Google preferred image SEO is the work of giving Google clear, consistent signals about the image that best represents a page. Google says its preview selection is automated, but publishers can influence it through schema.org markup or the og:image meta tag.
A preferred image does not guarantee a thumbnail in Search or Discover. It does reduce ambiguity when a page contains logos, author photos, icons, screenshots and several editorial images. This audit explains what to check across content, metadata, crawlability and preview controls.
Understand the preferred-image signals
Google’s image SEO best practices list two metadata routes. A page can specify primaryImageOfPage on a WebPage, or attach an image URL or ImageObject to the main entity, such as a BlogPosting. A page can also specify <meta property="og:image">.
These are inputs to an automated choice, not commands. Google may use other sources and select another image. The safest implementation makes the preferred image genuinely representative and keeps the structured data, Open Graph tag and visible page aligned.
Do not confuse a preferred editorial image with a favicon. A favicon represents the website beside a result, while a result thumbnail represents the individual page. Kayaar’s favicon SEO audit covers that separate asset.
Build the audit inventory
Choose representative templates: homepage, service page, category, article, product, author profile and any page type that receives organic traffic. Record the visible primary image, its URL, file type, dimensions, byte size, alt text, response code and whether it appears in the initial HTML.
Then capture every metadata candidate: schema.org image fields, primaryImageOfPage, og:image, social-card fields and theme defaults. Conflicting plugins often produce several tags or use a fallback logo when a featured image is missing.
Audit check
Preferred state
Typical problem
Visible content
The preferred image appears prominently and supports the page topic.
Metadata points to an unrelated or hidden asset.
Schema.org
One clear page or main-entity image relationship uses a crawlable URL.
Several entities claim different primary images.
Open Graph
og:image matches the intended editorial image.
A plugin falls back to the site logo.
Technical access
Page and image return successful responses and allow crawling.
Robots rules, hotlink protection or authentication blocks access.
Important detail sits near an edge or text becomes unreadable.
Choose a representative image
The selected image should explain the page at a glance. Avoid a generic logo, stock image with no connection to the topic, decorative background or image dominated by text. Google specifically recommends relevant, representative and high-resolution images while warning against generic, text-heavy and extremely proportioned assets.
Use descriptive filenames and useful alt text, but keep their jobs separate. A filename offers a light subject clue. Alt text describes the image for accessibility and understanding; it should not be a keyword list or a repeated headline when that does not describe the visual.
For pages with original diagrams, screenshots or photography, document ownership and permitted use. Do not select a large image merely because it looks attractive if the business cannot lawfully publish it.
Align schema and Open Graph metadata
Inspect the rendered HTML and original source. Confirm that the structured-data image URL and og:image resolve to the same intended asset unless a documented platform need requires a different version. Use absolute HTTPS URLs and avoid temporary query strings that expire.
For articles, check that the image belongs to the Article or BlogPosting entity associated with the page. Kayaar’s Article schema audit shows how main entity, author and page relationships should remain coherent.
If the CMS creates several sizes, choose a stable source with enough resolution and consistent accessibility. Do not declare a tiny thumbnail while displaying a larger editorial image. Test templates with and without a manually selected featured image so fallbacks remain intentional.
Check crawlability and delivery
Google recommends standard HTML image elements and can discover images in the src attribute of an img element. It does not index CSS background images in the same way. Ensure the preferred asset is in crawlable HTML and not available only after a user interaction.
Request the image URL without cookies. Verify a successful response, correct image content type, stable redirect behaviour and no blocking by robots rules, firewall, CDN protection or login. Keep the same image URL when the same asset is reused so crawlers can cache it efficiently.
Review loading performance as well. Compression, responsive sources and explicit dimensions can protect page experience without reducing the source below useful display quality. Kayaar’s website SEO audit provides a wider technical baseline.
Prepare the image for Discover
Google’s Discover guidance recommends large, high-quality images at least 1,200 pixels wide and more than 300,000 total pixels. Google tries to crop images for 16:9 use, so important subjects need safe spacing for a landscape presentation.
Large previews must be allowed with max-image-preview:large or AMP. Confirm that another robots meta tag or HTTP header does not impose a smaller preview rule. Eligibility does not guarantee Discover traffic, and Discover traffic is less predictable than keyword-driven Search visits.
Preview the crop on mobile and desktop. A wide image can still fail if the subject is too small, faces are cut off or overlay text becomes illegible. Use the metadata image as a clean representation, not a clickbait panel.
Validate and release safely
Test structured data syntax, inspect the live page and compare the page source with the rendered DOM. Check the selected image in social-preview debuggers only as a supporting test; Google’s final choice is independent.
Deploy to a small template group first. Crawl the pages, confirm that image URLs remain accessible and inspect priority URLs in Search Console after recrawling. Kayaar’s multimodal Search reporting guide can help connect image discoverability with broader visual-search performance.
Monitor thumbnail changes, image impressions, page clicks and page speed without claiming that metadata alone caused movement. For a template-level image and metadata review, contact Kayaar.
FAQs
What is a preferred image in Google Search?
It is the image a publisher indicates as the best page representation through schema.org metadata or the og:image tag.
Does preferred-image metadata guarantee a thumbnail?
No. Google’s image selection is automated, and metadata only influences which image may be chosen.
Should schema and og:image use the same image?
Usually yes. Consistent signals reduce ambiguity unless a documented platform requirement justifies a separate version.
How large should a Google Discover image be?
Google recommends at least 1,200 pixels wide, more than 300,000 total pixels and a composition suitable for 16:9 cropping.
Is alt text a preferred-image signal?
Alt text helps accessibility and image understanding, while schema.org and og:image are the documented preferred-image metadata sources.
