EN
RO
Image SEO comes down to three concrete practices: a descriptive alt text for every relevant image, a modern format like WebP wherever it brings real benefits, and compression calibrated so the file stays small without visible quality loss. Done right, these three measures cut page load time, help Google understand and index images in Google Images, and improve the Core Web Vitals scores that directly influence search ranking.
Many site owners treat images as a purely visual detail, not as a performance and indexing element. In practice, images are frequently the heaviest part of a web page, and an uncompressed file or a missing alt text means lost organic traffic, both from standard search and from Google Images.
This guide explains, step by step, how to write alt text that actually helps SEO and accessibility, when it's worth switching to WebP, which compression levels preserve visual quality, and which common mistakes cancel out the whole optimization effort.
Why image optimization matters for SEO
Images affect SEO along three distinct lines. The first is performance: heavy images slow down page load, which directly affects Largest Contentful Paint (LCP), one of the Core Web Vitals signals Google uses to evaluate page experience. The second is accessibility and semantic indexing: alt text lets search engines and assistive technologies understand the content of an image that would otherwise stay opaque to a crawler. The third is extra traffic from Google Images, especially relevant for eCommerce, recipes, tutorials and visually documented content.
According to Google's official image publishing documentation (Google Search Central – Image publishing guidelines), the quality of the surrounding text context, the file name and the alt text all count directly toward how Google interprets and ranks that image.
How to write effective alt text for SEO
The alt attribute describes an image's content for screen readers and search engines, when the image itself can't be displayed or visually interpreted. Good alt text is descriptive, specific and short, not a list of crammed-in keywords.
- Describe what's actually shown in the image, not what you want to rank for in Google.
- Include the relevant keyword naturally, only if it fits the real description.
- Avoid phrases like "image of" or "picture showing" — they're redundant.
- Keep it around 8-12 words for content images.
- For purely decorative images, use
alt="", not a forced description.
Weak example: alt="seo image". Correct example for an online shoe store: alt="men's running shoes flexible sole". The difference is that the second example describes exactly the product shown and uses terms a buyer would actually search for.
WebP versus JPEG and PNG: when to choose each
WebP is an image format developed by Google that typically produces a smaller file size than JPEG or PNG at comparable visual quality. That doesn't mean WebP is always the right choice for every image.
- WebP — the best default choice for photos and content images on the modern web; supported by all major current browsers.
- JPEG — still useful for detail/color-heavy photos where a universal fallback is needed, or when the existing pipeline doesn't easily support conversion.
- PNG — recommended for images with transparency, logos and graphics with few colors, where lossy compression would visibly affect sharp edges.
- SVG — for icons and vector illustrations, where lossless scalability matters more than compressing a raster image.
In practice, a safe strategy is to export photos and content images as WebP, keep PNG for elements with transparency, and leave SVG for icons.
Compression without sacrificing quality: which thresholds work
Compressing an image reduces file size by removing redundant or barely perceptible data. The goal isn't the smallest possible file, but the best ratio between file size and perceived visual quality.
- For WebP/JPEG photos, a quality level between 70-85% preserves good visual quality in most cases, with a significant reduction in file size.
- Resize the image to its actual display size on the page, don't load a 4000px file for an 800px slot in the layout.
- Use the
srcsetattribute to serve different sizes on desktop and mobile, instead of a single large image for every screen. - Check the visual result after compression, not just the file size — over-compression artifacts are most visible on gradients or text areas.
Useful tools for compression without visible quality loss: squoosh.app (Google), tinypng.com for PNG/JPEG, and native WebP converters built into most modern image editors and CMS platforms.
File naming, responsive sizes and lazy loading
Image SEO doesn't stop at alt text and compression. File naming, loading behavior and responsive sizing all influence both indexing and the actual speed users perceive.
- Descriptive file name — use hyphen-separated words (
mens-running-shoes.webp), not generic names likeIMG_2031.jpg. - Lazy loading — add the native
loading="lazy"attribute for images below the fold, so they don't slow down the page's initial load. - No lazy loading on the main image — the most important image at the top of the page (hero image) should not be lazy-loaded, to avoid delaying LCP.
- Explicit dimensions — set the
widthandheightattributes (or CSS equivalent) to prevent layout shifts on load, relevant to the CLS score.
When extra effort pays off, and when the minimum is enough
Not every site needs the same level of image refinement. The decision depends on how much images matter for that site's search intent.
- eCommerce with many products — invest in an automated compression and WebP conversion pipeline at upload, plus systematic alt text derived from the product title; high image volume makes manual optimization inefficient.
- Blogs and content sites — prioritize descriptive alt text and correct compression for article images, where Google Images traffic can bring extra visitors.
- Presentation/service sites with few images — manual optimization, done once at publish time, is enough; an automated pipeline makes no sense for a few dozen images.
Common mistakes that cancel out image optimization
Even teams that know the basic rules repeat a few mistakes that cancel out their image optimization effort.
- Alt text stuffed with keywords — Google treats this as a manipulative practice, and screen reader users get a useless description.
- Converting to WebP without resizing — changing format without also reducing the actual file size to the displayed size brings minimal gains.
- Lazy loading applied to the main image too — delays the perceived load speed and can hurt the LCP score.
- Missing width/height attributes — causes visual layout shifts during loading, penalized by the CLS score.
- Raw, untreated files uploaded straight from camera/phone — multi-megabyte images uploaded with zero compression remain the most common cause of slow pages.
Practical implementation plan for an existing site's images
For a site that already has unoptimized images published, the following plan can be applied gradually, without stopping the site's ongoing activity.
- Week 1 — audit: use PageSpeed Insights or Search Console to identify pages with the largest image files and the weakest LCP score.
- Week 2 — conversion and compression: convert content images to WebP and compress them at 70-85% quality, starting with the pages found in the audit.
- Week 3 — alt text and file names: rewrite alt text for images that have none or a generic description, and rename relevant files for SEO.
- Week 4 — technical validation: add
loading="lazy"to secondary images, setwidth/height, and retest Core Web Vitals scores for the updated pages.
| Format | When to use it | Transparency | Typical file size |
|---|---|---|---|
| WebP | Photos and content images, default choice | Yes | Small |
| JPEG | Photos, universal fallback | No | Medium |
| PNG | Logos, low-color graphics, transparency | Yes | Large |
| SVG | Icons and vector illustrations | Yes | Very small |
Frequently asked questions about image SEO
Is WebP always better than JPEG for SEO?
For most photos and content images, WebP delivers a smaller file size at comparable visual quality, which indirectly helps SEO through faster load times. For images with transparency or low-color graphics, PNG can remain the better choice.
How many words should good alt text have?
There's no strict limit, but alt text of roughly 8-12 words, descriptive and specific, works well in most cases. Length matters less than the accuracy of the description.
Does image compression affect the visual quality visitors see?
At a quality level between 70-85% for WebP/JPEG, the visual difference is usually imperceptible to the human eye, while file size drops significantly. Over-compression below that range can introduce visible artifacts.
Is it worth optimizing images if the site gets little Google Images traffic?
Yes, because the main benefit isn't just Google Images traffic — it's also page load speed, which influences Core Web Vitals and overall user experience, both relevant to standard search ranking.
Do I need to optimize every image manually, or can I automate the process?
For sites with a large image volume (eCommerce, catalogs), automating WebP conversion and compression at upload is recommended. For sites with few images, manual optimization done once at publish time is enough.
Conclusion
Image SEO means descriptive, specific alt text, the right format for each image type (WebP as the default for photos, PNG for transparency, SVG for icons), and compression calibrated at 70-85% quality, which reduces file size without visibly affecting quality. Combined with correct lazy loading and explicit dimensions, these measures improve both site speed and the odds of getting indexed in Google Images.
Having speed or image indexing issues on your website? Let's discuss a technical audit.
Image generated with AI, used for illustrative purposes.
Write a comment