UX Audit Checklist 2026: 30 Items to Test in 2 Weeks

A UX audit done well is one of the highest-ROI design investments a company can make. Done badly, it produces a 50-page PDF nobody reads. The difference is in the framework. This is the framework we use at UX Studio when running paid audits for clients — refined across 50+ engagements in 2024-2026.

The audit is structured as 30 items across six dimensions. We typically deliver against the full checklist in 1-2 weeks for a single product, with a prioritized roadmap as the primary output.

TL;DR — the 6 audit dimensions

1. Heuristic evaluation (5 items) — Nielsen's 10 heuristics applied
2. Accessibility (5 items) — WCAG 2.2 AA compliance check
3. Performance (5 items) — Core Web Vitals + perceived speed
4. Conversion (5 items) — funnel analysis, friction points, drop-offs
5. Content and copy (5 items) — clarity, voice, microcopy quality
6. AI and modern UX patterns (5 items) — 2026 expectations

Each item gets: severity (critical/high/medium/low), effort to fix (S/M/L), and recommended next step.

Dimension 1: Heuristic evaluation

1. Visibility of system status

Test: Do users always know what's happening? Loading states, success confirmations, error states, processing indicators visible at every interaction.

Common failures: 5+ second operations with no feedback. Form submissions that show no confirmation. Background processes the user can't track.

2. Match between system and real world

Test: Does the language and structure match user mental models? Are technical terms, jargon, or system-internal vocabulary leaking into the UI?

Common failures: "User-Account Settings" instead of "Profile". Database field names exposed in the UI. Concepts that make sense to engineers but not to users.

3. User control and freedom

Test: Can users undo, exit, or recover from any state? Are there clear escape hatches? Is destructive action confirmable?

Common failures: No undo for accidental delete. Modal dialogs you can't close. Multi-step forms that lose data on back-button.

4. Consistency and standards

Test: Are similar elements treated similarly? Do interaction patterns match platform conventions (web, iOS, Android)?

Common failures: Buttons styled three different ways for the same action. Custom date pickers when the native one works. Mixed iconography.

5. Error prevention and recovery

Test: Are common error states prevented before they happen? When errors do happen, are the messages clear, actionable, and human?

Common failures: "Error 500" with no explanation. Validation that fires after submit instead of inline. No suggestion for how to fix the problem.

Dimension 2: Accessibility (WCAG 2.2 AA)

6. Color contrast

Test: All text meets WCAG 2.2 AA contrast (4.5:1 for normal text, 3:1 for large). Test with axe DevTools or Lighthouse.

Common failures: Light gray placeholder text. Brand colors that don't pass on white backgrounds. Status indicators that rely only on color.

7. Keyboard navigation

Test: Every interactive element reachable and operable via keyboard. Visible focus indicator. Logical tab order.

Common failures: Custom dropdowns that don't open with keyboard. Modal traps that don't close with Esc. Hidden focus rings.

8. Screen reader compatibility

Test: All non-text content has alt text. Form fields have labels. ARIA roles where appropriate. Test with VoiceOver/NVDA.

Common failures: Decorative SVGs announced to screen readers. Form fields with placeholder-only labels. Generic "click here" link text.

9. Cognitive accessibility

Test: Reading level appropriate for audience. Forms broken into manageable chunks. Time-sensitive operations adjustable or pausable.

Common failures: Dense legal copy. 30-field forms on one screen. Carousels that auto-advance without controls.

10. Motor accessibility

Test: Click/tap targets at least 44×44px (WCAG 2.2 AAA). Drag interactions have keyboard alternatives. No precision gestures required.

Common failures: Tiny mobile buttons. Drag-only sorting. Hover-only menus.

Dimension 3: Performance

11. Largest Contentful Paint (LCP)

Test: LCP under 2.5s on mobile (4G). Use Lighthouse, PageSpeed Insights, or real-user monitoring.

Common failures: Unoptimized hero images. Above-the-fold render-blocking JS. Inefficient web fonts.

12. Cumulative Layout Shift (CLS)

Test: CLS under 0.1. No content jumping during load.

Common failures: Images without dimensions. Late-loading ads. Web fonts causing reflow.

13. Interaction to Next Paint (INP)

Test: INP under 200ms. Interactions feel snappy.

Common failures: Heavy JavaScript on input handlers. Synchronous API calls during typing. Unbatched re-renders.

14. Time to First Byte (TTFB)

Test: TTFB under 600ms. Server response is fast.

Common failures: Unoptimized backend queries. No CDN. Unnecessary server-side rendering for static content.

15. JavaScript bundle size

Test: Initial JS bundle under 300KB compressed. Defer non-critical JS.

Common failures: Loading entire libraries for one feature. Polyfills for browsers nobody uses. Tracking scripts that block render.

Dimension 4: Conversion

16. Funnel completion rate

Test: Identify the primary conversion funnel (signup, purchase, lead capture). Measure step-by-step completion rate.

Common failures: Funnels that lose 70%+ at one specific step (usually a friction-heavy form). No funnel measurement at all.

17. Form friction

Test: Number of fields, validation timing, error recovery, autofill compatibility.

Common failures: 12-field forms when 4 would do. Validation on blur instead of inline. Required fields not marked.

18. Call-to-action clarity

Test: Primary CTA visible above the fold on key pages. Action verb in CTA copy. Visual hierarchy makes primary action obvious.

Common failures: "Submit" instead of "Get my quote". Five competing CTAs on one screen. Primary CTA below the fold on mobile.

19. Empty states and zero-data

Test: Empty states pre-populate, suggest, or guide instead of just saying "nothing here".

Common failures: Bare empty states. No path to first action. Sample data confused with real data.

20. Trust signals

Test: Security badges, certifications, customer logos, reviews, testimonials placed at moments of friction.

Common failures: Trust signals only on the homepage. No social proof at checkout. Reviews not visible on product pages.

Dimension 5: Content and copy

21. Voice and tone consistency

Test: Read 10 random pages or surfaces. Does the brand voice come through consistently?

Common failures: Marketing-speak in product UI. Casual brand voice abandoned in legal copy. AI-generated copy that doesn't match brand.

22. Microcopy quality

Test: Button text, error messages, empty states, tooltips, form labels — are they clear, helpful, human?

Common failures: "Submit", "OK", "Click here". Apologetic error messages with no remediation. Tooltips that just repeat the label.

23. Information hierarchy

Test: Most important information most prominent. Less important information present but secondary.

Common failures: Equal visual weight for all elements. Critical price/CTA buried below the fold. Key facts hidden in long paragraphs.

24. Localization quality

Test: All translations are human-quality. Cultural conventions respected (date formats, currencies, name order).

Common failures: Google Translate output that reads stilted. Hardcoded English in non-English versions. Hreflang errors.

25. Reading level appropriateness

Test: Content matches audience reading level. Use Hemingway or similar tools.

Common failures: B2B SaaS at PhD reading level. Consumer products at university level. Legal copy unreadable to most.

Dimension 6: AI and modern UX patterns

26. AI feature integration

Test: If AI features exist, are they discoverable, contextual, and producing real value?

Common failures: AI chatbot bolted on as a marketing feature. No clear value proposition for AI use. Hallucinations not handled.

27. Command palette

Test: For complex products, is there a Cmd+K command palette? Discoverable, comprehensive, fast?

Common failures: No command palette in apps that need one. Command palette that only does navigation, not actions.

28. Loading and async UX

Test: Long operations have progress indicators, time estimates, and the ability to navigate away.

Common failures: Spinner with no progress information. Operations that block the entire UI. No retry mechanism on failure.

29. Mobile-first reality

Test: On mobile, is the experience first-class — not a stripped-down desktop?

Common failures: Desktop-first design that "responsive-shrinks" to mobile. Touch targets too small. Mobile flows that require horizontal scrolling.

30. Dark mode

Test: Dark mode supported and properly designed (not just inverted colors). Persistent across sessions.

Common failures: No dark mode in apps where users spend hours. Dark mode that breaks color contrast. Auto-switching that overrides user preference.

How to actually run the audit

The audit isn't done in 30 minutes by checking off items. The full process:

1. Setup (1 day): tools (Lighthouse, axe DevTools, BrowserStack, Hotjar/Fullstory if available), access (admin accounts, analytics, error tracking), context briefings with stakeholders.

2. Heuristic and accessibility pass (3-5 days): walk through all major flows and screens. Document issues with screenshots, severity, and recommendations.

3. Performance pass (1-2 days): run Lighthouse, PageSpeed Insights, WebPageTest. Pull real-user data from CrUX or analytics if available. Identify slowest pages.

4. Conversion pass (1-2 days): pull funnel data from analytics. Run heatmaps if available. Identify highest-impact friction points.

5. Content audit (1-2 days): read 20-30 pages or screens. Document voice issues, microcopy failures, hierarchy problems.

6. Synthesis and prioritization (1-2 days): rank all issues by impact × effort. Build a remediation roadmap. Identify quick wins (1-2 weeks) vs. major projects (1-3 months).

7. Presentation (half day): walk stakeholders through findings, demo critical issues live, align on next steps.

What to do with the audit output

The audit is worthless if it sits in a drawer. The strongest follow-up patterns:

• Quick wins sprint (2-4 weeks): fix all S-effort/H-impact issues immediately. Typically 30-50% of identified issues.
• Quarterly remediation for M-effort items, batched into focused 2-3 week sprints.
• Major refactors for L-effort items planned into the product roadmap with priority weighting.
• Re-audit at 6-12 months to measure progress and surface new issues.

---

If you want a UX audit run on your product against this framework, we offer a free 30-minute scoping call to size the engagement. No upsell — just an honest read on whether an audit is the right next investment for you.