WCAG dragging movements 2.5.7 is a Level AA success criterion added in WCAG 2.2, and its rule fits in one sentence: anything a user can do by dragging must also be doable "by a single pointer without dragging", unless dragging is essential or the browser itself provides the drag (W3C). A single pointer means one click or one tap at a time. So the fix is almost always the same move: next to every drag interaction, add a control you can click. A sortable list gets "Move up" and "Move down" buttons. A kanban card gets a "Move to…" menu. A slider jumps to wherever you click on its track. A map gets pan buttons. The one fix that does not count on its own is keyboard support, which surprises a lot of teams. W3C says plainly that keyboard equivalence for a drag "does not automatically meet this success criterion" (W3C). The rest of this guide covers what counts as dragging, how to fix the common patterns, where 2.5.7 sits legally, and what a scanner can and cannot tell you about it.
What does WCAG 2.5.7 Dragging Movements require?
The full criterion: "All functionality that uses a dragging movement for operation can be achieved by a single pointer without dragging, unless dragging is essential or the functionality is determined by the user agent and not modified by the author" (W3C).
It exists for people who can't hold a pointer down and move it accurately at the same time. W3C names the users: people on trackballs, head pointers, eye-gaze systems and speech-controlled mouse emulators (W3C). For someone using eye gaze, "press, hold, travel 400 pixels, release over the right column" is a hard task. Two separate clicks are not.
There is one sufficient technique, G219, "ensuring that an alternative is available for dragging movements that operate on content" (W3C G219). And one documented failure, F108: the only way to trigger a function is to drag a target from one position to another, with no single-pointer alternative (W3C F108). Its two examples are a drag-to-reorder list and a kanban board, which is a fair summary of where this bug lives in real products.
The alternative doesn't have to be one click. F108 explicitly allows "a series of single pointer interactions", for example selecting the item, opening a dropdown, and picking the destination from a list (W3C F108). Three clicks pass. One drag with no fallback fails.
What counts as a dragging movement, and what doesn't?
A dragging movement is a pointer-down on an element, a move, and a pointer-up somewhere else, where only the start and end points matter. That separates it from the gestures covered by 2.5.1 Pointer Gestures, a Level A criterion about "multipoint or path-based gestures" such as swipes, where the route or speed of the pointer matters (W3C). The two criteria point at each other: 2.5.1 "does not apply to gestures that involve dragging in any direction", and the single-pointer alternative you add for 2.5.7 can't itself be a swipe (W3C).
Sliders sit in between. Most use pointer capture, so the thumb keeps following you even if you wander off the track, and that is plain dragging. A slider that drops your grip as soon as the pointer strays from the track is path-based as well, and W3C notes it may fail both 2.5.1 and 2.5.7 (W3C).
Out of scope:
- Dragging the browser provides. Native scrollbars, touch scrolling, and pull-to-refresh are the user agent's job, not yours. That changes once you suppress those behaviors or build your own scrolling, because then the drag is yours (W3C).
- Essential dragging. A freehand drawing tool, or a signature pad, records the path itself, so there's nothing to replace it with. Be strict with yourself here. A drag-to-reorder list is never essential, because the outcome (the new order) can be reached another way.
Why doesn't keyboard support satisfy 2.5.7?
Because it's a pointer criterion, and the users it protects are pointer users. Someone driving a head pointer may have no practical keyboard at all. W3C puts it directly: this requirement "is separate from keyboard accessibility", and a keyboard equivalent only helps here if it also gives you controls that can be clicked or tapped (W3C).
You still need the keyboard path. 2.1.1 Keyboard applies to the same widget, and the two get evaluated separately. In practice the cheapest design covers both at once: visible buttons are focusable and clickable, so one "Move up" button satisfies the keyboard user and the eye-gaze user together. The space-to-lift, arrows-to-move keyboard mode that several drag-and-drop libraries ship is good keyboard work. It is invisible to someone who only has a pointer.
How do I fix the common drag patterns?
Here is the approach I use for each one. These are design patterns rather than W3C wording, except where a source is cited.
Sortable lists
Add explicit move controls to each row. Plain buttons with real accessible names:
<li>
<span id="item-3">Quarterly report</span>
<button type="button" aria-describedby="item-3">Move up</button>
<button type="button" aria-describedby="item-3">Move down</button>
</li>
After a move, keep focus on the button the user pressed so they can press it again, and announce the new position in a polite live region ("Quarterly report, position 2 of 5"). W3C lists up/down buttons as one of its examples for this exact case (W3C).
Kanban boards
Give each card a "Move to…" menu that lists the columns. Click the card's menu button, pick "Done", and the card moves. That is F108's own example of a passing multi-step alternative: activate the target, open a dropdown, choose the destination (W3C F108). Most teams already have a card detail view with a status field. If that field changes the column, you may already pass, as long as users can find it.
Sliders and range inputs
Make a click anywhere on the track move the thumb there. W3C gives this as its slider example (W3C). Native <input type="range"> controls generally behave this way when you click the track, but custom sliders built from divs often don't. Test each one in the browsers you support rather than assume. A text field for entering the exact value is also an acceptable alternative (W3C), and for price filters it's usually the better UI anyway.
For the keyboard side of a custom slider, the WAI-ARIA Authoring Practices pattern is the reference: role="slider" with aria-valuenow, aria-valuemin and aria-valuemax, arrow keys to step, Home and End for the limits (W3C APG).
Maps, carousels and crop tools
Maps get on-screen pan buttons and carousels get previous/next buttons; both are W3C examples (W3C). For an image crop box you drag into place, nudge buttons or numeric width/height/offset fields do the same job.
File drop zones
Pair every "drop files here" area with a real <input type="file"> behind a visible "Choose files" button. Most upload components already do this. The ones that don't usually removed the button for looks, so check the component before you trust it.
Is WCAG 2.5.7 legally required?
It depends on which version of WCAG your law points at, and today most of them point at 2.1, which doesn't include 2.5.7.
In the EU, the harmonised standard currently cited is EN 301 549 v3.2.1, adopted through Implementing Decision (EU) 2021/1339, and it maps to WCAG 2.1 AA. The new edition, EN 301 549 V4.1.1, moves web content to WCAG 2.2 and lists dragging movements as clause 9.2.5.7, one of six clauses that separate 2.1 AA from 2.2 AA (ETSI EN 301 549 V4.1.1). When I checked for this post I couldn't find a citation of V4.1.1 in the Official Journal yet, and until there is one, it doesn't carry the presumption of conformity. Our EN 301 549 guide and the post on what V4.1.1 changes cover that in detail, and the European Accessibility Act page explains how the standard connects to the law.
In the US, the DOJ's Title II rule for state and local government names WCAG 2.1 Level AA as its technical standard (ADA.gov), so 2.5.7 isn't part of it. Our ADA overview covers the rest of the US picture.
My view: fix it anyway, and fix it now. Drag-only interfaces lock out real users today, the fix is usually two buttons, and 2.2 is the version the next round of standards points at. I'd go further and design the click path first, then add drag on top as the shortcut for people who like it. Teams that build drag first tend to bolt the alternative on late, and it shows.
Can a scanner find dragging movement failures?
Mostly no, and I'd rather say that plainly. A scanner reads the DOM and the accessibility tree. It can see that an element is draggable="true" or has a pointer handler attached, but it can't tell whether a separate button somewhere on the page does the same job. G219's test procedure is two human steps: find the functions triggered by dragging, then check that a single-pointer alternative operates the same function (W3C G219). That's a judgment about functionality, not markup.
That fits the wider numbers. Across more than 2,000 audits, Deque found automated tools caught about 57% of issues by volume (Deque), and estimates by share of success criteria run closer to 30%. 2.5.7 sits in the manual remainder. What a scanner does well here is the neighborhood: unlabeled icon buttons on your new "Move up" controls, a custom slider missing aria-valuenow, a live region that never announces. The test itself is quick, though. Put the mouse in one hand, promise yourself you won't hold the button down, and try to do everything the screen offers.
Frequently asked questions
What is WCAG 2.5.7 Dragging Movements?
It's a Level AA criterion in WCAG 2.2 that requires every function operated by dragging to also work with a single pointer without dragging, unless dragging is essential or provided by the browser (W3C).
Does keyboard support satisfy 2.5.7?
No, not by itself. W3C says the requirement is separate from keyboard accessibility, so you need a click or tap alternative such as visible buttons, which also happen to work from the keyboard (W3C).
Can the alternative take more than one click?
Yes. F108 accepts a series of single-pointer interactions, for example selecting an item, opening a menu, and choosing a destination (W3C F108).
Do native range inputs pass 2.5.7?
Often, if clicking the track moves the thumb there, which is the single-pointer alternative W3C describes for sliders. Test yours in the browsers you support, or add a text field for the exact value, which W3C also accepts (W3C).
Is 2.5.7 the same as 2.5.1 Pointer Gestures?
No. 2.5.1 is Level A and covers path-based and multipoint gestures like swipes and pinches; 2.5.7 is Level AA and covers dragging, where only the start and end points matter (W3C).
Check the rest of your WCAG 2.2 gaps
Dragging is a manual check, but it rarely travels alone. Run a free scan to catch the machine-detectable issues around your drag components, like unnamed buttons and broken slider attributes, and work through the manual WCAG 2.2 criteria with our WCAG checklist. The scan finds what code can show. The one-handed, no-dragging test is yours to run.
Pavel Charkasau, founder, wcagc.com. Last updated 17 November 2026.
Sources
- Understanding SC 2.5.7: Dragging Movements, W3C WAI: criterion text, Level AA, intended users, examples, the user-agent exception, and the note that keyboard equivalence does not meet it on its own. Accessed 17 November 2026.
- Technique G219, W3C WAI: the sufficient technique and its test procedure. Accessed 17 November 2026.
- Failure F108, W3C WAI: the documented failure, the list and kanban examples, and the multi-step alternative. Accessed 17 November 2026.
- Understanding SC 2.5.1: Pointer Gestures, W3C WAI: how path-based gestures differ from dragging, and the slider edge case. Accessed 17 November 2026.
- Slider pattern, WAI-ARIA Authoring Practices Guide, W3C: roles, states and keyboard interaction for custom sliders. Accessed 17 November 2026.
- ETSI EN 301 549 V4.1.1: clause 9.2.5.7 and the WCAG 2.2 reference. Accessed 17 November 2026.
- Commission Implementing Decision (EU) 2021/1339, EUR-Lex: citation of EN 301 549 v3.2.1. Accessed 17 November 2026.
- Fact sheet: Title II web and mobile accessibility rule, ADA.gov: WCAG 2.1 Level AA as the technical standard. Accessed 17 November 2026.
- Automated testing identifies 57% of accessibility issues, Deque: automated coverage figures. Accessed 17 November 2026.