We are a headless CMS migration agency. We take sites that rank well and are hard to change off WordPress, Drupal, AEM or Sitecore, move the content into a headless CMS and rebuild the frontend. We own the whole project: the audit, the content model, the import scripts, the URL and 301 map, the cutover and the weeks after launch, so search traffic survives the switch. We also run frontend rewrites, framework upgrades and hosting moves when they come with the same risks.
This page covers how we run a migration as a project with a fixed scope. If you want to add a developer to your own team instead, see how to hire headless CMS developers. If you are replatforming a live site on any stack and want the full process with deliverables, read our website migration agency page. If you are still deciding whether to go headless at all, start with headless CMS development.
Which migrations do we run?
CMS replatforms
This is most of our work. A team has outgrown WordPress plugins, Drupal modules or an AEM licence, and editors wait on developers to publish. We move the content into Payload, Strapi, Sanity, Contentful or Storyblok and rebuild the frontend, usually in Next.js. The content model gets designed first. Import scripts come after it.
Frontend rewrites
Sometimes the CMS is fine and the frontend is the problem: a PHP theme that renders every page on request, or a single page app that search engines struggle to index. We rebuild the frontend with static generation or incremental regeneration and keep the CMS behind it. Headless WordPress is the common case: WordPress stays as the editing backend and only the frontend is rebuilt.
Framework upgrades
Vue 2 and Nuxt 2 have reached end of life, and Drupal 7 has too. Next.js apps on the Pages Router keep running but miss newer features. An upgrade looks like routine maintenance, yet it often touches routing, data fetching and rendering, so it needs the same URL checks as a replatform.
Hosting moves
A move to a new host with the same CMS and the same files is a job for the new host’s transfer service, and we say so. We take hosting moves when they come with a rebuild: a frontend moving to a CDN, a self-hosted CMS moving into containers, or a site leaving a managed platform plan.
Where are you moving from?
| From | To | Our page |
|---|---|---|
| WordPress | Payload, Strapi, Sanity, Contentful or Storyblok | WordPress to headless CMS migration |
| WordPress, keeping it as backend | Next.js frontend over WPGraphQL or the REST API | Headless WordPress development |
| Drupal 7 or an older Drupal | A current Drupal version, or a headless CMS | Drupal migration services |
| AEM on premises or 6.x | AEM as a Cloud Service, or a headless CMS | Adobe Experience Manager migration |
| Sitecore XP | A headless CMS, or Sitecore XM Cloud | Sitecore XM Cloud and headless work |
| Server-rendered or SPA frontend | Next.js or Astro with static generation | Next.js development |
| Any platform, new domain or URLs | Same content, new structure | SEO-safe website migration |
Two comparisons help before you commit to a route: WordPress vs headless CMS for teams on WordPress, and AEM vs headless CMS for teams paying for Adobe.
How does a migration run?
Every project follows the same seven steps. Each ends with a sign-off from a named person on your side.
- Audit. We crawl the site, export traffic and rankings from analytics and Search Console, and list every template, content type, plugin and integration. This becomes the baseline we measure against.
- Content model mapping. Each old content type is mapped field by field to the new schema. Fields nobody uses are dropped here, not after launch.
- URL and 301 map. Every old URL gets a target: keep, redirect or retire. We keep URLs that earn traffic and give the rest one-hop 301s.
- Content migration scripts. We write import code against the old database or API, run it on a copy of production data, and produce a diff report for editors. Nobody copies content by hand.
- Parallel run. The old site stays live while the new one runs on staging with real content. Editors try their usual publishing tasks there, and we crawl it against the old site.
- Cutover. A short content freeze, a final import, then a DNS or routing switch in a planned window, with a written rollback plan.
- Post-launch watch. We track Search Console coverage, server logs, rankings and conversions against the baseline and fix gaps as they show up.
The headless CMS migration checklist lists the same gates if you plan to run the project in house.
How do we protect SEO during a migration?
Keeping a URL is safer than redirecting it. Where a URL must change, it gets a one-hop redirect that lives in version control and is tested against a crawl of the old site. Links inside body copy are rewritten to the new URLs, so the site does not send its own visitors through redirects.
Titles, descriptions, canonicals, structured data and hreflang come over template by template, and a crawl diff compares them with the old site before launch. Analytics tags, events and conversion goals go into the QA plan, because they break without anyone noticing.
We also measure Core Web Vitals before and after. A migration that makes pages slower has failed, whatever the CMS.
How to choose a headless CMS migration agency
Most migration agencies can build a new site. Fewer can move a ranking site without losing traffic. Ask these questions before you sign with anyone, us included.
- Do they start with an audit and a baseline? A quote given before anyone has crawled the site and read Search Console is a guess. Ask what they will measure before and after launch.
- Who writes the URL and 301 map, and when? It should exist before any template is built, live in version control and be tested against a crawl of the old site.
- Is content moved by scripts or by hand? Import code that can run again after a fix, with a diff report for editors, is the only way to move thousands of pages and their media reliably.
- Can they show the content model first? The new schema decides how editors work for years. Ask to see a field by field mapping from your old content types.
- What is the rollback plan? A cutover without a written way back is a bet. Ask how long a rollback takes and who decides to use it.
- How long do they stay after launch? Coverage drops and broken redirects show up in the weeks after cutover, not on launch day. Ask who watches Search Console and logs, and for how long.
- Have they moved a site like yours? Ask for a migration with a similar platform, page count and number of languages, and for the numbers before and after.
What drives the risk and the cost?
We do not publish a price list, because two sites with the same page count can need very different work. Every project is scoped after a short audit and quoted as a fixed scope. These are the drivers we look at.
| Driver | Why it matters |
|---|---|
| Templates | Each template is a frontend component to build and a set of fields to map |
| Content model | Custom fields, page builder markup and shortcodes take the most import work |
| Changed URLs | Every changed URL needs a redirect, a test and a place in the watch list |
| Languages | Each locale adds hreflang, routing and review cycles |
| Integrations | Forms, CRM, search, analytics and single sign-on each need rewiring and testing |
| Design changes | A redesign done during the move adds scope and makes traffic changes hard to read |
| Sign-off speed | The content freeze date and how fast decisions come back set the end date |
The biggest single risk is changing too much at once. We keep the URLs, the structure and the design where we can, move the platform, and leave improvements for after the new site has proved itself.
Proof from recent migrations
Drupal 7 to Next.js and Payload for a European enterprise healthtech organization. The site carried over 1,200 content nodes across four languages and had to leave Drupal 7 before its end of life. A TypeScript pipeline pulled every node through the Drupal REST API with checksum validation, and locale-aware routing kept the URL structure and hreflang tags. LCP went from 4.3s to 1.5s, editor publish time fell by about 50%, and there were 0 SEO ranking regressions at cutover. Read the Drupal replatform case.
WordPress to Payload for a US-based Series-B fintech. Roughly 40 custom plugins sat on the old site, and every copy change needed a developer review. All existing permalink structures were carried forward, and compliance approvals moved into the CMS. Time to publish fell by about 65%, LCP dropped from 4.2s to 1.4s, and organic traffic held through cutover. See the fintech replatform case.
Start with an audit
Send us the site, the platform you have in mind and your launch window. We crawl it, record a baseline and come back with a fixed scope, a dated plan and the risks we see. Book a migration assessment.