Skip to content
Dev Ring Technologies
Let’s talk
Back to blog

Keyboard navigation: check a website from menu to content

A menu can look simple with a mouse yet be difficult to use with a keyboard. Before accepting a website, set aside a pointer-free check of a few useful journeys. This guide covers navigation between pages and components, complementing form checks. It proposes an acceptance method, not a full accessibility audit or a conformance statement.

1. Define a task and starting point

In our fictional example, a visitor needs a service description and a practical answer. Plan the journey home → menu → service → frequently asked question → return. Record the starting address, language, browser, window size and tested version. Include embedded content actually present, such as a video player or map.

WAI’s Keyboard criterion calls for keyboard operation, with an exception for functions that depend on the path of movement. For this check, try to reach and use the functions your task needs. Perform actions that can send a message or change data only in an environment intended for those tests.

Source : W3C WAI — Understanding Keyboard (2.1.1)

2. Follow controls in an understandable order

From a reloaded page, move forward with Tab and back with Shift + Tab. Try Enter on common controls; try Space on buttons too. Some components use arrow keys or other keys: consult their intended interaction rather than requiring Tab on every item. Keep the pointer aside throughout the journey.

WAI requires focus order to preserve meaning and operation; it need not reproduce the visual layout. In the example, note whether moving from the menu to the service remains understandable. Record an unreachable control or confusing jump with the exact keys, rather than asking for every paragraph to receive focus.

Source : W3C WAI — Understanding Keyboard (2.1.1) · Source : W3C WAI — Understanding Focus Order (2.4.3)

3. Locate the active control and obscured areas

Focus is the element receiving keyboard actions. WAI distinguishes a visible focus indicator from avoiding a component being entirely hidden by site content. The minimum in 2.4.11 allows partial obscuring; this is not advice to hide part of a control.

For your check, inspect the indicator on light and dark backgrounds and near a sticky header. Repeat after scrolling and with a narrower window. If you lose the active control, take a screenshot and record the last identifiable element. Request an observable correction: being able to find the control and understand the action that activating it will trigger.

Source : W3C WAI — Understanding Focus Visible (2.4.7) · Source : W3C WAI — Understanding Focus Not Obscured, Minimum (2.4.11)

4. Enter a component, then leave it

Open the example menu or dialog, explore its controls, close it and continue through the page. Try its intended exit method, often Escape or a button. Record where navigation resumes and whether the task remains possible. Repeat for each embedded component in scope, without making a real booking or sending anything.

WAI requires a keyboard exit from a component reached by keyboard, with instructions if leaving needs a nonstandard method. Temporarily keeping focus inside a modal dialog is not automatically a trap if it can be exited. Distinguish an expected loop while it is open from a blocker with no known way out.

Source : W3C WAI — Understanding No Keyboard Trap (2.1.2)

5. Repeat the journey and useful variations

Start again at the initial address, then repeat after browser back and reload. Try the menu closed and open, and every published language. For Arabic, judge the journey in its context; do not infer keyboard order solely from text alignment. Keep conditions consistent to compare before and after a correction.

Add a narrow window that triggers the compact menu. This layout check does not replace testing on a phone with appropriate equipment. If a control does not respond, check the browser and system keyboard-navigation settings, record them and reproduce. Do not turn an environment uncertainty into a passing result.

6. Report an issue someone can reproduce

Complete the worksheet below for each journey. For a blocker, give the address, starting state, ordered keys, expected result and observation. One fictional example: “menu open, three Tabs, close, Tab: focus indicator no longer visible”. A screenshot locates the problem; the sequence lets someone reproduce it.

After a correction, repeat the affected journey and a nearby journey sharing the menu or component. Keep the tested version and remaining limits. An automated test or one successful journey does not cover every page or disability. Plan further assessment with a qualified person when your scope requires it.

Keyboard navigation check worksheet

Keyboard navigation check worksheet
StageResult to observeEvidence to keep
StartKnown task and starting stateAddress, version and environment
JourneyUsable controls in an understandable orderKeys and destinations
IndicatorIdentifiable active controlScreenshot and position
ComponentPossible entry and exitSequence and resumption
VariationsReproducible resultLanguage and window
CorrectionBlocker resolved and neighbour checkedVersion and limits

Frequently asked questions

Should Tab reach all text?

No. Follow the controls needed for the task. The focus-order criterion addresses the order of elements that receive focus; it does not require every paragraph to be focusable.

Is a modal dialog retaining focus a defect?

Not necessarily. Check that a keyboard exit is available and understandable, and that the journey can continue. A loop while the dialog is open is not enough to identify a trap.

Does a successful journey prove conformance?

No. It provides evidence limited to the recorded tasks, versions and environments. A complete assessment needs a broader scope and checks.

Choose a first journey

Pick a frequent journey, complete the worksheet and have the person responsible for the correction reproduce any blocker. Then extend checks to components and pages sharing that journey.

Sources and scope

Primary sources checked on November 5, 2026. WAI Understanding documents explain WCAG criteria and are informative. The scenarios, worksheet and fictional journey are editorial proposals. No audit of the Dev Ring website, client outcome or conformance commitment is claimed.