Image SEO Checklist

Use this for: Pages containing images.

Inspect the page and identify whether the image explains something, shows a real product, performs an action or provides decoration. Keep an image when it supports a useful purpose. Choose a clearer diagram, photo or example if the current visual is generic or misleading. Do not present a generated illustration as a photograph of an actual customer or measured result.

1. Identify the image’s role on this page.
2. Check that its content supports the nearby answer.
3. Replace misleading or unhelpful visuals.

Use this for: Images sourced or created for publication.

Check where the image came from and whether your intended use is allowed. Keep the applicable creator, license or permission details with the project. Follow actual credit requirements and remove unnecessary private information from the delivered file. Do not assume an image found through search is free to reuse. Get specialist advice if rights are unclear.

1. Identify the image’s source.
2. Confirm permitted use and any credit requirements.
3. Keep accurate rights information with the asset.

Use this for: New image files or justified renaming work.

Before uploading a new asset, name it for its actual subject using ordinary words and the correct extension. Use a consistent separator convention such as hyphens. A filename is a light clue, not a keyword quota. Avoid renaming an established image address solely for cosmetics unless references and any necessary transition are handled.

1. Choose a descriptive name before upload.
2. Keep the correct format extension.
3. Preserve existing references when changing an established file.

Use this for: Informative images in page content.

In the page editor, put the image near the relevant heading and paragraph. Add a caption when it supplies useful context, a source credit or an important qualification. Do not repeat the same keyword sentence in the heading, caption and alternative text merely for SEO. Keep essential instructions available as text.

1. Identify the paragraph the image explains.
2. Place the image and useful caption near it.
3. Keep essential meaning available in page text.

Use this for: Informative images.

Open the image’s alternative-text field in the editor. State the important information someone misses if they cannot see it. Use a relevant subject term naturally, without a list of keywords or unnecessary “image of” wording. Avoid repeating a nearby caption when that repetition adds no useful meaning. There is no universal alt-text word count.

1. Identify the image’s essential meaning.
2. Write a concise contextual alternative.
3. Read it with the surrounding paragraph.

Use this for: Purely decorative images.

If the image adds no information beyond nearby text and serves no independent control function, use the editor’s decorative option or an empty alt attribute. Do not omit the attribute accidentally when the platform expects it. A linked icon may still need an accessible name, so decide from the full link/control context rather than appearance alone.

1. Decide whether the image is purely decorative.
2. Use the appropriate empty-alternative treatment.
3. Check that any surrounding control still has a name.

Use this for: Functional or linked images.

For an image that is the only content of a link or button, ensure its accessible name explains the destination or action. Use the relevant alternative-text or control-label setting. If adjacent link text already supplies the same name, avoid announcing it twice. Test the control with the image unavailable or with assistive technology when possible.

1. Identify image-only links and controls.
2. Provide the action or destination name.
3. Check that the full control remains understandable.

Use this for: Complex explanatory images or charts.

Use short alt text to identify a complex image and put its important details in nearby prose, a table or a full description. Do not squeeze an entire chart into an unreadable keyword-heavy alternative. Explain the takeaway and provide the values or steps a reader needs. Label hypothetical data clearly and do not invent a measured result.

1. Identify the information needed beyond a short label.
2. Add a nearby text or table equivalent.
3. Check that the reader can complete the task without seeing the image.

Use this for: Uploaded raster images.

Check the image’s approximate rendered dimensions in the layout, then resize the source in an image editor or the platform’s media pipeline. Allow appropriate resolution for higher-density displays and responsive layouts. Compare at the intended size. Do not upload a huge original without reason or force every diagram into a small blurry thumbnail.

1. Check the layout’s intended display size.
2. Prepare a suitable source or responsive versions.
3. Inspect detail at normal viewing size.

Use this for: Images being prepared or optimized.

Export a copy using a supported format and reasonable compression, then compare file size, sharpness, text and transparency with the original. Modern formats such as WebP or AVIF can be useful where the workflow supports them; keep required PNG or other format policies when applicable. No single quality percentage or kilobyte target fits every asset.

1. Export a smaller candidate copy.
2. Compare file size and visible quality.
3. Upload the suitable supported version.

Use this for: Sites with responsive image delivery.

Open the page at different widths and inspect the image request in browser Network tools. Ask the developer to review srcset and sizes or the platform’s image component if a phone downloads a needlessly large desktop asset. Responsive layout and responsive file selection are separate: an image can look small while still transferring a large original.

1. Compare small and large viewport requests.
2. Inspect the selected file and dimensions.
3. Correct responsive image configuration when needed.

Use this for: Images affecting page layout.

Check that the image template provides correct intrinsic width and height or an appropriate CSS aspect ratio. Reload the page and watch content around the image as it loads. If the reserved shape does not match the delivered crop, correct the template rather than adding an arbitrary fixed height that clips the image on another screen.

1. Inspect the image’s reserved dimensions.
2. Reload and observe nearby content.
3. Correct the dimensions or aspect ratio.

Use this for: Templates controlling image loading.

Identify whether the image is a main above-the-fold or LCP element using performance diagnostics. Do not lazy-load that important image indiscriminately. Lower-priority images can use suitable lazy loading, but content needed for search must load without requiring a special click or other unsupported interaction. Test the actual output rather than enabling a global setting blindly.

1. Identify the important initial image.
2. Review loading behavior by role.
3. Retest loading and rendered availability.

Use this for: Images intended for public discovery.

Inspect the published HTML for a normal img element with a valid src, or have the developer confirm equivalent crawlable output from the platform. Google does not index CSS background images as standard image content. Open the actual image URL and review relevant crawl restrictions when an important asset cannot be fetched. Private media should remain appropriately protected.

1. Inspect the rendered content-image markup.
2. Open the delivered file.
3. Investigate unintended blocks or missing output.

Use this for: Pages with missing or broken images.

Open DevTools → Network, reload the page and select the failing image request. Check its status, final address and response type, then compare the source reference with the intended file. A missing file, wrong path, access restriction and HTML error returned as an image need different fixes. Also check transformed thumbnails, not just the original asset.

1. Reload with Network open.
2. Inspect the failed image request.
3. Correct the specific path, access or transformation problem.

Use this for: Sites with an identified image-discovery need.

If important public images are difficult to discover through normal page crawling, review whether the platform can include them in an image sitemap. Associate the actual image URLs with their relevant pages and check fetch access, including a separate image host when used. A sitemap is a discovery aid, not a fix for a broken or blocked asset.

1. Identify a genuine discovery need.
2. Generate accurate image/page entries when useful.
3. Check the listed public image URLs.

Use this for: Images with relevant credit or licensing information.

When image licensing information is relevant, provide accurate creator, credit and license details through the supported visible fields and structured or IPTC metadata. Check the delivered file or page output because processing can remove embedded fields. Never invent a creator or add location metadata as a ranking trick. Metadata eligibility does not guarantee a search feature.

1. Confirm the actual credit or licensing facts.
2. Add supported relevant metadata and visible credit.
3. Check what remains in the delivered output.

Use this for: Sites with available image-search data.

In Search Console Performance, choose the Image search type and review relevant page/query data using comparable periods. Keep the filter with your observations. Image impressions and clicks answer a different question from page speed or alternative-text quality. A change in demand or results can affect the numbers, so do not attribute every movement to one filename edit.

1. Select Image search type.
2. Review the relevant pages and comparable periods.
3. Combine search observations with actual image-quality checks.