How to Add Google’s “Preferred Sources” Button to Your Website

How to Add Google’s “Preferred Sources” Button to Your Website

A Complete 2026 Guide for Publishers

What it does, why it matters, and exactly how to embed it, with code.

Example of the Add to Preferred Sources button placed on a website homepage, showing the actual Google button with its multicolor logo

AI Overviews are quietly stealing clicks that used to land on your site. Google’s answer is a button: drop it on your pages, and a reader can mark your site as a trusted source in one tap, no detour to Google’s settings required.

This guide covers what it does, whether it’s worth your five minutes, and exactly how to install it, code included.

Quick Summary

  • What it is: a button readers tap to mark your site as trusted, boosting your visibility in Top Stories, AI Overviews, and AI Mode.
  • Launched: August 20, 2026, as an embeddable upgrade to a feature that’s existed since mid-2025.
  • Setup: one script tag, one div. No redesign, no dev sprint.
  • Payoff: Google says readers are about 2x more likely to click through to a source they’ve marked preferred.
  • Scale: 600,000+ unique sources selected as of late August 2026, up from ~90,000 in December 2025.
  • Catch: it’s not a ranking factor. It builds repeat visibility with readers who already trust you.

What Is the Preferred Sources Button?

Preferred Sources is a personalization feature in Google Search. A reader can flag a specific website as one they trust, and Google uses that signal to show more of that site’s content when it’s relevant, particularly inside Top Stories, AI Overviews, and AI Mode, which happen to be the surfaces where organic clicks have taken the biggest hit from AI generated answers.

Before this button existed, adding a preferred source meant leaving whatever you were reading, opening Google’s own source preferences page, and searching for the publication by name.

Most readers never bothered to finish that. The new button collapses the whole thing into one click. A reader taps “Add to Preferred Sources” on your page, confirms in a small pop-up, and lands right back where they were reading, usually within a couple of seconds.

Diagram comparing the old multi-step flow to add a Google preferred source with the new one-tap button flow
Figure 1. The old multi-step flow compared to the new one-tap flow.

A Quick History of Preferred Sources

Preferred Sources didn’t appear overnight. It has been rolling out in stages for over a year, and the embeddable button is really the latest step in a longer push to turn it into something ordinary readers actually use, not just something that exists in a settings menu nobody visits.

Timeline of Google Preferred Sources rollout milestones from mid-2025 to August 2026
Figure 2. Key milestones in the Preferred Sources rollout.

Adoption has picked up at each stage, and it jumped noticeably once the button removed the friction of Google’s settings page:

Bar chart showing growth in unique sources selected as preferred, from 345,000 in May 2025 to over 600,000 in August 2026
Figure 3. Growth in unique sources selected as preferred, based on Google’s own reported figures.

Why Add It to Your Site?

Embedding the button costs one script tag and a few minutes of your time. Here’s what publishers stand to get back for that small effort.

Infographic summarizing four benefits of the Preferred Sources button: 2x repeat click-through, more Top Stories visibility, more AI Overviews visibility, one-tap setup
Figure 4. A summary of the main benefits Google and early adopters have reported.
  • More visibility where it counts. Preferred sources get surfaced more often in Top Stories, and since May 2026, inside AI Overviews and AI Mode answers too, two places that have otherwise been hard for publishers to influence directly.

 

  • Readers come back. Google reports that people are about twice as likely to click through to a source again once they’ve marked it as preferred, compared with sources they haven’t.

 

  • It’s a real, stated signal, not a guess. Instead of Google inferring your relationship with a reader from indirect click data, the reader is telling Google outright that they trust you. That tends to be more reliable than anything inferred.

 

  • Low effort, no redesign. The button is a drop-in script and a single div. You don’t need to touch your site’s design system or rebuild any pages to use it.

 

  • It keeps readers on your site. Because the confirmation flow drops readers back on the exact page they were on, you don’t lose the session the way the old settings-page flow used to.
Worth knowing
Preferred Sources isn’t a direct ranking factor. Adding the button won’t move your position in regular search results. Its value is in repeat visibility with readers who already know your work, especially inside AI generated answers where organic links are otherwise scarce.

Step-by-Step: How to Embed the Button

The whole implementation comes down to two pieces of code: a script that Google hosts, and a container element that tells the script where to render the button.

Step 1: Load Google’s publisher library

Add this script tag once, ideally in your page’s <head>, or wherever your site loads third-party scripts:

<script async
  src="https://news.google.com/swg/js/v1/publisher.js">
</script>

It loads asynchronously, so it won’t block the rest of your page.

Step 2: Place the button

Add this div wherever you want the button to appear. Common spots are near the byline, at the end of an article, or in the site footer:

<div google-add-preferred-source-btn></div>

The script detects that attribute automatically and renders the button in place. You don’t need to hard-code your own domain. Google reads it from the page.

Step 3 (optional): Match your site’s theme

The button defaults to a light theme. If it’s sitting on a dark background, switch it with a data attribute:

<div google-add-preferred-source-btn data-theme="dark"></div>

Step 4 (optional): No-JavaScript fallback

If your CMS doesn’t allow custom scripts, you can link readers to Google’s source preferences page instead, using a deeplink. It’s a bit clunkier since the reader still has to search for your site by name once they land there, but it works anywhere you can place a normal link, including newsletters and social posts:

<a href="[Google source-preferences deeplink for your domain]">
  Add us as a preferred source
</a>

You’ll find the exact deeplink format in Google’s Search Central documentation for this feature.

Step 5: Test it

Load the page and click the button yourself. Confirm two things: the “Add as a Preferred Source on Google” screen shows your publication correctly, and clicking “Add” sends you back to the same page you started on.

Step 6 (optional): Track clicks

If you use Google Tag Manager, its built-in listener should pick up button clicks on its own. If you’re running gtag.js without GTM, fire the event manually:

window.gtag('event', 'preferred_source_button_click');

Confirm the event fires correctly using GTM Preview mode or GA4 DebugView before you rely on it, then treat it as a normal engagement metric in your reporting.

Where to Put It (and How to Get Readers to Notice)

A button nobody sees doesn’t do much good. A few placement and promotion tips worth considering:

  • Put it somewhere readers naturally pause, like near the byline, at the end of an article, or in a persistent footer, rather than burying it in a sidebar.
  • Use it once per page. Competing calls to action tend to reduce clicks on all of them, this one included.
  • Tell your existing audience it exists. A short newsletter mention or an on-site banner for a week or two will do more than the button sitting there quietly on its own. Most readers won’t notice a new element unless you point it out.
  • Prioritize pages with returning visitors, like your homepage or most-read articles. The button is really about deepening a relationship a reader already has with you, not creating one from scratch.

A Few Things to Keep in Mind

  • It’s not a ranking factor. It changes how often you’re surfaced to people who’ve already chosen you, not your position in standard organic results.
  • It depends on existing trust. The button turns loyal readers into a stronger signal, but it doesn’t generate loyalty on its own, so it works best on sites that already have a returning audience.
  • It requires JavaScript. Use the deeplink fallback if your platform restricts custom scripts.
  • Some in the SEO industry have pointed out a bit of tension here. Google already infers a lot about reader preference from click-through behavior, so asking readers to state it outright is, in a way, formalizing something Google could partly see already. That’s a fair point, but it doesn’t really change the practical upside. A clearly stated signal is still more reliable for everyone involved than one that’s guessed at.

Frequently Asked Questions

What is Google’s Preferred Sources button?

It’s an embeddable button that publishers add to their websites so readers can mark that site as a preferred source directly from the page they’re reading, without navigating to Google’s settings.

When did the Preferred Sources button launch?

Google released the embeddable button on August 20, 2026. The underlying Preferred Sources feature has existed since mid-2025, first as a Labs experiment.

Does adding the button improve my Google rankings?

No. It doesn’t directly affect ranking in standard organic search results. It increases how often your content is shown to readers who’ve marked you as preferred, especially in Top Stories, AI Overviews, and AI Mode.

How do I add the button to my website?

Add Google’s publisher script to your page, then place a div with the google-add-preferred-source-btn attribute wherever you want the button to appear. No further configuration is required.

What if my website can’t run custom JavaScript?

Use a deeplink to Google’s source preferences page instead. It requires the reader to search for your site by name once they land there, so it’s less smooth, but it works anywhere a normal link can go.

How many people are using Preferred Sources?

Google reported more than 600,000 unique sources selected as of late August 2026, up from around 90,000 in December 2025.

Conclusion

The Preferred Sources button is a small, low-cost addition. One script, one div, a few minutes of testing. It won’t move your organic rankings, but in a search landscape where AI generated answers are absorbing a growing share of clicks, giving your existing readers an easy, one-tap way to keep choosing you is a reasonable bet. Especially given that Google says people who use it are about twice as likely to come back.

If you decide to add it, don’t stop at the code. Tell your readers it’s there. The button only works if people actually see and use it.

Sources: Google Search Central documentation; Google’s official blog post on Search/News/Discover personalization; reporting from Search Engine Roundtable, PYMNTS, Search Engine Journal, TechWyse, and Technology.org (August 2026).

Is SEO Still Relevant for Generative AI Search?

ChatGPT Image May 19 2026 12 56 38 AM

Search engines have always evolved. From ten blue links to featured snippets, voice search to personalised results, the way people find information online has never stood still. But the rise of generative search has sparked a question that marketers, business owners, and content creators are all asking right now: does SEO still matter?

The short answer is yes. But the longer answer is more interesting, and more useful. If you are working with or looking for an SEO agency in London, understanding how generative search actually works will help you invest your efforts in the right places.

What Is Generative Search and How Is It Different?

Traditional search returns a list of links. You search, you scan, you click. Generative search does something different. It reads across multiple sources, synthesises the information, and produces a direct answer, often without the user needing to click anywhere at all.

Google’s Search Generative Experience and similar tools from Bing and others are already changing how search results pages look and feel. Instead of ten ranked links, users might see a conversational summary at the top, with source links appearing below or alongside it.

This has understandably made some people nervous about organic traffic. If the answer is served directly on the results page, will anyone click through to your website?

Why SEO Is More Important, Not Less

Here is the thing that gets missed in most conversations about this shift. Generative search does not create information. It pulls from existing content on the web, evaluates its credibility, and surfaces what it deems most trustworthy and relevant.

That means the websites and pages being cited inside these generated answers are the ones that have strong SEO fundamentals behind them. Quality content, strong authority signals, clean technical structure, and genuine topical depth are exactly what gets referenced.

If your content is weak, vague, or poorly structured, generative search will not find it worth citing. If it is well-optimised, well-written, and genuinely useful, it stands a strong chance of appearing not just in traditional results but inside the generated answer itself.

What Changes and What Stays the Same

Some things genuinely are shifting. Here is what you need to think about differently:

  • Clicks may decrease for simple informational queries. If someone asks a basic question and gets the answer immediately, they may not visit your site. This is not new, featured snippets have done this for years.
  • Brand visibility inside answers matters more. Being cited in a generated response builds awareness even without a click. The long-term brand value of that is significant.
  • Content depth now carries more weight. Thin content that exists purely to rank for a keyword has less chance of being picked up as a trusted source. Substance matters more than volume.
  • Structured data and schema markup become more valuable. Helping search engines understand exactly what your content contains makes it easier for generative systems to pull from it accurately.

What stays exactly the same:

  • The need for a technically sound website
  • Building genuine authority in your niche
  • Earning quality backlinks from credible sources
  • Writing content that actually serves the reader
  • Optimising for search intent, not just keywords

The Trust Factor Has Never Been Bigger

One thing generative search has amplified is the importance of trust. These systems are not just looking for pages that mention the right words. They are looking for sources that demonstrate expertise, experience, authority, and trustworthiness.

Google has been signalling the importance of these qualities through its guidelines for years. Generative search has made them even more central to how content gets surfaced and cited.

Businesses that have invested in building real credibility online, through consistent publishing, genuine expertise, and solid technical foundations, are far better placed in this new landscape than those chasing shortcuts.

What Good SEO Looks Like Right Now

If you want your website to perform well across both traditional and generative search, the priorities are clear:

  • Create content that answers real questions thoroughly and honestly
  • Build topical authority by covering your subject area in depth, not just broadly
  • Make sure your website loads quickly, works on mobile, and is easy to crawl
  • Use structured data to help search engines understand your content
  • Earn links from websites that are genuinely relevant and respected in your space
  • Keep your content fresh and accurate, outdated information loses credibility fast
  • Write for people first and search engines second, always

The Businesses That Will Win

The shift toward generative search is not a threat to good SEO. It is a filter. It is removing the advantage that used to come from publishing large volumes of low-quality content optimised purely around keywords. It is raising the bar for what it means to be a trusted, visible presence online.

Businesses that have always focused on quality, real value, and genuine expertise will find this transition works in their favour. Those relying on outdated tactics will feel the squeeze.

The role of a knowledgeable SEO partner has not diminished through this change. If anything, it has become more strategic. Getting the fundamentals right, understanding how search intent is evolving, and ensuring your content is genuinely worthy of being cited, these are not simple tick-box tasks.

A Final Word

Search is changing, but the goal of search has never changed. People want reliable, relevant, useful answers. The businesses that provide those answers clearly and credibly will always find a way to be visible, whatever form the results page takes.

SEO is not becoming irrelevant. It is becoming more meaningful.

Gemini SEO: How to Rank on Google Gemini in 2026

freepik a closeup shot of a circuit board with the letters 64024

The “Blue Link” is dead.

If you’re still checking your Google Search Console and obsessing over whether you’re #1 or #3 for a specific keyword, you’re measuring the height of a sinking ship. In early 2026, the digital marketing world hit a collective wall. Recent data shows that click-through rates (CTR) for traditional organic search results have plummeted by over 50% in just six months.

Why? Because Google Gemini doesn’t want your users to leave.

With the introduction of the “See More” redirect, Google has effectively moved the goalposts. When a user clicks for more info, they aren’t sent to your website – they are pushed deeper into AI Mode, a high-speed, conversational interface where Gemini summarizes your content, answers the user’s follow-up questions, and even completes the purchase for them.

For businesses trying to stay visible in this rapidly changing search environment, working with an experienced SEO agency in London has become increasingly important. Modern SEO strategies are no longer only about ranking pages but about ensuring your brand becomes a trusted source that AI systems reference.

Welcome to the era of Zero-Click Authority. If you aren’t the source Gemini cites, you don’t exist.

The New Reality: Data Points for 2026

We are no longer optimizing for humans who scroll; we are optimizing for LLMs (Large Language Models) that synthesize. To survive, you need to ditch “Rankings” and start tracking Model Citation Rate (MCR).

The 2026 Benchmarks

  • Model Citation Rate (MCR): This is your new North Star. It measures how often Gemini (or GPT-5) pulls your brand as the primary factual source in an AI response. A “Good” MCR in 2026 is 12% or higher for your niche.
  • Gemini 3 Flash vs. GPT-5: Our internal testing shows a massive divergence. Gemini 3 Flash favors “ecosystem signals” (YouTube, Google Maps, Chrome history), while GPT-5 leans heavily on “Direct Expert Attribution” (Whitepapers and deep-domain forums).
  • Agentic Commerce Impact: 40% of all e-commerce “searches” are now handled by AI Agents. The user says, “Gemini, find me a sustainable running shoe under $150 and buy it.” If your product isn’t formatted for Agentic Parsing, you lose the sale before the user even sees a photo.

Picture1

Visualizing the “Great Crossover” where AI-generated answers became the primary click-destination over traditional organic listings.

The Core Framework: The AI Signals Framework

To rank in 2026, you need to feed the machine what it craves. We call this the AI Signals Framework. It’s not about tricking an algorithm; it’s about becoming the most “digestible” truth on the internet.

Picture2

  1. Semantic Saturation (Knowledge Clusters)

Keywords are for 2019. Gemini 3 Flash thinks in Knowledge Clusters.

If you want to rank for “New Delhi Real Estate,” you can’t just repeat the phrase. You must achieve “Semantic Saturation” – covering the air quality index of specific neighborhoods, the proximity to the new Rapid Rail corridors, and the legal nuances of the RERA Act. Gemini looks for the breadth of your expertise, not the density of your keywords.

  1. Structured Data 3.0

Schema.org isn’t optional anymore—it’s the “API of the Web.” In 2026, we use Schema 3.0, which includes:

  • isVerifiedSource: A cryptographic tag linking your content to a verified human author.
  • agentAction: Code that tells Gemini exactly how an AI agent can interact with your pricing or booking system.

If your site isn’t “Agent-Readable,” Gemini will skip you for a competitor who is.

  1. Multi-Platform Citation (The Reddit/YouTube Alpha)

Gemini is a “Social Searcher.” It prioritizes sources that are being discussed in real-time on high-trust platforms.

  • The 5x Rule: A mention on a high-upvote Reddit thread or a citation in a YouTube transcript is now 5 times more valuable for your Gemini ranking than a traditional backlink from a guest post.
  • Why? Because Gemini uses these as “Human Sentiment Signals” to ensure it’s not recommending “AI-slop” to its users.

Localized AI: Winning in Delhi, New York, and Beyond

Gemini’s “Near Me” logic has evolved. In 2026, it doesn’t just look at distance; it looks at Live Context.

If a user in Delhi asks, “Where should I go for lunch?” Gemini considers:

  1. Zomato/Google Map sentiment (The last 24 hours of reviews).
  2. Current traffic on the Outer Ring Road.
  3. Real-time menu availability (parsed via Agentic AI).

To rank locally, your “Google Business Profile” must be updated daily, not monthly. The AI favors businesses that show “Active Life.”

The ‘Agentic’ Checklist: Your 2026 Action Plan

Picture3

Stop auditing for 2026. Start optimizing for the Agent. Use this technical checklist:

  • Enable “AI-Readiness”: Ensure your robots.txt explicitly allows GoogleOther and Gemini-Bot to crawl your full JavaScript renders.
  • Deploy Knowledge Graphs: Use internal linking to create a “topic map” that even a basic LLM can map in under 50ms.
  • Optimize for “Query Fan-Out”: Write your headers as answers to the 5 most likely follow-up questions a user would ask Gemini.
  • Human-in-the-Loop Verification: Add “Expert Review” timestamps to every page. Gemini filters for “freshness” more aggressively than the old core updates ever did.
  • Video-to-Text Mapping: Ensure every YouTube video you produce has a timestamped, keyword-rich transcript on your site. Gemini loves cross-referencing video data with text data.

The Final Word: Be the Source, or Be Forgotten

The shift to Gemini Search isn’t a “tweak” to the algorithm; it’s a total replacement of the search experience. The websites that thrive in 2026 won’t be the ones with the best “SEO hacks.” They will be the ones that Gemini trusts to provide the “Final Answer.”

Picture4

A flow chart showing how Gemini identifies a need, scans verified Knowledge Clusters, checks social sentiment on Reddit, and completes a transaction via an Agentic API.

The era of the blue link is over. The era of the Cited Authority has begun. Are you in the summary, or are you in the graveyard?

How to Optimize for Google’s AI Overviews (Without Falling for the AI-SEO Myths)

seo

Editor’s note on scope: This guide covers the core mechanics of Google’s AI Overviews and AI Mode – what they are, how they actually work, and what’s proven to influence whether your content gets cited. For the newest 2026 agentic-search and Gemini-specific tactics, see our companion guide, Gemini SEO: How to Rank on Google Gemini in 2026.

Most “how to optimize for AI search” content repeats the same five bullet points: write conversationally, add schema, build E-E-A-T, diversify formats, speed up your site. None of it is wrong exactly, but almost none of it is sourced either. So before adding another generic checklist to the pile, it’s worth answering a more basic question: what has Google actually said about this?

What AI Overviews Actually Are

AI Overviews are AI-generated summaries that appear above traditional results on some Google searches. They started as an experiment called Search Generative Experience (SGE), shown publicly at Google I/O in 2023, and launched to US users under the AI Overviews name in May 2024. The rollout has expanded since, alongside a more conversational version called AI Mode.

Mechanically, AI Overviews work on a retrieval-augmented generation (RAG) pattern: Google retrieves relevant, already-indexed pages for a query, evaluates them against its normal quality and relevance signals, and passes the top candidates to a Gemini model, which generates a synthesized answer and links back to the sources it drew from.

That single fact undercuts a lot of the advice floating around the internet, including, honestly, our own earlier version of this article.

The Myth Worth Killing First

Here’s the part most SEO content skips: Google’s own documentation is direct about this. In its “AI features and your website” guidance, Google states that the best practices for regular SEO remain relevant for AI features in Search, and that there are no additional requirements to appear in AI Overviews or AI Mode – no separate index, no dedicated schema type that unlocks citations, no secret AI-ranking algorithm.

That doesn’t mean there’s nothing to do. It means the leverage is in doing standard SEO fundamentals well enough to be a genuinely good candidate for retrieval and summarization – not in chasing tactics that only exist in third-party blog posts.

What Actually Moves the Needle

1. Rank well first. Structure second.

If your page isn’t competitive in the top organic results for a query, restructuring it for “AI readability” won’t fix that. AI Overviews cite from pages that are already earning trust and relevance signals in regular search. Treat everything below as a way to make an already-strong page easier to summarize accurately – not a shortcut around ranking.

2. Answer the question early and directly

Pages that bury the answer under three paragraphs of preamble are harder for a model to extract cleanly. Put a clear, self-contained answer to the main query in the first few sentences, then expand with detail, context, and nuance afterward. This also just makes the page better for human readers who are scanning, not reading start to finish.

3. Cover the topic completely, not across five thin pages

Thin, narrow pages that each answer one sliver of a topic are less useful to a retrieval system than one well-structured page that covers the topic’s main facets and links out to deeper supporting content. If you have multiple posts competing for the same query intent, that’s not additional coverage – it’s the same authority getting split three ways. Consolidate before you expand.

4. Make E-E-A-T verifiable, not just claimed

Writing “we follow E-E-A-T best practices” doesn’t create E-E-A-T. It’s demonstrated through specifics: a named author with a real, checkable credential; original data, testing, or a documented outcome rather than paraphrased advice; independent sites linking to or citing the page; and visible transparency about who wrote the content and why. The checklist below is a working self-audit, not a promise.

 

 

eeat-checklist-generative-search

 

 

5. Use structured data for clarity, not as a citation hack

Article, FAQ, and Organization schema help machines – both Google’s systems and other AI crawlers – parse what a page is and what it’s answering. Google is explicit that this doesn’t guarantee an AI Overview citation. Use it because it removes ambiguity, not because it’s a lever you can pull for guaranteed visibility.

6. Keep technical SEO boring and solid

None of the above matters if the page can’t be crawled and rendered properly. Confirm important pages aren’t blocked by robots.txt, avoid unnecessary JavaScript rendering delays, keep URLs clean and descriptive, and make sure load speed and mobile experience are actually good – not just checked off a list.

How the Process Fits Together

 

how-ai-overviews-generate-answers
how-ai-overviews-generate-answers

Myths We’re Retiring

We got a few things wrong in earlier versions of this kind of advice, and it’s worth naming them plainly rather than quietly editing them out.

 

ai-search-myth-vs-reality
ai-search-myth-vs-reality

Frequently Asked Questions

Do I need different schema markup specifically for AI Overviews? No. Google Search Central does not publish a separate schema type for AI Overviews. Standard Article, FAQ, and Organization markup – the same structured data used for regular Search – is what’s recommended.

Will optimizing for AI Overviews hurt my regular organic rankings? Not if you’re following genuine SEO fundamentals. The signals used to rank organic results and the signals used to select AI Overview sources overlap heavily, since both draw from the same index.

Can a new or small website appear in AI Overviews? Yes, but the same trust-building rules apply as with regular search: sites with thin history, weak authority, or no demonstrated expertise are less likely to be retrieved as a trusted source, regardless of how well-formatted the page is.

How do I know if my content is actually being cited in AI Overviews? Google Search Console provides some visibility into AI Overview impressions for your pages. Cross-reference that data against your top-performing organic queries to identify where you’re already close and what’s missing.

Should I rewrite every page for “AI readability”? No. Prioritize pages that already rank well but bury the answer, or pages competing against your own content for the same query. Rewriting everything at once is a worse use of time than fixing the handful of pages with the clearest upside.

Where This Leaves You

If there’s one thing worth taking from this: stop optimizing for a system that doesn’t exist. There’s no separate “AI SERP algorithm” to reverse-engineer. There’s the same search index, the same quality bar, judged by a model that’s now doing the summarizing instead of a blue-link list doing the sorting. The fundamentals didn’t change. What changed is how unforgiving they are of thin, unverifiable content.

If you want a second set of eyes on where your content is actually losing ground – whether that’s cannibalization between your own pages, missing E-E-A-T signals, or technical issues blocking retrieval – that’s the kind of audit a PPC and SEO agency in London can walk through with you directly, using your actual Search Console data rather than generic advice.