Image Optimization

Images are the part of a web page most likely to be quietly sabotaging it. They are almost always the heaviest thing on the page, the single biggest cause of slow loading, and at the same time they are nearly invisible to a search engine, which cannot really see a picture the way you can. That combination, heavy and opaque, is exactly why image optimization matters, and why it has two distinct halves that are easy to confuse. One half is about weight: making images small and fast. The other is about understanding: making images legible to machines that cannot look at them. Do only one and you have half-optimized images. Doing both is the whole job.

Why it matters

Image optimization pays off in three connected ways, which is a lot of return for an unglamorous task. The first is speed. Because images are usually the heaviest thing on a page, optimizing them is typically the single biggest speed win available, and speed feeds directly into user experience and rankings. The second is accessibility and understanding. Described images can be read by screen readers for people who cannot see them, and that same description tells a search engine what the image is about, turning an opaque blob into meaningful content. The third is image search traffic. Properly optimized images can rank in image search, quietly bringing their own visitors to your site, an entire channel that heavy, undescribed images forfeit completely.

Put together, the case is lopsided. Well-optimized images make pages fast, accessible, understandable, and independently discoverable. Neglected images make pages slow and tell search engines nothing, wasting both performance and an entire traffic channel. Few technical tasks have such a clearly positive balance, which is why image optimization is one of the highest-value habits in technical SEO despite being nobody's idea of exciting.

Half one: the weight

The first half is a short, concrete checklist, and it is where the biggest, fastest gains live. Compress your images, reducing file size without a meaningful loss of visible quality, because most images are shipped far larger than they need to be. Size them for how they are actually displayed, rather than uploading a giant original and letting the browser shrink it on the fly, which forces visitors to download all that wasted weight. Use a modern, efficient format, which delivers the same picture in far fewer bytes than the older formats many sites still default to. And lazy-load the images below the fold, so images further down the page only load when the visitor scrolls toward them, rather than all at once on arrival.

None of this changes what your images look like; it changes what they cost. A page carrying a few carefully optimized images can be several times lighter than the same page carrying the raw originals, and because images dominate the weight of most pages, this is usually the most impactful performance work you can do. If you only ever do one thing for page speed, taming your images is the one to pick.

Half two: the understanding

The second half is about making a picture legible to a machine that cannot see it, and it rests on the labels you attach. The most important is alt text, a written description of the image, important enough to get its own section below. Beyond that, a few smaller signals add up. Descriptive filenames help: an image named for what it actually shows tells a search engine more than a meaningless string of characters from a camera. Surrounding context matters too, because a search engine reads the text near an image, and a relevant caption or nearby paragraph reinforces what the picture is about. And for certain kinds of images, structured data can add explicit machine-readable detail, though this is more specialised.

The unifying idea is that every one of these is a way of telling the blind traveller what is in the case. On their own each is a small signal; together, consistently applied across a site, they turn your images from opaque decorations into understood, indexable content. This half is less about performance and more about meaning, which is why it is so often neglected, the page looks fine without it, so people skip it, and quietly forfeit everything images could tell a search engine.

Alt text done right

Alt text deserves its own section because it does double duty and is constantly done badly. Its original and most important purpose is accessibility: screen readers read the alt text aloud to people who cannot see the image, so it is genuinely how blind and low-vision users experience your pictures. That same text is also a primary way search engines understand what an image depicts. Serve both by writing alt text that plainly and honestly describes the content and purpose of the image, as you would describe it to someone on the phone who cannot see it.

The failures are predictable. Leaving alt text empty forfeits both the accessibility and the search signal. Stuffing it with keywords instead of an honest description helps neither a screen-reader user nor a search engine, and reads as exactly the manipulation it is. Writing something useless like the image's file number tells nobody anything. The right instinct is simple and humane: describe the picture clearly and truthfully, for a person who cannot see it. Do that, and the search-engine benefit comes along for free, because a clear description is exactly what a machine wants too. Purely decorative images, which carry no meaning, are the one exception and can have empty alt text so screen readers skip them.

Here is how the topic sits in US search data.

This is a solid, evergreen topic at mid difficulty, the kind of subject where a clear, complete guide that actually covers both halves, weight and understanding, can earn its place, because a surprising number of pages on this topic cover only compression or only alt text and miss that it is really one job with two sides.

Serving the right size

One refinement ties the two halves together and is worth understanding on its own: serving a differently-sized image to different devices. A phone on a slow connection and a large desktop monitor do not need the same file, and sending the giant desktop image to the phone wastes exactly the weight you worked to cut. Modern responsive-image techniques let you provide several sizes of the same picture and let the browser choose the one that fits the device and screen, so a small screen downloads a small file and a large screen downloads a large one.

This is packing-light taken to its logical conclusion: instead of one compromise size for everyone, each visitor gets the lightest version that still looks sharp for them. It matters most on image-heavy, mobile-heavy sites, where the savings compound across every picture and every phone visitor. You do not need to hand-build this for every image if your platform supports it, but knowing that a page can and should serve right-sized images to each device is what separates good image optimization from merely acceptable image optimization.

Images and AI answers

Image optimization carries into the AI era mostly through its understanding half, and it is worth being clear about why. The systems building AI answers work primarily from text and from the descriptions attached to content, so a well-described image, with honest alt text, a sensible filename and supportive surrounding context, is far more legible to them than an unlabelled one. Your descriptions are, in effect, how a machine knows what your visuals contain, and that is as true for an answer engine as for image search.

The weight half helps too, in the same quiet way it helps everywhere: a light, fast page is easier for any crawler to fetch and process, the AI ones included. So the two halves of image optimization keep their split personality even here: packing light keeps the page accessible to the machines, and labelling clearly makes your visuals meaningful to them. As with so much good technical SEO, there is no separate AI tactic, just the same honest, thorough work paying off for a wider audience of readers, human and machine alike.

Mistakes to avoid

The neglect shows up in familiar ways.

Shipping huge uncompressed images, the most common single cause of a slow page.
Uploading giant originals and letting the browser shrink them, forcing visitors to download wasted weight.
Leaving alt text empty, forfeiting both accessibility and the search signal.
Stuffing alt text with keywords instead of honestly describing the image.
Doing only one half, compressing but never describing, or describing but never compressing.