Google Discover Optimization: How Content Gets Picked Up

Google Discover is the feed of content Google shows people based on their interests, without a search query. Google’s documentation says content “is automatically eligible to appear in Discover if it is indexed by Google and meets Discover’s content policies”, that no special tags or structured data are required, and that being eligible “is not a guarantee of appearing” (Discover and your website). There is no submission process and no way to request inclusion. What a publisher can control is whether its content is eligible, whether its images and titles make it a good fit for an interest feed, and how it reads the performance data, which since 2026 includes a separate view for generative AI features in Discover.

How Discover differs from Search

Google advises considering Discover traffic “as supplemental to your keyword-driven search traffic”.

What Google recommends

Google’s Discover page lists practices to increase the likelihood of appearing:

• No clickbait. Avoid misleading or exaggerated details in titles, snippets or images, and avoid withholding information needed to understand the content.
• Titles that capture the essence of the content.
• No sensationalism that plays on morbid curiosity, titillation or outrage.
• Timely, well-told or uniquely insightful content. News publishers’ timely articles can also reach Top stories and the News tab, which have their own policies.
• Large, high-quality images. Google recommends images at least 1,200 pixels wide, with more than 300,000 total pixels and a 16:9 aspect ratio, enabled through the max-image-preview:large setting or AMP. Specify a representative image with schema.org markup or the og:image tag, avoid generic images such as a site logo, and avoid text-heavy images.
• A good page experience overall.

Google adds that Discover uses many of the same signals and systems as Search to identify helpful, people-first content, and that it filters out content not suited to interest feeds, such as job applications, petitions, forms, code repositories or satire without context.

Why Discover traffic changes

Google lists three reasons:

• Changing interests: if someone’s interest in a topic declines, their feed shows other content.
• Content types: Discover adjusts which types of content appear in the feed.
• Updates to Google Search: because Discover is part of Search, ranking updates can change Discover traffic.

A drop in Discover traffic is therefore not necessarily a problem with the site. It can reflect a change in audience interest or in the feed itself.

Reading the Discover performance report

Search Console’s Discover performance report shows clicks, impressions and CTR for content shown in Discover. Its help page notes (Discover performance report help):

• the report appears only after a property reaches a minimum number of Discover impressions;
• an impression is counted when the item is scrolled into view, once per result per session;
• data is aggregated by page, and metrics are assigned to the canonical URL;
• rows below an impression threshold are not shown individually but are included in totals;
• Discover traffic is “somewhat less predictable than Search traffic”.

There are no query or position dimensions because Discover has no queries.

Generative AI features in Discover

On June 3, 2026, Google announced Search Console reports for generative AI features, “including dedicated reports for Search and Discover”, and rolled them out to all websites as of August 31, 2026 (Google’s June 2026 announcement). The announcement lists impressions, pages, countries and dates for these reports, with devices available for Search results. Keep two things separate:

• The ordinary Discover report, with clicks, impressions and CTR for Discover content.
• The generative AI report for Discover, with impressions of your URLs in generative AI features within Discover.

Google states that generative AI data is also included in the overall performance reporting; the dedicated report is a separate view of it. Do not add the two together.

The Search generative AI control in Search Console also covers generative AI features in Discover. Excluding a site there removes its content from those features, not from the ordinary Discover feed.

A Discover performance interpretation checklist

• [ ] Is the content indexed and free of Discover policy issues? Check the Manual Actions report for Discover-specific violations.
• [ ] Do articles specify a large, representative image, and is max-image-preview:large allowed?
• [ ] Are titles accurate rather than sensational?
• [ ] When traffic drops, does it coincide with a Google Search update on the Search Status Dashboard, or with a topic losing interest (check Google Trends for the topics)?
• [ ] Is the drop spread across all content or concentrated in one topic or format?
• [ ] Are ordinary Discover numbers and generative AI Discover numbers kept in separate rows of any report?
• [ ] Is Discover treated as supplemental traffic in forecasts, not as a baseline?

A Discover drop traced to image previews

Picture a lifestyle publisher whose Discover clicks fall by half in October while Search traffic is stable.

• Policy check: no manual actions.
• Images: a CMS update removed max-image-preview:large from article templates; previews now show small thumbnails.
• Topics: Google Trends shows steady interest in the site’s main topics.
• Fix: restore max-image-preview:large and confirm og:image points to a large, relevant image on each article.
• Follow-up: check in URL Inspection that the robots meta tag is restored, then watch Discover clicks over the following weeks.
• Limitation: Discover remains less predictable than Search; the fix restores a condition Google recommends, it does not guarantee previous traffic.

Discover promises to avoid

Discover cannot be guaranteed, scheduled or bought through SEO work. Eligible, well-presented content can appear; whether it does depends on people’s interests and Google’s systems.