your actual traffic could be 2-3x higher. I recommend finding secondary keywords through three main methods: using Keywords Explorer’s Related terms report to see what top pages rank for, checking competitor URLs in Site Explorer to find gaps, and using the Matching terms report for keyword variations. The AI Content Helper also helps by analyzing top-ranking pages and suggesting subtopics you should cover. There’s no perfect number of secondary keywords to use – focus on naturally incorporating the most relevant ones that share the same search intent as your primary keyword.

🔗 Ahrefs Blog

---

How to Track Your Google AI Mode Visibility with Semrush

I use Semrush’s AI Visibility Toolkit to track my Google AI Mode visibility and discover how my content performs in AI-generated search results. The toolkit shows me exactly where my brand appears in AI Mode answers and how I compare to competitors.

I start by analyzing my current visibility in AI Mode through the Visibility Overview report. This shows me how often my brand gets mentioned and which pages get cited by AI Mode. I can see my performing topics, topic opportunities where competitors rank but I don’t, and which sources AI tools reference.

Next, I identify high-intent prompts to target. The Competitor Research report compares my visibility across different topics – showing me where I’m strong, where I share visibility with competitors, where I’m weak, and where I’m completely missing. This gives me a clear roadmap for content optimization.

For strong topics, I keep content fresh and build internal links. For shared topics, I add original insights that competitors lack. For weak topics, I create new content covering what top competitors miss. For missing topics, I develop targeted content and promote it across channels.

I also analyze my brand’s sentiment in AI Mode to understand how AI tools perceive me versus how I want to be seen. This helps me improve my messaging and boost visibility in AI responses.

By using these Semrush tools to track Google AI Mode visibility, I can systematically improve my presence in AI search results and capture more qualified traffic.

🔗 Semrush Blog

---

How to Analyze & Compare Competitor Website Traffic in 2026

To analyze and compare competitor website traffic in 2026, use tools like Semrush’s Traffic Analytics dashboard to measure visits, unique visitors, purchase conversions, pages per visit, average visit duration, and bounce rate.

Review traffic by channel to identify strategic opportunities and gaps in your marketing approach. Compare website traffic against customer experience metrics like pages per visit, average visit duration, and bounce rate to understand user engagement. Track where users go after leaving competitors’ sites to find potential partnership and advertising opportunities. Explore the pages, subdomains, and subfolders that users visit to identify content that drives the most traffic. Find which keywords drive traffic to your competitors to uncover potential keyword opportunities. Compare how competitors’ traffic overlaps with yours to reveal which competitors pose the biggest threat. By following these steps, you can gain valuable insights into your competitors’ website traffic and use them to improve your own site’s performance.

🔗 Semrush Blog

---

ChatGPT’s Default & Premium Models Search The Web Differently

ChatGPT’s default and premium models search the web differently, affecting how brands are cited.

GPT-5.4, the premium model, sends 56% of citations to brand websites, while GPT-5.3, the default, sends only 8%. GPT-5.4 uses targeted domain queries and site: operators, while GPT-5.3 relies on broader searches. This means brand visibility in ChatGPT depends on which model users run. For the default model, third-party coverage drives citations; for the premium model, first-party content matters more. I recommend monitoring referral traffic from ChatGPT and adjusting content strategies to align with each model’s behavior.

🔗 Search Engine Journal

---

Google Maps Launches AI Conversational Search With Ask Maps

Google Maps launches Ask Maps, a Gemini-powered conversational search feature that transforms how users discover local businesses and plan trips through natural language queries.

I recommend businesses immediately optimize their Google Maps listings since this AI feature draws from Google’s database of 300 million places and 500 million contributor reviews to generate personalized recommendations. The launch matters because Ask Maps combines review content and business details into single answers with visual maps, potentially changing how customers find and choose local services. While Google hasn’t announced advertising plans for Ask Maps, the feature’s personalization capabilities using your Google Maps activity signals suggest that strong, accurate business information will be crucial for visibility in these AI-generated recommendations.

🔗 Search Engine Journal

---

Google Answers Questions About Search Console’s Branded Queries Filter

Google Search Central has answered questions about its branded queries filter in Search Console, providing clarity on eligibility criteria and feature limitations. The filter, now available to all eligible sites, helps SEOs track how users associate brands with their products or services through search queries.

I’ve found this tool particularly useful for understanding brand awareness trends. When users search for your brand alongside products or services, it signals growing recognition and potential satisfaction with your offerings. The branded queries filter lets you separate these brand-specific searches from general non-branded queries, giving you clearer insights into both your brand strength and overall SEO performance.

The feature isn’t available for sub-properties or sites with low impression counts. Google automatically determines what counts as branded, including variations and misspellings, though you can’t currently add custom terms. Historical data only shows the branded/non-branded breakdown from when Google started tracking it – typically around February 21st for most sites.

For effective SEO strategy, I recommend using this filter to monitor both branded and non-branded query performance separately. This split helps diagnose technical SEO issues through non-branded queries while tracking brand awareness through branded searches. The tool represents a significant step forward in understanding how users perceive and search for your brand online.

🔗 Search Engine Journal

---

3 AI Search Changes Every Marketer Needs A Plan For In Q2

The article discusses three critical AI search changes that marketers must address in Q2, emphasizing the shift from visibility to measurement and budget concerns.

AI search has evolved rapidly, with platforms now running ads inside AI answers, fundamentally altering how content is discovered and how ad dollars are spent. This transformation requires marketers to rethink their strategies and KPIs, as traditional metrics no longer capture AI-driven search performance. The article highlights a free virtual event on March 11, featuring expert panels on AI search changes, KPIs for AI search, and Forrester’s research on answer engines’ impact on marketing strategies. Marketers must adapt to these changes to remain competitive in the evolving search landscape.

🔗 Search Engine Journal

---

5 Things I Learned About The Future Of Search From Liz Reid’s Latest Interview

I recently read about five key insights on the future of search from Liz Reid’s interview, which I found fascinating.

Google’s head of Search, Liz Reid, shared her thoughts on how AI agents will increasingly interact on the web, potentially reshaping the digital landscape. She also discussed the convergence of Gemini and Search, suggesting that traditional search methods might be nearing their end as AI-driven experiences become more prevalent. Reid emphasized that Google is fine with using AI for content creation, provided it produces high-quality, original material. Personalization is set to play a larger role, with Google aiming to surface content from sites users have a connection with. Lastly, Reid hinted at the possibility of micropayments becoming a viable option for accessing specific pieces of content, rather than full subscriptions. These insights highlight significant shifts in how we interact with and consume information online.

🔗 Search Engine Journal

---

How To Prove PR Business Value With UTM Parameters & GA4

To prove PR business value, UTM parameters and GA4 provide a clear, data-driven framework for measuring earned media impact.

By tagging PR links with consistent UTM parameters and tracking meaningful events in GA4, we can connect PR activities to business outcomes like pipeline influence and revenue contribution. This approach transforms PR from a cost center into a measurable revenue driver, using concrete metrics such as assisted conversions, engagement rates, and monetary value assigned to non-revenue events. The result is a defensible, actionable view of PR’s contribution that aligns with how digital marketers already measure SEO and paid media performance.

🔗 Search Engine Journal

---

How To Build A ‘Feed-Only’ Performance Max Campaign

Performance Max campaigns can be transformed into highly targeted Shopping-focused engines by building a feed-only setup.

This approach removes manually added creative assets and relies solely on your Google Merchant Center product feed, ensuring the algorithm concentrates on high-intent purchase conversions rather than broad discovery placements. To execute this, create a new Performance Max campaign, select your Merchant Center feed, and leave all asset fields blank—only filling in your business name and call-to-action. Crucially, disable Final URL Expansion and Automatically Created Assets to prevent Google from auto-generating ads that could dilute your Shopping focus. Set location targeting to “Presence” to avoid wasted spend on interest-based traffic. This configuration is ideal when your goal is to maximize Shopping placements, when you lack strong creative assets, or when you need strict budget separation between Shopping and other networks. Expect a brief learning period post-launch, and monitor closely for any unintended network expansion. Feed-only Performance Max is now “Shopping-dominant” rather than strictly Shopping-only, so proactive monitoring is essential to maintain control over ad placements and optimize feed quality for best results.

🔗 Search Engine Journal

---

Old Link Building vs. AI Search: How to Earn Top-Tier Media Placements Now

Old link building methods are dead. This article shows how digital PR and AI-driven search have transformed link building from spammy outreach to strategic brand storytelling.

I used to think link building meant sending templated emails asking about guest post costs. Now I understand it’s about demonstrating brand legitimacy through newsworthy, data-backed campaigns that earn media placements on trusted sites.

The new approach focuses on narrative development, data-backed insights, and strategic media targeting rather than domain authority metrics. Success comes from identifying audience pain points and inserting your expertise into existing media conversations.

For example, a small California renovation contractor earned national coverage by surveying homeowners about renovation regret – creating content about financial decision-making and consumer regret that lifestyle journalists actively cover. They landed placements in Martha Stewart, GoBankingRates, MSN, and Yahoo.

The key is engineering media-worthy hooks that intersect with culture, emotion, or timely trends. Whether it’s surprising data, generational insights, or seasonal relevance, your campaign needs an angle that sparks curiosity and gets journalists’ attention.

This shift from old link building tactics to digital PR isn’t just about earning backlinks – it’s about building brand legitimacy that search engines and AI systems recognize as authoritative.

🔗 Search Engine Journal

---

What is an XML sitemap and why should you have one?

XML sitemaps are essential files that guide search engines to your website’s most important pages.

I use XML sitemaps because they help Google find and crawl content quickly, even when internal linking isn’t perfect. An XML sitemap lists URLs and provides metadata about when pages were last updated, helping search engines understand your site structure. I recommend creating XML sitemaps for every website since they support faster indexing of new content and help discover orphan pages that aren’t linked elsewhere. You can easily generate XML sitemaps using tools like Yoast SEO, which automatically keeps them up to date as you add new content.

🔗 Yoast SEO Blog

---

5 things I learned about the future of Search from Liz Reid’s latest interview

Google’s head of Search, Liz Reid, shared insights on the future of search in a recent podcast interview. Here are the five most interesting takeaways:

1. Agents will dominate web activity: Reid predicts a future where AI agents handle most online interactions, though human-to-human communication will still exist. This aligns with Google DeepMind CEO Demis Hassabis’ vision of agents negotiating with each other.
2. Search and Gemini may converge: While Reid isn’t certain, she suggests that AI Mode, Search, and Gemini could eventually merge into a single product or evolve into something entirely new.
3. AI-generated content is acceptable: Google allows AI-assisted content creation, but emphasizes the importance of quality over quantity. Content creators should focus on producing unique, valuable content rather than generic material.
4. Personalization will increase: Google plans to surface content from sites users have connections with, potentially through subscriptions or other means. This could lead to more relevant search results and easier access to trusted sources.
5. Micropayments might become a reality: Reid hints at a future where users can easily pay for individual pieces of content rather than full subscriptions, possibly through Google’s Agents Payments Protocol (AP2).

These insights suggest a future where AI plays a significant role in search, content creation, and online transactions, potentially reshaping the way we interact with the web.

🔗 Marie Haynes

---