Image SEO: The Part of SEO Most Teams Get Half Right

Image SEO is the practice of optimising visual assets so search engines can index them, understand what they show, and surface them in both image search and standard organic results. Done properly, it improves page load speed, boosts accessibility, adds keyword relevance signals, and opens up an additional traffic channel that most competitors have left half-optimised.

Most teams treat image optimisation as a checklist item: add an alt tag, compress the file, move on. That approach captures maybe 40% of the available value. The rest sits in file naming, structured data, page context, and the relationship between your images and the content around them.

Why Image SEO Gets Underinvested

When I was running an agency that had grown from around 20 people to over 100, we did a content audit for a mid-market e-commerce client who was convinced their SEO had plateaued. They were spending a reasonable budget on content and link acquisition. Their on-page fundamentals were solid. But their organic traffic had been flat for eight months.

We pulled a crawl. They had over 4,000 product images. Fewer than 200 had descriptive alt text. File names were almost universally strings of numbers and letters from the product management system, something like IMG_8843291.jpg. Page weight on product pages was running at an average of 3.8MB. None of it was in a sitemap.

The fix was not glamorous. It was a systematic pass through their image library, a change to their CMS image upload workflow, and a sitemap submission. Within three months, image search impressions had increased substantially and page speed scores had improved enough to move them into a better performance bracket. The traffic plateau broke.

Image SEO underperforms because it sits in the gap between content teams and technical teams. Content writers do not always know how to name files or write alt text correctly. Developers optimise for build speed, not SEO. Nobody owns the problem end-to-end, so it gets done partially by everyone and completely by nobody.

If you are building a complete SEO strategy, image optimisation belongs in the technical layer, not the content layer, and it needs someone accountable for it. The Complete SEO Strategy hub covers the full architecture of how these layers connect, from technical foundations through to content and authority building.

What Google Actually Does With Your Images

Google uses a combination of signals to understand an image: the file name, the alt attribute, the surrounding text, the page title, structured data, and increasingly its own machine vision to interpret the content of the image itself. The machine vision part is impressive but it does not replace the text signals. Google still weights what you tell it alongside what it can see.

The crawl sequence matters. Googlebot typically discovers images through the HTML of a page or through an image sitemap. It requests the image file, reads the HTTP headers (which include content type and caching instructions), and then processes the image in conjunction with the page context. Images that are lazy-loaded via JavaScript can cause problems if the implementation is poor, because Googlebot may not execute the JavaScript correctly and the image may never be indexed.

Google also uses image data as a quality signal for the page itself. A page with high-quality, relevant images that load quickly is treated differently from a page with broken image references or images that are technically present but inaccessible. This is one of the reasons image SEO affects overall page performance, not just image search visibility.

File Names: The Signal Nobody Bothers With

The file name is one of the first pieces of information Google has about an image. It reads it before it reads the alt text, before it reads the surrounding content. A file named red-leather-chelsea-boot-womens.jpg tells Google something meaningful. A file named DSC00291.jpg tells it nothing.

The rules are straightforward. Use hyphens to separate words, not underscores. Google treats hyphens as word separators and underscores as word joiners, so womens-chelsea-boot is three words and womens_chelsea_boot is one. Use lowercase throughout. Keep the name descriptive but not padded: red-chelsea-boot-womens is good, red-leather-stylish-womens-chelsea-boot-buy-now is not.

For teams working at scale, this means the file naming convention needs to be built into the upload workflow, not left to individual judgement. If your CMS allows you to rename files on upload, build a naming convention into your editorial guidelines. If files come from a product database with auto-generated names, the fix sits in the export configuration, not in manual renaming.

Alt Text: Accessibility First, SEO Second

Alt text exists primarily for accessibility. Screen readers use it to describe images to visually impaired users. That framing matters because it tells you exactly how to write it: describe what is in the image clearly and specifically, as if you were describing it to someone who cannot see it.

The SEO benefit is real but it is a consequence of doing the accessibility job well, not a separate task. An alt text that says “red leather womens chelsea boot with block heel on white background” is good accessibility and good SEO. An alt text that says “chelsea boot womens boots leather boots buy chelsea boots” is bad accessibility and bad SEO, and it will be treated as keyword stuffing.

A few specific rules. Decorative images (dividers, background textures, abstract design elements) should have empty alt attributes (alt=””) so screen readers skip them. Do not write “image of” or “photo of” at the start, screen readers already announce that it is an image. Keep alt text under 125 characters where possible. If the image contains text, include that text in the alt attribute.

One thing I have seen cause problems in larger organisations: alt text gets treated as a copywriting task and gets written by people who do not understand the accessibility context. The result is alt text that describes the marketing intent of the image rather than its visual content. “Our award-winning product range” is not an alt text. “Flat lay of five skincare products on a marble surface” is.

Image Compression and Format: The Performance Layer

Page speed is a ranking factor for Google, and images are the dominant contributor to page weight on most content and e-commerce sites. This is not a marginal issue. A product page that loads in 2 seconds and a product page that loads in 6 seconds are not equivalent from a ranking or conversion standpoint.

The format question has become clearer in recent years. WebP delivers smaller file sizes than JPEG or PNG at comparable quality, and it is now supported across all major browsers. AVIF is a newer format with even better compression but adoption in CMSs is still catching up. For most teams right now, WebP is the right default for photographs and complex images, and SVG is the right choice for logos, icons, and simple graphics.

Compression needs to be automated, not manual. No team is going to manually compress every image before upload at any kind of scale. The right approach is to build compression into the CMS pipeline: either through a plugin (if you are on WordPress, there are several reliable options), through a CDN that applies compression at the edge, or through an image optimisation service that sits between your storage and your delivery layer.

Serving images at the correct display size is a separate issue from compression. An image that is 2400 pixels wide but displayed at 600 pixels is carrying three times the weight it needs to. Responsive images (using the srcset attribute) allow you to serve different image sizes to different devices. This is a development task but it has a direct impact on both page speed scores and Core Web Vitals, specifically Largest Contentful Paint, which is one of the metrics Google uses as a ranking signal.

Structured Data for Images

Structured data does not directly cause images to rank higher, but it gives Google explicit context that can lead to enhanced search appearances. For certain content types, this matters a great deal.

Product pages with ImageObject markup and the correct image properties are more likely to appear in Google Shopping surfaces and product rich results. Recipe pages with image markup are eligible for the visual rich results that dominate the top of recipe searches. Article pages with the image property in their Article schema are more likely to have their images pulled into Top Stories and Discover.

The ImageObject schema properties worth knowing: url (the image URL), width and height (in pixels), caption (a human-readable description), and representativeOfPage (a boolean that tells Google this is the primary image for the page). These are not complex to implement but they are often missing from schema setups that were done quickly or templated without thought.

For e-commerce teams in particular, I would prioritise product image schema alongside price and availability markup. The combination is what drives the rich result appearances that generate significantly higher click-through rates than standard blue links.

Image Sitemaps: Telling Google What Exists

An image sitemap is an extension of your standard XML sitemap that explicitly lists the images on each page. It is particularly valuable when images are loaded via JavaScript (where Googlebot may not discover them through normal crawling) or when you have a large image library that you want indexed quickly.

The format is an extension of the standard sitemap protocol. Each URL entry in your sitemap can include one or more image:image elements, each containing the image location, caption, title, and licence URL if applicable. Most major SEO plugins handle this automatically if configured correctly.

One practical note: Google does not guarantee it will index every image in your sitemap. The sitemap is a request, not a command. But for large sites with significant image libraries, submitting an image sitemap through Google Search Console is the most reliable way to ensure comprehensive coverage and to identify indexing issues when they arise.

Lazy Loading: Get It Right or Pay the Price

Lazy loading defers the loading of images that are below the fold until the user scrolls toward them. It is a legitimate performance technique that reduces initial page load time. But it is also one of the most common sources of image indexing problems.

The issue is that Googlebot’s rendering of JavaScript is not identical to a browser’s rendering. If lazy loading is implemented in a way that requires user interaction to trigger, or if it relies on scroll events that Googlebot does not replicate, the images may never be seen by the crawler and therefore never indexed.

The safe implementation uses the native HTML loading=”lazy” attribute on the img tag. This is a browser-native feature that Googlebot handles correctly. JavaScript-based lazy loading libraries are riskier and should be tested with Google’s URL Inspection tool in Search Console to confirm that images are being rendered and indexed as expected.

For above-the-fold images, particularly hero images and the primary product image on e-commerce pages, do not lazy load at all. These images should load immediately. The Largest Contentful Paint metric is often determined by the primary above-the-fold image, and lazy loading it will hurt your Core Web Vitals score.

Google Image Search as a Traffic Channel

Image search is a real acquisition channel that most SEO strategies treat as incidental. For certain industries, it is genuinely significant: e-commerce, food and recipe publishing, interior design, fashion, travel, and any category where the visual quality of a product or outcome is part of the purchase decision.

The traffic from image search behaves differently from standard organic traffic. Users who click through from an image result are often in an earlier stage of consideration. They saw something visually appealing and clicked to find out more. The page they land on needs to immediately deliver context and a clear path forward, whether that is a product page, a recipe, or an article.

To optimise for image search specifically: ensure images are high quality and genuinely useful (Google’s image ranking considers the quality and relevance of the image itself, not just the surrounding text signals), include the image in an image sitemap, use descriptive alt text and file names, and ensure the landing page is fast and contextually relevant to what the image shows.

One thing worth monitoring: Google Search Console separates image search impressions and clicks from standard web search. If you are not already looking at this data, filter by search type in the Performance report. You may find you are generating significant image impressions with very low click-through rates, which often points to a mismatch between the image and the landing page experience.

Original Images vs. Stock: The Quality Signal

Google has been explicit that it values original, high-quality images. Stock photography is not penalised directly, but it does not add anything distinctive to a page. If the same stock image appears on 300 other websites, it carries no uniqueness signal and contributes nothing to your content differentiation.

For content marketing and editorial SEO, original images, whether photographs, custom illustrations, data visualisations, or screenshots, are worth the investment. They are more likely to be linked to by other sites (because they cannot get the same image elsewhere), more likely to appear in image search for competitive terms, and they signal to Google that the page contains genuinely unique content.

I have seen this play out in content audits more than once. Pages with custom-created visuals consistently outperform near-identical pages with stock images across a range of metrics, including dwell time and backlink acquisition. The images are not the only variable, but they are a consistent differentiator.

For teams with limited budgets, the priority should be original images on the pages that matter most: pillar content, high-converting product categories, and any content where you are trying to rank for competitive terms. Stock photography is fine for lower-priority supporting content.

A Practical Audit Framework

If you want to assess the current state of your image SEO, the audit does not need to be complicated. A standard crawl tool will surface most of the issues. Here is what to look for, in order of impact.

First, missing alt text. Pull a report of all images with empty or missing alt attributes. Prioritise fixing images on your highest-traffic pages and your most important commercial pages. Second, file names. Sample your image library and check whether file names are descriptive. If they are not, build the fix into your upload workflow going forward rather than trying to rename existing files retroactively (which can break URLs and cause 404s if not handled carefully).

Third, page weight. Use a tool like Google PageSpeed Insights or a similar performance auditing tool to identify pages where images are contributing disproportionately to load time. Fourth, format. Check whether your site is serving WebP or whether it is still defaulting to JPEG and PNG for all images. Fifth, sitemap coverage. Confirm that your image sitemap is submitted and that Google is indexing the images you expect it to index.

A good SEO audit process, as covered in resources like Moz’s SEO auditing framework, treats image optimisation as part of the technical health layer, not a separate workstream. That framing is right. Images are not a standalone discipline. They are part of the technical and content infrastructure of the page.

For teams building out their broader SEO capabilities, tools and communities like Search Engine Journal provide useful context on how search engine behaviour evolves over time, including how image search has developed as a channel in its own right.

Understanding how users actually interact with your pages, including where they look and what they engage with, is part of the optimisation picture too. Platforms like Hotjar can show you how users engage with image-heavy pages, which is useful data when you are deciding where to invest in original photography versus stock.

Content strategy and image strategy are more connected than most teams treat them. Resources from Copyblogger on content marketing strategy are worth reading alongside your image SEO work, particularly when you are thinking about how visual content supports broader content goals.

Image SEO is one component of a complete technical and content strategy. If you want the full picture of how it connects to keyword research, technical foundations, and authority building, the Complete SEO Strategy hub brings all of those elements together in one place.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions