Accessibility testing often begins with standards, automated scans, and technical checks. Those are important, but they don’t always tell you what it actually feels like to move through a website using assistive technology.
As part of our ongoing accessibility work, we recently conducted another round of testing on the ArtVersion website with people who use assistive technologies in their everyday lives.
The testing was structured around specific tasks rather than simply asking participants to look for accessibility problems. We wanted to understand whether people could find information, navigate interfaces, understand what was happening on a page, and complete tasks without unnecessary friction.
Overall, the site performed well. Most tasks were completed without significant issues. What was more interesting were several smaller observations that made us reconsider how some familiar design patterns behave when experienced differently.
Carousels Introduce More Context Than We Realize
Carousels were one area where we found opportunities for improvement.
Visually, their behavior is usually easy to understand: there is another slide, an image has changed, or additional content exists outside the current view. With assistive technology, those relationships need to be communicated much more deliberately.
The testing made us look again at controls, labeling, sequence, and how clearly a user can understand where they are within a collection.
It also exposed another consideration: alt text can become too much of a good thing.
We spend considerable time making sure imagery has appropriate alternative text. But within a carousel or image gallery, several highly descriptive alt texts encountered in sequence can quickly become an information overload.
That doesn’t mean providing less useful information. It means becoming better at deciding what someone actually needs to know about an image in its particular context. More description isn’t automatically more accessible. Sometimes a shorter, more purposeful description creates a better experience.
Long Pages Change the Cognitive Load
Another observation had little to do with a specific accessibility defect.
Some of our pages are intentionally content-rich. When viewing a page visually, it is relatively easy to scan ahead, jump between areas, and develop a sense of its overall size.
Sequential navigation changes that experience.
As pages become longer, headings, links, controls, and content sections accumulate. The issue becomes not only whether each individual element is accessible, but whether someone can maintain a clear sense of where they are and where the information they need sits within the larger page.
That has us looking more carefully at hierarchy, section length, navigation, and comprehension—not necessarily reducing information, but making longer experiences easier to understand and traverse.
Decorative Video Presents an Interesting Tradeoff
Background video gave us another question we are still working through.
Our background videos have no audio and are used as visual atmosphere rather than to communicate information. For people sensitive to movement, however, the ability to pause motion is important.
At the same time, someone who cannot see that background video may encounter a pause or play control for something that otherwise has no meaning within their experience.
That creates an interesting design problem: how do we give people who need to stop motion an obvious way to do so without making purely decorative media unnecessarily prominent to everyone else?
There isn’t much value in adding controls simply to satisfy a pattern if those controls introduce new confusion. We are continuing to rethink how we handle that balance, including the role of reduced-motion preferences and how motion controls are presented.
Accessibility Is More Than Finding Errors
That may have been the most useful takeaway from the testing.
Automated tools can identify many important problems. Manual review can uncover considerably more. But real-user feedback introduces questions that are harder to express as a pass or fail.
Is this description useful, or just long?
Is this control necessary, or is it another thing someone has to navigate?
Is the page technically structured correctly but still difficult to comprehend after several minutes of moving through it?
Those are design questions as much as accessibility questions.
The testing didn’t leave us with a long list of things that were broken. It gave us something more useful: a clearer understanding of several places where an already functional experience could become more thoughtful.
And that is exactly why we intend to keep doing it.