
Adding WCAG 2.2 AA compliance to a typical SaaS in 2026 costs $8,000 to $150,000+ depending on what you ship. A small marketing site runs $8,000 to $25,000. A mid-size SaaS app with dashboards, forms, and modals runs $40,000 to $80,000. A complex multi-product enterprise platform runs $100,000 to $250,000+. Audit costs ($5,000 to $20,000 from firms like Deque or Level Access) sit on top of remediation.
The cheaper number assumes you build accessibility in at the component layer from day one. The bigger numbers are what you pay to retrofit it later. That gap (often 5x) is the most important budget decision a founder makes about accessibility.
The European Accessibility Act (EAA) enforcement date was June 28, 2025, and the first wave of penalties has landed in Q1 2026. Any SaaS sold into the EU (including B2B SaaS billed in USD if EU customers can sign up) is now in scope. Penalties vary by member state, but Germany alone allows fines up to €100,000 per violation per product.
US litigation has continued climbing. Roughly 4,500 ADA Title III digital lawsuits were filed in 2024, and Q1 2026 is tracking higher. Most settle for $10,000 to $50,000, but the remediation work that follows usually costs 3x to 10x the settlement.
The point is simple: WCAG 2.2 AA is no longer a "nice to have" for SaaS founders selling to mid-market or enterprise. It is a procurement gate (most enterprise security questionnaires now ask for a VPAT or ACR) and a litigation surface.
Almost every legal mandate (ADA, EAA, Section 508, AODA) targets WCAG 2.2 AA, not AAA. AAA exists, but it requires sign-language interpretation for prerecorded audio, 7:1 color contrast, and reading-grade restrictions that make most real product copy non-compliant.
Stop at AA. Hitting AAA on a SaaS app typically doubles the budget for zero additional legal protection and zero additional addressable market. The W3C itself explicitly recommends against requiring AAA across an entire site.
The only common exception: government tenders in specific jurisdictions (some Canadian provinces, some EU public sector RFPs) ask for AAA. If you're not chasing those, AA is the ceiling.
Founders who say "add WCAG" usually mean five distinct workstreams priced separately: code-level remediation (semantic HTML, ARIA, focus management, keyboard support), design and content fixes (contrast, alt text, captions), audit (axe-core plus manual review by trained auditors), VPAT / ACR documentation ($3,000 to $10,000 outsourced), and ongoing compliance (CI checks, backlog, annual re-audit). Founders who only budget code-level remediation end up at 60% of their real cost.
| Approach | Cost (small marketing site) | Cost (mid SaaS app) | Timeline | Pros | Cons |
|---|---|---|---|---|---|
| US full-time accessibility engineer | $180k+/yr salary | $180k+/yr salary | 3 months to hire, 6+ months to ship | Owns it long-term | Overkill unless you have multiple products |
| Accessibility consultancy (Deque, Level Access, TPGi) | $15k audit + $30k remediation | $20k audit + $80k–$150k remediation | 8–16 weeks | Deep expertise, defensible audit reports | High hourly rate ($250–$400), slow procurement |
| Generalist dev agency | $10k–$20k | $40k–$100k | 6–12 weeks | Familiar engagement model | Most don't know WCAG well; quality varies |
| Freelancer (Upwork / Toptal) | $5k–$15k | $25k–$60k | 4–10 weeks | Cheaper hourly | Hard to vet expertise; no audit deliverable |
| Cadence | $500–$2,000/wk | $500–$2,000/wk × 8–16 wks | 48-hour trial then ship | Weekly billing, replace any week, AI-native by default | Not a substitute for a formal audit firm if you need a notarized VPAT for enterprise procurement |
The Cadence row is the unbundled option: book a senior engineer to do the remediation work weekly, then pay a specialist firm separately for the formal audit and VPAT if you need one for procurement. That combination usually lands at 40 to 60 percent of an end-to-end consultancy quote.
A WCAG 2.2 AA audit from one of the well-known firms breaks down roughly like this in 2026:
| Scope | Deque | Level Access | TPGi | Boutique firms |
|---|---|---|---|---|
| Single marketing page | $2,500–$5,000 | $3,000–$6,000 | $3,000–$5,000 | $1,500–$3,000 |
| Small site (10–25 pages) | $8,000–$15,000 | $10,000–$18,000 | $8,000–$14,000 | $5,000–$10,000 |
| Mid SaaS app (key flows) | $15,000–$30,000 | $18,000–$35,000 | $15,000–$28,000 | $10,000–$20,000 |
| Full enterprise audit | $40,000–$100,000+ | $50,000–$120,000+ | $40,000–$90,000+ | $25,000–$60,000 |
What you get for that money:
You can DIY most of this with axe DevTools, WAVE, and Lighthouse, but you cannot DIY the legal defensibility. If you're ever in court, "we ran Lighthouse" is not a defense. "Deque audited us in March" is.
This is the single most expensive misunderstanding in accessibility budgeting. Axe-core, Lighthouse, WAVE, and the rest catch roughly 25 to 35 percent of WCAG failures. Deque's own published research puts axe at about 57% (best in class), but that number assumes you're running it on every page, every state, including modals, error messages, and focus traps.
What automated tools miss:
role="link" and onclick handler still passes axe)A senior engineer with a screen reader (NVDA on Windows, VoiceOver on Mac, JAWS for enterprise) catches the other 70%. Plan for at least 40 hours of manual keyboard and screen-reader testing per major user flow. That's the number most cost calculators leave out, and it's why DIY accessibility projects always overshoot.
The biggest cost lever in accessibility is when you add it.
At build time, accessibility costs roughly 5 to 10 percent of feature development. You pick a component library that's already accessible (Radix UI, Headless UI, React Aria), you write semantic HTML, you test with a keyboard as you go. A senior engineer adds about 4 hours per feature for accessibility work.
As a retrofit, the same work costs 5x to 10x more. Why:
<div onclick> need to become real <button> elements (often a cascade of CSS rewrites)Concrete example: rewriting a custom date picker to be accessible from scratch takes a mid engineer about 8 hours. Retrofitting an existing custom date picker that has been in production for 18 months, with three other features hooked into it, takes 40 to 80 hours.
This is the same dynamic as adding security or observability after the fact, and the cost calculus is identical to what you see when migrating database engines: the migration itself is small; the surface area you have to touch is the multiplier.
If you're early stage, the cheapest accessibility decision you'll ever make is picking Radix, React Aria, or a similar headless library this week.
Real numbers for a typical mid-size React or Next.js SaaS app, assuming a senior engineer at $150/hour blended rate (or 1 week of a Cadence senior at $1,500):
| Feature | Hours | Cost (consultancy at $300/hr) | Cost (Cadence senior at $1.5k/wk) |
|---|---|---|---|
| Navigation and menus | 8–16 | $2,400–$4,800 | 0.5 week ($750) |
| Forms (login, signup, settings) | 24–48 | $7,200–$14,400 | 1–1.5 weeks ($1,500–$2,250) |
| Modal / drawer focus management | 16–24 | $4,800–$7,200 | 0.5 week ($750) |
| Data tables (sortable, paginated) | 24–40 | $7,200–$12,000 | 1 week ($1,500) |
| Charts and dashboards | 40–80 | $12,000–$24,000 | 1.5–2 weeks ($2,250–$3,000) |
| Rich text editor | 40–60 | $12,000–$18,000 | 1.5 weeks ($2,250) |
| Date / time pickers | 16–32 | $4,800–$9,600 | 0.5–1 week ($750–$1,500) |
| File upload UI | 8–16 | $2,400–$4,800 | 0.5 week ($750) |
| Notifications and toasts | 8–16 | $2,400–$4,800 | 0.5 week ($750) |
| Video player + captions | 16–32 + caption costs | $4,800–$9,600 | 0.5–1 week |
| Color contrast audit + fixes | 16–24 | $4,800–$7,200 | 0.5 week ($750) |
| Image alt text (across product) | 8–24 | $2,400–$7,200 | 0.5 week ($750) |
A mid-size SaaS with most of those features lands at roughly 220 to 410 engineering hours of remediation. That's $66,000 to $123,000 at consultancy rates, or about 6 to 10 weeks of a Cadence senior at $9,000 to $15,000 in labor (plus a separate $15,000 to $25,000 for the formal audit and VPAT).
Five moves cut total accessibility spend by 40 to 70 percent without compromising compliance:
A practical 3-step plan, sized for a SaaS doing $500k to $10M ARR:
That sequence puts you at roughly $25,000 to $50,000 total for a mid-size SaaS to reach defensible WCAG 2.2 AA, versus $80,000 to $150,000 for an end-to-end consultancy engagement. Same outcome, half the cost, faster timeline.
The pattern (book the building work weekly, buy the certification separately) is the same playbook founders use to cut admin dashboard build costs or scope a Shopify app cheaply. Unbundle commodity certification from custom engineering work, then pay market rate for each piece.
If you're already getting RFPs that ask for a VPAT, the slowest path is recruiting a full-time accessibility engineer (3+ months). The fastest path is booking a senior on Cadence this week, shipping remediation in 8 weeks, then commissioning a scoped audit. See what a senior costs on Cadence before you sign any consultancy SOW.
For a small marketing site, 2 to 4 weeks of engineering. For a mid-size SaaS app, 8 to 16 weeks if you work in parallel. Big-bang releases take longer; weekly shipping is faster because regressions surface as you go.
The ADA Title III applies to "places of public accommodation" and courts have repeatedly applied it to websites. Roughly 4,500 digital ADA lawsuits were filed in 2024. WCAG 2.1 AA is the de facto standard courts reference, with WCAG 2.2 AA as the current ceiling. You're exposed even if you never sell into the EU.
No. Automated tools catch about 25 to 35 percent of WCAG failures. The rest requires manual keyboard testing, screen-reader testing (NVDA, VoiceOver, JAWS), and judgment about whether alt text actually describes the image. Lighthouse is necessary but nowhere near sufficient.
A VPAT is the template; an ACR (Accessibility Conformance Report) is a VPAT that's been filled out for your specific product. Enterprise procurement teams ask for both terms interchangeably. Expect $3,000 to $10,000 if you commission one from a specialist firm, or self-author one for free if you trust your internal audit.
Contract for the initial remediation, then keep one in-house engineer as the long-term owner who spends ~4 hours a week on reviews. Full-time accessibility specialists make sense at 50+ engineers or for regulated products. Below that, the workload is too lumpy to justify the salary.
Yes, meaningfully. Cursor and Claude Code can refactor inaccessible components, generate ARIA patterns, write alt text, and draft VPAT prose. Expect 2x to 3x speed on repetitive remediation work. Judgment calls still need a human. Every engineer on Cadence is AI-native by default; the speed compounds across an 8-week project.