全部技能
Web Accessibility (A11y)
⬢ 層級 3領域
中等
薪資影響
12 個月
學習時長
中等
難度
12
職業
一覽
Web accessibility (a11y) makes products usable for the 1B+ people with disabilities: vision, hearing, motor, cognitive. Built on WCAG 2.1 AA standard (50 success criteria), enforced by ADA lawsuits in the US (10k+ filed in 2023), the European Accessibility Act (mandatory June 2025), and gov/enterprise procurement. Career path: A11y-aware Frontend ($95-130k) → A11y Specialist ($130-170k) → Accessibility Lead ($170-220k). The gap between 'knows ARIA roles' and 'can audit + remediate a 1000-page enterprise site' is worth $40-60k.
什麼是 Web Accessibility (A11y)
Build accessible products: screen readers, keyboard navigation, WCAG compliance. Serve 1B+ people with disabilities. Legal requirement for many companies. Learning Curve: Medium (empathy + technical standards)
🔧 工具與生態
axe DevToolsLighthouseWAVEPa11yNVDAJAWSVoiceOverTalkBackStarkAccessibility InsightsDeque axeStorybook a11y addon
💰 各地區薪資
| 地區 | 初級 | 中級 | 高級 |
|---|---|---|---|
| USA | $95,000 | $135,000 | $185,000 |
| UK | £50,000 | £75,000 | £110,000 |
| EU | €55,000 | €80,000 | €115,000 |
| CANADA | CA$100,000 | CA$140,000 | CA$190,000 |
🎓 認證
🎯 使用 Web Accessibility (A11y) 的職業
❓ 常見問題
Why does a11y matter beyond 'being nice'?
Three reasons: (1) Legal, ADA lawsuits target accessible-failed sites; settlements run $25k-$500k+. EAA in EU is mandatory June 2025 for digital products. (2) Procurement, gov, edu, big enterprise require VPAT (a11y conformance doc) before signing. No VPAT = locked out of $$ contracts. (3) Reach, 15% of users globally have a disability; another 30% benefit from a11y features (captions, dark mode, keyboard nav). a11y is market expansion, not charity.
What's the difference between WCAG A, AA, and AAA?
A = bare minimum (alt text, captions, keyboard access), no one settles for this. AA = the legal/professional standard (color contrast 4.5:1, focus indicators, error identification), what ADA, EAA, Section 508 require. AAA = aspirational (contrast 7:1, sign-language for video), most sites can't meet AAA across the board. Target AA + AAA on individual criteria where you can. AA is what hiring managers mean when they say 'WCAG-compliant.'
How do I test a11y as a frontend dev without a screen reader?
Three layers: (1) Automated: axe DevTools or Lighthouse catches ~30% of issues, color contrast, missing labels, ARIA misuse. Run on every PR. (2) Manual keyboard: Tab through every flow, can you reach everything? Esc close modals? (3) Screen reader: NVDA + Firefox on Windows or VoiceOver + Safari on Mac, both free. Spend 1 hour/week navigating your app with eyes closed. The 70% automated tools miss = the experience for real users.
ARIA, when do I use it vs native HTML?
Rule #1 of ARIA: 'don't use ARIA.' Native HTML elements (button, label, input, dialog) come with accessibility built in. Use ARIA only when no HTML element does the job (custom widgets like comboboxes, tabs, trees). Misused ARIA is worse than no ARIA, wrong roles confuse screen readers more than missing roles. Memorize the 5 ARIA rules from the W3C ARIA Authoring Practices Guide before adding role= or aria-* anywhere.
What's a VPAT and why do enterprise clients ask for one?
VPAT (Voluntary Product Accessibility Template) = standardized doc declaring how your product conforms to WCAG/Section 508. Procurement at Fed gov, universities, big banks and large enterprises often requires it before purchase. Honest VPATs (with known gaps + remediation roadmap) win deals; bullshitted VPATs lose contracts when audited. Generating a VPAT requires a real audit ($10-30k for an enterprise app), outsource to Deque, Level Access, or TPGi for legal-grade output.
Can I rely on overlay tools (AccessiBe, UserWay)?
No. Overlay tools (the 'a11y widget on the corner') promise instant compliance but ship broken: they break screen readers, miss 70% of real issues, and have been sued (AccessiBe lost ADA suits 2022-2024). Real a11y is built into the product, semantic HTML, focus management, proper labels, not bolted on. Avoid overlays in resumes/projects; sophisticated hiring managers see them as a red flag.
How does AI/automation change a11y in 2026?
Two angles: (1) AI helps, auto-generated alt text (with human review), real-time captions (Otter, Otter, Zoom), AI image-recognition for VoiceOver users. (2) AI hurts, generative UI (LLM-built components) often skips labels, focus, semantics. New a11y job in 2026 = AI-assisted UI auditing: catching what the AI generator missed. Demand for human a11y experts is growing because 80% of LLM-generated UI code fails axe out of the box.
不確定這項技能適不適合你？
做一下職業匹配，我們會推薦合適的方向。
找到最適合我的技能 →找到理想的職業方向
在 2,521 個職業中進行技能匹配。免費。
免費做職業匹配 →