How to prepare article images for Google Discover and image search

A publisher can spend hours polishing an article and then upload a tiny, generic image. The page may still rank, but it gives visual discovery systems less useful material to show. Google Discover and Google Images are not the same product, yet both rely on accessible, meaningful images associated with credible pages. The practical goal is eligibility and clarity, not a formula promising impressions.

Google's Discover guidance says indexed content that meets Discover policies is automatically eligible without special tags or schema, but eligibility is not a guarantee of being featured. It recommends large, high-quality images and gives concrete specifications. Treat those as publication checks, not ranking factors with invented weights.

Choose an image that explains the article

Use a genuine photograph you have rights to, an accurately licensed illustration or a clearly conceptual original. The image should relate to the topic and remain legible at reduced card sizes. Avoid generic thumbnails, misleading composites or a heavy text poster that duplicates the headline. For a real event, do not use a synthetic picture as though it documented the scene.

Keep the important action away from extreme edges because crops can change. If an article concerns a product interface, use an official screenshot only when rights and context are appropriate; label a conceptual representation honestly. Maintain an internal record of source, license, edits and credit obligations.

Meet the documented size and preview requirements

Google recommends Discover images at least 1200 pixels wide, higher than 300,000 total pixels, with suitable landscape composition. Its documentation discusses a 16:9 crop and recommends enabling max-image-preview:large or using AMP, as applicable. An example size of 1600 by 900 is comfortably above the listed thresholds, but dimensions alone do not produce Discover placement.

Specify an appropriate representative image through og:image or supported structured data, as the Discover help page describes. Make sure the referenced URL serves the intended image and can be fetched without authentication or accidental bot blocking. Verify that responsive versions do not quietly replace the large image with a tiny thumbnail in the metadata.

Help search engines discover the file

The Google image SEO guide emphasizes discoverability and landing-page context. Use descriptive filenames where practical, concise accurate alternative text, sensible captions when they add information and nearby prose that explains the image. Do not stuff the alt attribute with keyword variants or use it as an invisible promotional paragraph.

Prefer standard image elements and crawlable URLs over images that exist only as CSS backgrounds or inaccessible scripts when the image carries meaning. Keep appropriate HTTP responses, avoid blocked paths and confirm that image files are not omitted by overaggressive CDN settings. Compress efficiently while preserving subject clarity.

Compare representation rather than obsessing over one ranking

Inspect how an article looks in social previews, browser cards and mobile crops. Verify the largest intended image is selected and that credits are present where required. Use Search Console's Discover report when your property has sufficient data to receive it, but do not treat the absence of that report as proof of a technical violation.

Consider an article whose main image is a logo with a line of text. Replacing it with a clear subject-specific visual may improve readers' understanding even if Discover traffic never changes. That editorial benefit is real and independent of any search algorithm claim.

Prepare an image record alongside each article: a stable file name, rights owner, license terms, crop-safe master, intended displayed image URL, alt text and any attribution obligation. This prevents a future editor from using a licensed crop in an unintended context. It also makes corrections easier when a previously published graphic turns out to contain an error.

Consider the difference between a hero image and a logo. A brand mark may identify the publication, but it tells readers little about the subject of a specific story. A conceptual diagram can help explain a technical guide, while a verified photograph may be appropriate for a reported event. Which one works depends on the article's purpose and rights, not a blanket rule that photographs always outperform illustrations.

To audit image delivery, inspect the actual page source and response after deployment. Confirm that the image URL points to the expected dimensions, returns a successful response, and is not blocked by authentication or robot restrictions. If the same article exists in six languages, the shared image can be reused while the surrounding caption and alternative text are localized accurately. Do not change the factual representation of the image between languages. Finally, allow time for search systems to recrawl changed files rather than expecting immediate visual updates everywhere.

Create a repeatable image checklist

Before publishing, confirm rights, factual accuracy, file dimensions, accessible URL, correct og:image, descriptive alt text and crop safety. After publishing, recheck the article publicly, including cached and localized variants. If an image changes materially, confirm that all versions reflect the correction.

This checklist complements our faceted crawling guide and traffic investigation guide. Both deal with discoverability from different angles. No single featured image can fix weak reporting, a blocked page or a misleading headline. A good one gives solid work a fair opportunity to be found.