Discover is the feed of articles and videos that Google shows in the Google app, on the google.com mobile homepage and in Chrome on mobile, with no query typed. It is part of Google Search, but it works more like a social feed: Google predicts what someone is interested in and shows them content before they search for it. For news sites, publishers and content-heavy blogs, it can rival Search as a traffic source. For most B2B and ecommerce sites it is a bonus, not a strategy.
How Discover differs from Search#
| Search | Discover | |
|---|---|---|
| Trigger | A query | The user's interests, activity and follows |
| Content | Evergreen and fresh | Mostly fresh, timely or newly relevant content |
| Traffic shape | Steady, compounding | Spiky: a story can surge for a day, then fade |
| What you optimise | Relevance to queries | Topic expertise, freshness, image and title appeal |
Google states that Discover traffic is "less predictable or dependable" than keyword-driven Search traffic, because interests change, content types shift and Search updates apply (Get on Discover). Google has also said since 2023 that Discover uses many of the same signals as Search, so the quality work in E-E-A-T & Quality Guidelines carries over. Plan for Discover as upside, and never forecast a business on it.
Eligibility and content policies#
There is nothing to apply for. Content is eligible when it is indexed by Google and meets Discover's content policies; no special tags or structured data are required. If a page cannot rank in Search because of a crawling or indexing problem, it will not reach Discover either.
The Discover content policies sit on top of Google's general spam policies and cover dangerous content, deceptive practices, harassing content, hateful content, manipulated media, medical content, sexually explicit content, violent extremist content, violent and gory content, and vulgar language and profanity. Two rules catch ordinary publishers most often:
- Misleading previews. Google prohibits preview content that "misleads users to engage with it by promising details which aren't reflected in the underlying content."
- Transparency for news. News sources should provide clear dates and bylines, information about authors, the publication and the publisher, and contact information.
Violations can lead to a Discover manual action, reported under Security & Manual actions in Search Console. Google's documentation has also said since 2021 that Discover filters out content that would confuse people in an interest feed, such as job applications, petitions, forms, code repositories, and satire stripped of its context.
Images: the biggest technical lever#
Discover is a visual feed, and large images are what make a card stand out. Google's requirements and recommendations (Get on Discover):
- Use high-quality images that are at least 1200px wide, over 300,000 pixels in total, ideally 16:9.
- Allow large previews with the
max-image-preview:largerobots setting (or by using AMP). - Tell Google which image you prefer with schema.org markup or the
og:imagemeta tag. Since March 2, 2026 Google documents that it uses both as sources for thumbnails in Search and Discover. - Avoid generic images such as your site logo, and avoid text-heavy images.
<meta name="robots" content="max-image-preview:large" />
<meta property="og:image" content="https://example.com/images/ai-crawler-study-1600x900.jpg" />
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "We blocked 12 AI crawlers for 90 days. Here is what changed.",
"image": ["https://example.com/images/ai-crawler-study-1600x900.jpg"],
"datePublished": "2026-10-01T09:00:00+00:00",
"author": { "@type": "Person", "name": "Sam Rivera", "url": "https://example.com/about/sam-rivera" }
}
</script>If you run Next.js, set robots.googleBot["max-image-preview"] = "large" in your root metadata export and give every article an openGraph.images entry at 1200px or wider. Check your CMS too: some themes still output max-image-preview:standard or tiny social images by default. More on image delivery in Image & Video SEO.
Titles and content: what the February 2026 Discover core update rewards#
On February 5, 2026 Google released its first core update that applied only to Discover. It rolled out to English-language users in the US and completed on February 27, 2026, longer than the "up to 2 weeks" Google estimated (Search Status Dashboard). Google said it would show (Google, Feb 5, 2026):
- more locally relevant content from websites based in the user's country;
- less sensational content and clickbait;
- more in-depth, original and timely content from websites with expertise in a given area.
Expertise is judged "on a topic-by-topic basis". Google's example: a local news site with a dedicated gardening section can have gardening expertise, while a movie review site that wrote one gardening article likely does not. That is topical authority applied to a feed. Google said it would expand the update to all countries and languages "in the months ahead"; as of October 5, 2026 we found no announcement confirming that expansion, so check the Search Status Dashboard before blaming it for a non-US drop.
Translated into title advice, which matches Google's own guidance to "use page titles and headlines that capture the essence of the content":
| Clickbait (demoted) | Discover-friendly |
|---|---|
| "You won't believe what Google just did to SEO" | "Google's February 2026 Discover update: what changed for US publishers" |
| "This one setting is killing your traffic" | "max-image-preview:large: the one-line fix for small Discover thumbnails" |
| "SHOCKING: AI crawlers are stealing your site" | "We blocked 12 AI crawlers for 90 days. Here is what changed." |
The good versions still create curiosity, but the article delivers exactly what the headline promises. Specific numbers, named entities and first-hand results help. For the mechanics of writing titles, see Titles & Meta Tags and preview them with the SERP preview tool.
Following, Search profiles and preferred sources#
Discover is personalised, so getting people to signal interest in you is a direct lever:
- Search profiles. Google's documentation (updated September 16, 2026) says that when people follow your Search profile, your content from your website and from Instagram, TikTok, YouTube, X and Facebook becomes "more likely to appear for your audience on Google Discover". Eligible owners can add a Search profile badge to their site that links to
https://profile.google.com/@handle(Search profiles). - Follow in Discover. Trade press reported in September 2025 that Google added Follow buttons for publishers and creators in Discover, alongside more posts from X, Instagram and YouTube Shorts. Google separately removed its older Discover Follow documentation (the RSS-based feature) on November 19, 2025, so ignore guides that tell you to optimise a feed for it.
- Preferred sources are a different feature. Readers can pick your site as a preferred source so it shows more often in Top Stories, AI Overviews and AI Mode. Google's preferred sources documentation (updated September 18, 2026) does not mention Discover, and only domain or subdomain sites are eligible (Preferred sources). Promote it for Search, not as a Discover tactic. More in Digital PR & Brand Signals.
AI features in Discover and the opt-out#
Discover now mixes generative AI into the feed. In July 2025 Google began showing AI-written summaries for trending topics in Discover in the US, each citing several publishers (TechCrunch, July 15, 2025), and in September 2026 trade press reported a "Dive deeper" test that shows a short AI topic overview with links to related stories. A summary card can answer the reader without a click, just as AI Overviews do in Search.
You now have a first-party control. In Search Console, Settings > Search generative AI lets you choose Include (the default), Exclude, or Inherit from parent. It covers AI Overviews, AI Mode and generative AI features in Google Discover. Excluding removes your links and content from those features and from grounding their answers; Google says it is not used as a ranking or inclusion signal for other parts of Search, so regular Discover cards are unaffected. Most content drops out within one or two days. It does not limit AI training, which is controlled with the Google-Extended robots.txt token (Search Console Help); see AI Crawlers & Bot Access.
Measuring Discover#
The Discover performance report in Search Console shows clicks, impressions and CTR for Discover. It appears only once your property reaches a minimum number of Discover impressions. An impression is counted once per session when a card scrolls into view, and only clicks (not shares) count as clicks (Search Console Help).
The Generative AI performance report for Discover (global since August 31, 2026) shows impressions when your content appears in Discover's generative AI features. It reports impressions only, with no clicks or position (Google).
Habits that make the data useful:
- Read Discover by page and by day, not by month. Spikes are normal. Look for which topics and formats get picked up repeatedly.
- Compare Discover hits against your topic clusters. Pages that win in Discover usually sit in sections where you publish regularly and in depth.
- Annotate updates. Add Search Console annotations for Discover core updates and for big image or title changes.
- Do not rely on GA4 to split Discover from Search. Use Search Console as the source of truth for Discover clicks.
This completes the Platforms & Surfaces module. Revisit YouTube SEO and Reddit, Forums & Communities, or go to SEO for AI Mode & AI Overviews for the AI side of Search.