Skip to content
Tutorials Center
My learning

Topics

Lesson 37 of 60 ยท Manual Testing in Practice

Web UI, Usability and Accessibility Checks

Practical manual checks for web forms, navigation, responsive layouts, keyboard and screen-reader access, plus usability heuristics and a contrast-ratio calculator.

Most users never see your code or your API. They see a page: a form to fill in, a menu to follow, a layout that either fits their phone or does not. UI testing checks that this surface works and looks right; usability testing asks whether people can actually get things done with it; accessibility testing asks whether people with disabilities, or with a keyboard instead of a mouse, can do the same.

A tester with a short, well-practiced checklist finds a surprising number of real problems in minutes, with no tools beyond a browser. This lesson gives you those checklists, explains the reasoning behind them, and includes two small stand-ins that calculate things you would otherwise judge by eye. Test design ideas such as equivalence partitioning apply to forms too, so combine them with the lists below.

Three lenses on one page

Functional UI Does every control do its job?
Usability Can people succeed without help?
Accessibility Can everyone use it, including with a keyboard?
1 / 2

Checking forms

Forms are where users hand over data and where many defects hide. Work through these for every form:

  • Labels and hints. Every field has a visible label, and the format is explained ("MM/DD/YYYY") before the user fails.
  • Required fields. Leaving one empty gives a clear message next to that field, not a generic banner or a silent failure.
  • Validation. Try the boundaries and odd values you know from Boundary Value Analysis: empty, one character, maximum length, spaces only, emoji, pasted text.
  • Error messages. They say what is wrong and how to fix it. "Invalid input" is a defect; "Enter a date in the future" is a message.
  • Keeping data. After an error, the fields the user already filled in are still filled in. Losing a long form is a classic cause of abandoned sign-ups.
  • Double submit. Click the submit button twice quickly. Does it create two orders?
  • Browser help. Does autofill work? Can the password field be pasted into? Does the right mobile keyboard appear (numeric for phone numbers)?
  • Success feedback. After submitting, does the page confirm what happened?

Checking navigation and content

  • Every link and button goes where its text says, and none lead to an error page.
  • The back button and refresh behave sensibly. Refreshing after a form submit should not silently repeat the action without warning.
  • The current location is visible (highlighted menu item, breadcrumbs, page title).
  • Page titles are meaningful and unique, which matters for browser tabs, history and screen readers.
  • Text has no spelling mistakes, placeholder text such as "Lorem ipsum" is gone, and images load with sensible alternatives when they do not.
  • A custom 404 page exists and offers a way back.
  • Long content wraps; long words and URLs do not break the layout.

Responsive and visual checks

Resize the window (or use the browser's device toolbar) and look for breakage at the widths your users have: about 360 to 430 px for phones, 768 to 1024 px for tablets, 1280 px and wider for desktops.

  • No horizontal scrolling on a phone, and no text cut off.
  • Buttons and links are large enough to tap with a thumb (a common guideline is at least about 44 by 44 CSS pixels).
  • Zoom the browser to 200%. Content should still be readable and usable without losing features.
  • Rotate to landscape, and check dark mode if the product supports it.
  • Look for overlapping elements, clipped images and text that grows when translated. Compatibility and Localization Testing returns to this.

Keyboard checks

Many people cannot use a mouse, and many more simply prefer the keyboard. Unplug the mouse (or just do not touch it) and use the page.

  • Tab moves forward and Shift+Tab backward through every interactive item: links, buttons, fields, menus.
  • The focus indicator is always visible. If you cannot see where you are, a keyboard user cannot either.
  • The tab order matches the visual reading order, and nothing is skipped or unreachable.
  • Enter activates links and buttons; Space toggles checkboxes and buttons; Escape closes dialogs and menus.
  • A dialog traps focus while open and returns it to the control that opened it when closed.
  • There is no keyboard trap: you can always Tab out of any widget.
  • A "skip to main content" link near the top lets keyboard users bypass the menu.

The first stand-in simulates one rule worth knowing. In HTML, elements with a positive tabindex jump ahead of everything in natural order, which usually produces a confusing path. The code orders a small invented page the way a browser would.

JS Tab order and positive tabindex

Press Run (or Ctrl + Enter) to see the console output here.

No output.

This is a simplified model: real browsers also include only visible, enabled controls, and they order equal positive values by position in the page. It still shows the finding a tester would report: the skip link, which should be first, is third.

Screen-reader basics

A screen reader announces the page as speech or braille. You do not have to become an expert, but a five-minute check teaches a lot. Try the reader built into your system (Narrator on Windows, VoiceOver on macOS and iOS, TalkBack on Android) with the screen off or ignored.

  • Images that carry meaning have a text alternative; decorative ones have an empty one, so they are skipped.
  • Form fields are announced with their labels. If the reader says only "edit text", the label is not connected.
  • Buttons and links have meaningful names. "Click here" and "Read more" repeated ten times are useless when heard in a list of links.
  • Headings form a sensible outline, so users can jump between sections.
  • Changes such as error messages and "added to cart" notices are announced without the user hunting for them.
  • Custom widgets (tabs, menus, sliders) announce their role and state, for example "expanded".

Automated tools catch only part of these issues. They cannot judge whether alternative text is meaningful, which is why manual checks remain essential.

Colour and contrast

Text must stand out from its background. The Web Content Accessibility Guidelines (WCAG) define a contrast ratio from 1:1 (no difference) to 21:1 (black on white). At level AA, normal text needs at least 4.5:1, and large text (roughly 24 px, or 19 px bold) needs at least 3:1. Also never rely on colour alone: an error shown only by turning a border red is invisible to many users.

The ratio comes from a formula based on the relative luminance of each colour. The second stand-in implements it, so you can check any pair of hex colours instead of guessing.

JS WCAG contrast ratio calculator

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Light gray on white looks elegant and fails both thresholds; orange on white is similar. Tools such as browser developer panels show the same ratio when you inspect a text element. Remember that the formula covers text; other graphics and form borders have their own (3:1) requirement.

Accessibility testing checks that people with disabilities can use the product: people who are blind or have low vision, who are deaf or hard of hearing, who cannot use a mouse, who have motor, cognitive or attention differences, and people with temporary limits (a broken arm, bright sunlight, a noisy room). Accessible products are also better for everyone: captions help in a loud train, keyboard shortcuts help power users.

WCAG in brief

The Web Content Accessibility Guidelines (WCAG), published by the W3C, are the main technical standard. Their four principles are summarised by the word POUR:

  • Perceivable: information can be perceived, with text alternatives for images, captions for audio and video, sufficient contrast, and content that adapts to different presentations.
  • Operable: everything can be operated with a keyboard, users have enough time, nothing flashes dangerously, and navigation is clear.
  • Understandable: text is readable, behaviour is predictable, and forms help users avoid and fix mistakes.
  • Robust: content works with current and future assistive technologies, which means valid, semantic markup.

Success criteria have conformance levels: A (minimum), AA (the usual legal and commercial target) and AAA (the highest, not required for whole sites). WCAG has versions (2.1, 2.2 and later drafts); check which version applies to your product.

ARIA and semantic HTML

WAI-ARIA adds attributes (role, aria-label, aria-expanded, aria-live) that tell assistive technology what a custom widget is and what state it is in. The first rule of ARIA is do not use it when a native element does the job: a real <button> already has the role, keyboard behaviour and focus handling, while a <div role="button"> needs all that built by hand. Testers check that:

  • every control has an accessible name (label, aria-label or text);
  • custom widgets expose their role and state, and update them (for example aria-expanded toggles);
  • dynamic messages (errors, toasts) are announced with live regions;
  • headings, landmarks (main, nav) and lists reflect the real structure.

What to test

Area Check
Text alternatives Meaningful images have useful alt text; decorative ones have empty alt
Contrast Text 4.5:1 (3:1 large); UI components and focus rings 3:1
Keyboard Everything reachable and operable with Tab, Enter, Space, arrows, Esc; no keyboard traps; visible focus; logical order
Screen reader Page title, headings, labels, errors, dynamic updates are announced sensibly
Forms Visible labels linked to fields, errors named in text and tied to the field, required fields marked
Zoom and reflow 200% zoom and a narrow window do not hide or overlap content
Colour Information is not conveyed by colour alone
Media Captions, transcripts, no autoplay sound, controls usable
Motion Animation can be paused; respects reduced-motion settings
Touch targets Big enough and spaced on mobile

See the keyboard, screen-reader and contrast sections earlier in this lesson for hands-on steps.

Tools and manual testing

  • Automated checkers (axe, Lighthouse, WAVE and similar) find roughly a third to a half of issues: missing alt text, low contrast, missing labels. They cannot judge whether alt text is meaningful or a focus order is logical.
  • Manual checks: keyboard-only navigation, zoom, and reading the page with a screen reader (NVDA or JAWS on Windows, VoiceOver on macOS and iOS, TalkBack on Android).
  • Real users: the most valuable feedback comes from people who use assistive technology daily.

Build checks into the process: design reviews, automated accessibility tests in the pipeline, and a manual pass before release.

Many countries require accessibility, and the rules differ and change, so confirm with legal advice for your market. Common reference points:

  • ADA (Americans with Disabilities Act) in the United States, applied by courts to websites and apps, usually with WCAG AA as the benchmark.
  • Section 508 of the US Rehabilitation Act for federal agencies and their suppliers.
  • European Accessibility Act (EAA) and the EN 301 549 standard in the European Union, covering many consumer products and services.
  • AODA in Ontario, Equality Act in the United Kingdom, and similar laws elsewhere.

A VPAT (voluntary product accessibility template) or accessibility conformance report documents how a product meets a standard and is often requested in procurement.

JS A tiny accessibility audit of HTML text

Press Run (or Ctrl + Enter) to see the console output here.

No output.

A script like this resembles what automated checkers do, and it shows their limit: it can tell that alt exists, not whether "Sales chart: up 12 percent in Q3" is the best description.

Usability heuristics

Usability testing is usually done with real users performing tasks while you observe. But even without them, you can review a screen against well-known rules of thumb. The ten heuristics published by Jakob Nielsen are the classic list; here is a compact version you can apply during any session:

  • Visibility of status: the system tells you what is happening (loading, saved, failed).
  • Match with the real world: words and icons are familiar to the user, not internal jargon.
  • User control: undo, cancel and back are available. Mistakes are recoverable.
  • Consistency: the same thing looks and behaves the same everywhere.
  • Error prevention: confirm destructive actions; disable impossible choices.
  • Recognition over recall: options are visible, so users do not have to remember them.
  • Flexibility: shortcuts for experts without hiding things from beginners.
  • Minimalism: no clutter competing with what matters.
  • Helpful errors: plain-language messages that suggest a fix.
  • Help when needed: documentation and hints that are short and searchable.

When you log a usability problem, describe the user task and the obstacle ("I could not find where to change my email after 2 minutes"), rather than your opinion ("the layout is ugly"). Report it with the structure from Writing Excellent Bug Reports.

Accessibility is also a legal requirement in many countries and for public-sector sites. If you are unsure of the rules for your product, ask. A tester's job is to raise the question early, not to interpret the law.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. A form shows 'Invalid input' above the page when the email field is wrong. What is the best bug report?

    Good error messages say what is wrong and how to fix it, near the field. A vague banner is a usability defect.
  2. 2. What should you see when you press Tab through a page?

    Keyboard users rely on a visible focus indicator and a logical order. Missing focus styles or a jumbled order are common accessibility defects.
  3. 3. Which text colour pair satisfies WCAG AA for normal-size text?

    AA requires at least 4.5:1 for normal text (3:1 for large text). Visual judgement varies between people and screens, which is why a measured ratio is used.
  4. 4. A page has a button with tabindex=5 and a skip link with the default order. What is the likely problem?

    Positive tabindex values move an element ahead of all natural-order elements, which usually creates a confusing path.
  5. 5. Which finding would an automated accessibility scan most likely miss?

    Tools can detect missing attributes but cannot judge whether the text is meaningful. That needs a human.

Summary

  • UI checks cover forms, navigation, content and responsive behavior; usability asks if users can succeed; accessibility asks if everyone can.
  • For forms, test labels, required fields, error messages, kept data, double submits and boundary values.
  • Keyboard checks: Tab order, visible focus, Enter/Space/Escape, no traps, a skip link.
  • Screen readers need labels, meaningful link and button names, headings and announced changes.
  • WCAG AA contrast needs 4.5:1 for normal text and 3:1 for large text; never use colour alone to carry meaning.
  • Review against usability heuristics and describe problems as tasks and obstacles, not opinions.

Next up: Compatibility and Localization Testing.