Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
Discover Is an Interest Feed, Not a Search Query
Google Discover recommends content based on a person’s interests and Web and App Activity. A page can be automatically eligible when it is indexed and follows Discover policies. There is no special schema, application, or request that places it in the feed.
That changes the content decision. Search begins with expressed demand; Discover tries to anticipate what a person may want before they type. A narrow answer to a stable keyword can perform well in search while offering little reason to interrupt a feed. A timely investigation, strong visual story, surprising dataset, or distinctive firsthand lesson may have greater Discover fit.
Eligibility is only the floor. Google explicitly says it does not guarantee appearance, and Discover traffic is less predictable than keyword-driven search. Build the program so a useful page still deserves to exist when feed distribution never arrives.
Use a Discover-Fit Editorial Test
| Question | Strong evidence | Weak substitute |
|---|---|---|
| Why now? | New development, seasonal moment, or changing behavior | A date inserted into an unchanged evergreen article |
| Why this publisher? | Firsthand access, original data, expertise, or clear analysis | A rewrite of already circulating coverage |
| Why open it? | Specific promise expressed honestly | Withheld detail, shock, outrage, or exaggerated stakes |
| Why keep reading? | Strong story, evidence, structure, and visual pacing | A thin answer stretched around advertisements |
| Why trust it? | Byline, sourcing, methods, publisher, and corrections context | Generic authority claims with no accountable creator |
Google recommends content that is timely, tells a story well, or provides unique insights. It also advises against misleading previews and sensational tactics. The title and image should earn attention by representing the actual value, not by manufacturing an information gap the article never closes.
Build the Story Around Original Value
A Discover candidate needs an editorial reason beyond topic popularity. Interview someone with direct experience. Analyze proprietary or public data with a defensible method. Photograph the process. Compare outcomes after a real implementation. Explain what changed and why it matters to the audience now.
- State the concrete reader promise before drafting.
- Identify the evidence only this team can contribute.
- Verify names, dates, numbers, captions, and source context.
- Pair every strong claim with visible support or clear attribution.
- Give the reader a satisfying conclusion rather than a teaser loop.
Clear authorship and publication information matter. Discover’s policies call for dates, bylines, information about authors and the publisher, company or network context, and contact information for news sources. Sponsored relationships should be plainly disclosed; paid material should not masquerade as independent editorial work.
The content-operations system turns those requirements into source, verification, approval, and update gates instead of relying on a last-minute quality check.
Treat the Lead Image as Editorial Packaging
Google recommends compelling, relevant, high-quality images for Discover, especially large images. Its current specifications call for at least 1200 pixels of width, more than 300,000 total pixels, and a composition that works in a 16:9 crop. The page must permit large previews through max-image-preview:large or use AMP.
Specify the preferred representative image through appropriate schema.org markup or og:image. Avoid a generic logo, decorative stock image, or text-heavy graphic. A feed thumbnail is small and context-light; one clear subject, strong contrast, and safe crop space usually communicate better than a collage.
| Image check | Release standard |
|---|---|
| Relevance | Represents the specific story rather than the brand |
| Resolution | At least 1200px wide and more than 300,000 pixels |
| Crop | Main subject survives a landscape 16:9 presentation |
| Selection | Consistent preferred URL in page metadata |
| Delivery | Crawlable, fast, stable, responsive, and dimensioned |
The image SEO guide covers responsive delivery, crawl access, filenames, alt text, image sitemaps, and performance. Do not trade a compelling thumbnail for a slow or unstable page.
Make the Preview and Page Agree
A feed card creates an expectation through its headline, image, and source. The opening screen should confirm that promise immediately. Put the key subject and evidence near the top, make the byline and date clear, and avoid an intrusive interstitial or ad stack between the tap and the story.
Google advises titles that capture the essence of the content and warns against misleading or exaggerated preview details, sensationalism, and withholding crucial information. The title and description guide explains how visible headings, title elements, and other signals can influence Google’s generated presentation.
Use accurate Article or NewsArticle markup when it matches the page, but do not invent dates, authors, or images for eligibility. Markup can describe the page; it cannot turn a weak or noncompliant story into a recommended one.
Protect the Mobile Reading Experience
Discover is consumed as a feed, so the transition into the article must be fast, stable, and readable on a phone. Reserve image and ad dimensions, keep the main visual discoverable early, reduce heavy scripts, and ensure the first interaction is responsive.
Page experience is broader than one speed score. Google asks whether pages have good Core Web Vitals, use secure delivery, display well on mobile, avoid excessive ads and intrusive interstitials, and make the main content easy to distinguish. The Core Web Vitals playbook gives a template-level diagnosis for LCP, INP, and CLS.
Measure Discover as a Volatile Portfolio
Search Console shows a Discover Performance report only after the property reaches a minimum impression threshold. It includes impressions, clicks, and CTR, can group by page, country, appearance, and date, and credits data to canonical URLs. Google’s Discover documentation says the report can cover the last 16 months.
Do not benchmark Discover like steady query traffic. A topic can disappear from audience interest, Google can adjust which content types appear, and Search updates or reporting anomalies can create sharp changes unrelated to one publishing decision. Review four-week and longer portfolios rather than treating each daily spike as a repeatable acquisition channel.
- Track eligible published stories by format and topic.
- Record impressions, CTR, clicks, engaged visits, and outcomes.
- Compare original evidence and image treatments, not only headlines.
- Check manual actions and official data anomalies before diagnosing a drop.
- Keep keyword-driven search and direct audience growth as the baseline.
Add Discover to the review cadence in the Search Console operator’s guide, but keep it separate from Web Search. The surfaces have different demand mechanics and volatility. Publishers should also isolate the dedicated reporting and technical requirements described in the Google News field note.
Questions About Google Discover SEO
- How do I submit a page to Google Discover?
There is no Discover submission process. Content is automatically eligible when Google has indexed it and it follows Discover content policies. No special tag or structured data is required, and eligibility never guarantees an appearance.
- Does content have to be brand new to appear?
No. Google says older content may appear when it remains helpful and relevant to someone’s interests. Timely stories are a strong fit for an interest feed, but original evergreen explainers, analysis, and videos can also surface. Update dates only when the underlying content changes substantially.
- Why did Discover traffic suddenly disappear?
Discover is less predictable than keyword-driven search. Audience interests, content types, Search updates, reporting issues, and Google’s feed systems can change traffic even when publishing quality is steady. Check manual actions and data anomalies, then evaluate longer trends rather than rewriting pages after one drop.
- Do I need Article schema for Discover?
No special markup is required for Discover eligibility. Accurate Article markup can help Google understand an article and can specify representative images, but it does not unlock or guarantee Discover. Visible authorship, dates, publisher information, and page content must remain accurate.
Primary Sources Reviewed
- Google Search Central: Discover and your website
- Google Search Help: Discover content policies
- Google Search Central: Helpful, reliable, people-first content
- Google Search Central: max-image-preview specification
- Google Search Console: Discover Performance report
Reviewed July 28, 2026. Discover interfaces, policies, reporting, and distribution systems change frequently; verify current requirements and treat feed traffic as supplemental.
