Migrate Optimizely CMS 12 to Next.js + Supabase

Enterprise teams are leaving Optimizely CMS 12—the PaaS evolution of what was once Episerver—in unprecedented numbers. It bundled CMS, commerce, and personalization into a managed Azure environment, and that worked for a while. But it's 2026 now: monolithic .NET roots, escalating licensing tiers, and a shared-infrastructure model that creates performance ceilings your engineering team literally can't break through.

We've migrated enterprise teams off Optimizely DXP who were paying $200K–$500K annually in licensing alone. They needed multi-channel delivery, sub-300ms TTFB, and the ability to hire React developers instead of .NET specialists. The math stopped working.

The Real Pain Points with Optimizely DXP

Licensing That Scales Against You

Optimizely's tiered pricing—Group, Corporate, Enterprise—charges per environment and per Web App. Add commerce modules, personalization add-ons, and custom integrations, and you're looking at six-figure annual renewals that creep up every contract cycle. When your marketing team wants a new microsite, the budget conversation starts with licensing. Not design. Licensing.

Performance Ceilings Baked Into the Architecture

Optimizely DXP runs on shared Azure PaaS infrastructure. Your Integration, Preproduction, and Production environments share a single Web App and search index. Real-world Lighthouse mobile scores land between 45–65 for most Optimizely sites we audit. The SCA case study—widely cited by Optimizely themselves—showed improvement from 43 to 70. That's progress, sure. But it's not competitive. Modern headless builds hit 95–100 consistently.

Developer Experience Friction

The .NET-only templating system means your frontend developers can't use the tools they actually know. Every UI change requires a full-stack .NET deployment cycle. Try hiring senior .NET CMS developers in 2026—that talent pool is shrinking while React/Next.js developers are everywhere and cheaper.

Migration Tooling That Creates More Problems

Optimizely's own Project Migration tool for CMS 11→12 upgrades shows the cracks. Spinning up new Linux Docker DXP instances, copying blobs via Azure Storage Explorer with temporary SAS links, database copies that take 20+ minutes for large sites—and that's just moving between Optimizely versions. Think about what a full exit looks like without the right agency.

Vendor Lock-In by Design

Content models, block structures, page hierarchies, taxonomy trees—all of it lives in proprietary formats. There's no "Export to JSON" button. Every year you stay, the switching cost grows. That's not accidental.

What You Get with Next.js + Supabase

The target architecture is a composable stack: Next.js 15 on Vercel for the frontend, Supabase (PostgreSQL + Auth + Storage + Edge Functions) as the backend and content layer, and Cloudflare for edge caching and DDoS protection.

Performance That Wins Core Web Vitals

Next.js App Router with ISR (Incremental Static Regeneration) delivers sub-300ms TTFB globally. Static pages generate at build time; dynamic content revalidates on demand. Lighthouse mobile scores of 95–100 are the baseline, not the aspiration.

Open-Source Economics

Supabase Pro at $25/month replaces a $200K/year DXP license. Vercel hosting at $20/site replaces $10K+/month Azure PaaS fees. Three-year TCO drops from $900K+ to roughly $300K, including the migration itself. ROI turns positive within 12 months.

Real Headless, Not Headless-Adjacent

Optimizely CMS 13 marketed headless capabilities, but it's still PaaS-first with an API bolted on. Supabase is API-native. Every piece of content is accessible via REST or GraphQL, secured with row-level security policies. Your mobile app, your web app, your kiosk—same content API, zero additional licensing.

Modern Developer Workflow

React components, TypeScript, Git-based deployments, preview environments per PR, edge functions for server logic. Your team ships features in days, not sprint cycles. And the hiring pipeline opens up to the largest frontend developer community in the world.

Our Migration Process

Aryan Shah, our Platform Migrations Lead, has built a repeatable playbook for Optimizely DXP exits that preserves every piece of content, maintains SEO equity, and achieves zero downtime.

Phase 1: Deep Audit (Weeks 1–3)

We map your entire Optimizely content model—page types, block types, content areas, taxonomy trees, media assets, URL structures, and integration touchpoints. We export a complete inventory and flag content that needs restructuring versus content that maps 1:1 to Supabase schemas.

Phase 2: Schema Design & ETL Pipeline (Weeks 4–8)

Custom Node.js ETL scripts extract content from Optimizely's database and transform it into Supabase PostgreSQL tables. Taxonomy hierarchies are preserved with recursive table structures and RLS policies that mirror your existing RBAC permissions. Media assets migrate from Azure Blob Storage to Supabase Storage. We've handled catalogs of 50K+ assets in under an hour using parallel upload pipelines.

Phase 3: Frontend Build (Weeks 6–14)

Next.js 15 App Router frontend, built component-by-component against your existing design system. We don't just replicate—we optimize. Every Optimizely block type becomes a React Server Component. Navigation, search, and filtering move to Supabase Edge Functions, replacing Optimizely's shared search index with dedicated, fast infrastructure.

Phase 4: SEO Preservation (Parallel Track)

This is where migrations fail. It's where we don't. Every URL gets mapped and 301 redirects are configured at the edge via Cloudflare. XML sitemaps regenerate automatically from Supabase content. Structured data (JSON-LD) is rebuilt per page type. We monitor Google Search Console throughout the migration and for 90 days post-launch to catch any indexing anomalies.

Key SEO actions:

• Complete URL audit and redirect mapping
• Canonical tag preservation
• Internal link structure validation
• Meta title/description migration with programmatic verification
• Schema markup rebuild
• Robots.txt and sitemap.xml regeneration

Phase 5: Zero-Downtime Cutover (Week 15–16)

Blue-green deployment. The new Next.js + Supabase stack runs in parallel on Vercel while the Optimizely site stays live. Final content sync runs 2 hours before cutover. DNS switches via Cloudflare with instant propagation. Traffic shifts with zero downtime. We keep the old environment warm for 30 days as a safety net.

Phase 6: Post-Launch Optimization (Weeks 17–20)

Performance monitoring, CWV tracking, editor training on Supabase Studio (or a lightweight CMS layer like Payload CMS if your editors need a visual interface), and handoff documentation.

Timeline and Investment

A typical enterprise Optimizely DXP exit takes 14–20 weeks from audit to launch. Pricing ranges from $150K–$350K depending on content volume, number of sites, integration complexity, and localization requirements.

For context: your next Optimizely DXP renewal is probably $200K+. The migration pays for itself before year two.

Who This Is For

Enterprise teams running Optimizely CMS 12 (or legacy Episerver) who are staring down a renewal decision, fighting performance problems, or just can't ship fast enough. If your engineering team spends more time wrestling the platform than building features, it's time to go.

We don't do generic re-platforms. Aryan and the team know Optimizely's internals cold—the content model, the blob storage quirks, the deployment pipeline limitations. That's why we can guarantee taxonomy preservation and zero downtime. This isn't our first exit.