Research institute · CMS Migration
A CMS migration that didn't make the editors relearn their job.
The situation
A research institute had grown into a tangle of legacy systems. The editorial team published thousands of pages of research, analysis, and case data, but the CMS underneath had been bolted onto for over a decade. Adding a new content type meant adding new fields by hand across dozens of templates. Editors were doing structural work that should have been platform work.
What we found
The editors needed to keep publishing during the migration. They held important knowledge about the content and its relationships, so their working process helped shape the new system.
We began by defining the content types and relationships the platform would need to support.
What we built
We mapped the existing content into a clean structured model: research artifacts, case data, expert profiles, classifications, and the relationships between them. Then we migrated to Sanity, a headless CMS that supports structured content natively. The editorial UI was custom-built to match the team's existing mental model, not the CMS's defaults. Workflows that took six clicks before took two. Field-level versioning and a clear approval path replaced the email-based revision flow.
The migration ran in parallel with live publishing. The editorial team never had a day where their old tool went away before the new one was ready.
What changed
The team kept their daily rhythm and got faster instead of slower. New content types ship in hours, not weeks. The structured model became the foundation for downstream uses: an API for partner sites, a search index keyed to research types, and content syndication that doesn't require manual export.