Accessible Star Ratings and Review Widgets
Star ratings are how shoppers judge products, and how review widgets most often fail assistive technology users. Making ratings readable, and review forms fillable, by everyone.
A star rating looks like the simplest element on a product page: five shapes and a number. Underneath, it is frequently an image with no text, a set of clickable icons with no names, or a form control that exists only for mouse users. The information shoppers rely on most is delivered in the format assistive technology handles worst.
This guide covers the display side, making ratings readable, and the input side, making leaving a review possible, plus the checks that keep third party review widgets honest.
Ratings must exist as text
Every rating shown as stars needs a text equivalent in the reading order: 4.5 out of 5 stars, 212 reviews, available to screen readers and visible or reachable for everyone. An average conveyed only by filled and empty star shapes is information a share of shoppers simply never receive.
The same rule applies in listing grids and search results, where ratings are compressed hardest. If space forces a choice, the number survives and the decoration goes, not the other way round.
Star input needs real controls
Leaving a rating should use a genuine radio group or equivalent native pattern: five labelled options, keyboard selectable, with the chosen value clear in text. A row of clickable star images with click handlers is not a form control, and it fails keyboard users at the exact moment they try to contribute.
Half star precision, hover effects, and animated fills are decoration layered on top. Build the accessible control first, then style it, and test that the styling never becomes the only way the control can be seen or used.
Review forms carry the usual duties
The written review form is a form, with the full set of obligations: labelled fields, instructions in text, errors that are announced and specific, and a confirmation that the submission worked. Review forms fail these basics often because they are treated as a widget rather than a journey.
Photo upload, helpful votes, and sorting controls inside review sections need the same pass. Each is a control a shopper may need, and each is commonly shipped as an unlabelled icon by the widget vendor.
Audit the widget, then configure or replace it
Most review displays come from a third party widget, which puts the code outside the stores control but not outside its responsibility. Run the keyboard and screen reader pass, document the gaps, and put them to the vendor in writing with dates.
Check the vendors configuration options before accepting a verdict of impossible: many widgets offer text alternatives and labelling options that are simply off by default. What configuration cannot fix becomes a documented factor at renewal time.
Related reading
If this topic is useful, these earlier guides go deeper on connected questions: Making Sale and Clearance Pages Accessible Under Time Pressure, Accessible Appointment and Consultation Booking, Step by Step, and Accessible Product Videos: Captions, Transcripts, and Controls.
Frequently asked questions
How should a star rating be presented to screen readers?
As text in the reading order: the average, the scale, and the review count, such as 4.5 out of 5 stars based on 212 reviews. The star shapes can remain as decoration for sighted users, but the information must exist as text.
What is the accessible pattern for leaving a star rating?
A real radio group with five labelled options that can be selected by keyboard and touch, with the current selection clear in text. Clickable star images without control semantics exclude keyboard users entirely.
Who is responsible for a review widgets barriers, store or vendor?
The vendor owns the code, the store owns the shopper experience and the decision to use the tool. Documenting gaps, configuring available options, and weighing accessibility at renewal are the stores levers.
Do ratings in search results and listings need text too?
Yes. Listings are where ratings do their hardest work, and the compressed format is where text equivalents are most often dropped. The number and count should survive even when space is tight.