Alt text is the written description carried in an image’s alt attribute. Screen readers announce it in place of the image, browsers display it when an image fails to load, and search engines and AI systems use it to understand what the image shows. It is one line of markup, it takes seconds to write, and it is skipped more often than almost any other accessibility requirement.
Alt text exists primarily so that people who cannot see the image still get its meaning. The search benefit is real but secondary, and treating it the other way round produces alt text that helps nobody, which is the mistake worth correcting first.
The accessibility obligation
A screen reader user navigating your site hears the alt text where a sighted visitor sees the picture. If it is missing, some screen readers read out the filename instead, which turns a photograph of your team into an announcement of letters and numbers. If it is unhelpful, the information in the image is simply lost.
This is codified. WCAG success criterion 1.1.1, Non-text Content, is a Level A requirement, the most basic tier of conformance. It is also among the most frequently cited issues in accessibility complaints, precisely because it is easy to test for at scale. See ADA compliance.
The rule that makes it easy
Describe what the image conveys in this context, not what it depicts in the abstract. The same photograph needs different alt text on different pages.
A photo of a finished kitchen on a portfolio page might be “Renovated galley kitchen with white shaker cabinets and butcher block counters.” The same photo illustrating an article about lighting might be “Under-cabinet LED strips lighting a kitchen worktop.” The image has not changed; what it is there to communicate has.
The test: if you removed the image and replaced it with your alt text, would the page still make sense to the reader?
The four cases
Informative images
Images carrying content. Describe what matters, specifically, and stop. Around 125 characters is a reasonable working ceiling, not because of a hard limit but because screen reader users do not want a paragraph read at them.
Decorative images
Background textures, dividers, abstract shapes, an icon sitting beside text that already says the same thing. These take an empty alt attribute: alt="". That is deliberately different from omitting the attribute. An empty alt tells assistive technology to skip the image silently. A missing alt leaves it to guess, and it often guesses badly. This distinction is the single most useful thing to know about alt text, and the one most often got wrong.
Functional images
An image that is a link or a button. Describe the destination or action, not the picture. A logo linking to the homepage should be “City of Oaks Marketing home,” not “logo.” A magnifying glass button should be “Search,” not “magnifying glass icon.” What the user needs to know is what happens when they activate it.
Complex images
Charts, diagrams, infographics. Alt text alone cannot carry a data table. Give a short alt naming what it shows, then provide the detail in nearby visible text or a linked description. Putting the figures in text serves screen reader users, search engines, and AI systems that cannot read images, all at once.
The keyword question, answered honestly
Alt text is a minor signal for image search and a small piece of context for the page. It is not a keyword slot.
Writing “raleigh web design raleigh wordpress developer raleigh seo agency” into an alt attribute is keyword stuffing, it violates Google’s guidelines, and a screen reader user hears that string read aloud. It fails both audiences simultaneously.
The useful guidance is unglamorous: describe the image accurately and naturally, and if a relevant term genuinely belongs in that description, it will appear on its own. Accurate descriptions tend to contain the right words because the image is about the thing.
What not to write
“Image of” or “picture of” at the start. Screen readers already announce it as an image, so this is redundant on every single one.
The filename. Never useful.
The same text as a visible caption. The user hears it twice. If the caption already describes the image, an empty alt is usually correct.
Nothing at all on a meaningful image. The default failure.
A description of an image’s appearance when its function is what matters, as with icons and logo links.
Where it breaks in practice
Even on sites where people write good alt text, a few patterns quietly lose it.
Lazy-loading placeholders. Some lazy loading implementations render a placeholder image first and swap in the real one with JavaScript, and the placeholder can carry an empty alt while the real description never reaches the markup. Logos and badge images are common casualties. Always check the rendered output, not just what the media library holds.
CSS background images. An image set as a background has no alt attribute and is invisible to assistive technology entirely. If it carries meaning it should be a real <img> element. Backgrounds are for decoration.
Page builders and icon libraries, which insert images and inline SVGs with no description and no obvious place for an editor to add one.
Bulk uploads, where fifty product images arrive with alt fields nobody filled in.
WordPress caption confusion. The media library has separate fields for alt text, title, caption and description, and only the first is the alt attribute. Filling in the title field and assuming it covers you is a common and invisible error.
The AI dimension
Systems that read pages to summarize or cite them mostly cannot see images. For them, alt text and surrounding body copy are the only account of what an image contributes. If your key evidence sits in a chart with no textual equivalent, it is invisible to them exactly as it is to a screen reader. The practices that make a page accessible make it machine-readable, which is a recurring pattern and the reason semantic HTML and alt text keep turning up in the same conversations.
Auditing it
Automated tools find missing alt attributes quickly. A site crawler such as Screaming Frog will list every image without one, and browser tools including Lighthouse and axe flag them per page.
What automation cannot judge is whether the alt text is any good. “image1” passes an automated check. So does a keyword-stuffed string. The practical approach is to use tooling to find the gaps, then review the important templates by hand: the header, the homepage, service pages, and anywhere images carry real information.
Where this fits
Image alt audits are part of our technical SEO services and of our accessibility reviews, and we build alt fields into every custom theme so editors are prompted rather than having to know. If you are unsure how your site would fare on a basic accessibility check, book a discovery call.
