Google Discover SEO: how pages get into the feed, the image rules, and how to read the traffic

Discover SEO is the work of getting your pages shown in Google Discover, the feed of articles and videos that Google shows people, mostly on their phones, based on their interests rather than on a search they typed. There is no special tag to apply for it. According to Google's own documentation, any page that is indexed by Google and meets Discover's content policies is automatically eligible. What moves you from eligible to actually shown is the same thing that wins in search, helpful content, plus three things Discover cares about more than search does: a large, relevant image, an honest headline, and timeliness.

This page walks through how Discover picks content, the exact image specification Google publishes, the habits that get a site filtered out, and how to measure what Discover sends you.

How Discover differs from normal search

In normal search somebody types a query and Google ranks pages against it. In Discover nobody types anything. Google builds a feed for each person from their Web and App Activity, so the question is not "which page best answers this query" but "which page would this person want to read right now".

That one difference explains most of the advice:

• There are no keywords to target. You cannot pick a Discover keyword in a research tool. You pick a topic your audience follows and publish well on it.
• The preview does the selling. A person sees a headline and an image in a scrolling feed. If either is weak, they scroll past.
• Timing matters. Google says it favours content that is timely for current interests, tells a story well, or gives unique insight. Older content can still appear when it is helpful and relevant to someone's interests.

Google also states that Discover uses many of the same signals and systems as Search to decide what is helpful, people first content. So a site that struggles in search will rarely do well in Discover. Fix the basics first.

The five things to get right on a page

1. Be indexed and inside the content policies

Discover only draws on indexed pages. Check a page is indexed with the URL Inspection tool in Search Console before you worry about anything else. Google also runs separate Discover content policies. If a site breaks them, a Discover manual action can appear in Search Console under Security and Manual actions, and that is the first place to look when Discover traffic stops entirely.

2. Use a large image, and allow Google to show it large

This is the most specific advice Google gives, and the one most often missed. Google recommends images that are:

• at least 1200 pixels wide
• more than 300,000 total pixels (a 1280 by 720 image has 921,600, which passes)
• in a 16 by 9 aspect ratio

Large images also need permission. Google lists two ways to enable them: the max-image-preview:large robots setting, or AMP. The robots setting is a single meta tag:

<meta name="robots" content="max-image-preview:large">

Then tell Google which image you want, using schema.org markup or the og:image meta tag. Our guide to Open Graph tags covers how to set that tag correctly. Google asks you to avoid your site logo and text heavy images in that slot. If you crop images yourself, crop for landscape and keep the important detail inside the 16 by 9 frame, because Google will otherwise crop automatically.

3. Write a headline that describes the article

Google's documentation is direct here. Use titles that capture the essence of the content. Avoid clickbait, which it describes as misleading or exaggerated detail in the title, snippet or image, or holding back the information a reader needs to understand what the page is about. Avoid sensationalism that plays on morbid curiosity, titillation or outrage. A headline that tells the truth about the article and still makes someone curious is the target.

4. Publish the kind of content the feed is built for

Discover is designed for interest based feeds, such as articles and videos. Google says it may not recommend job applications, petitions, forms, code repositories, or satire without context, and that it filters content that might be shocking or unexpected on top of SafeSearch. Google lists sports, health, entertainment and lifestyle among the content it regularly shows. A product page or a contact form is not going to appear. A well reported article or a useful explainer can.

5. Give a good page experience

Google points to its page experience guidance. In practice that means a page that loads quickly on a phone, does not jump around while loading, and does not bury the article under pop ups. Most people meet Discover on a phone, so test there.

Discover SEO checklist

Check	Where to verify it	What good looks like
Page is indexed	URL Inspection in Search Console	"URL is on Google"
No Discover manual action	Search Console, Security and Manual actions	No issues detected
Large image allowed	Page source, robots meta tag	max-image-preview:large present
Preferred image named	og:image or schema.org image	Relevant photo, 1200px or wider, 16 by 9, not a logo
Headline	Read it beside the article	Describes the article, no exaggeration
Mobile experience	A real phone	Fast, stable, article visible without fighting pop ups

Measuring Discover traffic

Search Console has a separate Performance report for Discover. It shows impressions, clicks and click through rate for any of your content that appeared in Discover in the last 16 months, once your data passes a minimum threshold of impressions. If you do not see the report at all, it usually means you have not reached that threshold yet. Google notes that the report includes traffic from Chrome and covers every surface where people use Discover. Our Search Console guide explains how the other reports in the tool measure their numbers.

When you read the report, look at it by page rather than in total. Discover tends to send a burst of visits to a handful of articles and very little to the rest. The pages that got picked up tell you which topics and which kinds of image your audience responds to.

Why Discover traffic goes up and down

Google is unusually open about this. It says Discover traffic is less predictable than keyword driven search and that you should treat it as supplemental to search traffic. It gives three reasons traffic moves:

• Interests change. When people stop searching for a topic, their feed moves on, and so does your traffic.
• Content types shift. Google keeps adjusting which kinds of content appear in the feed.
• Search updates. Discover is an extension of Search, so core updates can move Discover traffic too.

Google also says sites can see changes that are unrelated to the quality or publishing frequency of their content. So plan your business around search and other steady channels, and count Discover as the welcome extra it is.

Questions about Google Discover SEO

Do I need structured data to get into Discover?

No. Google says no special tags or structured data are required. Indexed content that meets the content policies is eligible. Markup helps only in naming your preferred image.

Can a small site or a business blog appear in Discover?

Yes. There is no size requirement in Google's documentation. What matters is whether the content suits an interest feed. A business blog that publishes useful, timely articles with strong images can be shown. Service pages and sales pages generally are not.

Does publishing more often get me more Discover traffic?

Not on its own. Google says Discover changes can be unrelated to publishing frequency. A few strong, timely articles with good images do more than many thin ones.

Why does my Discover report not appear in Search Console?

The report only shows once your content has reached a minimum number of Discover impressions. Until then there is nothing to show.

Is Google Discover the same as Google News?

No. Google News is its own product built around news. Discover is a personal feed inside Google Search that can show news but also evergreen articles, lifestyle content and videos matched to a person's interests.

The feed Discover does not cover

Discover is one example of a bigger change: people are given answers without typing a search. The same thing now happens when somebody asks ChatGPT, Gemini or Perplexity a question about your field. The assistant writes a short answer and names a few businesses or sources, and there is no report in Search Console that shows you whether yours was one of them. If your articles are good enough to be picked for Discover, it is worth asking those assistants the questions your readers ask and seeing whose name comes back. AI Knows Us is built to measure that answer and work on it.