Headless CMS: Strapi, Directus, Sanity, Contentful, Drupal

Traditional CMS works fine until a designer says "I want scroll animation with parallax," frontend says "we need React," and an SEO specialist says "why is TTFB 3.4 seconds." At that moment, monolithic architecture becomes a problem for everyone.

Headless CMS separates content management from presentation. Editors work in a convenient interface, developers get data through an API and build the frontend on any stack. Sounds simple. In practice, CMS selection, data modeling, and API setup take a significant portion of the project.

Choosing the right tool for the task

There is no universal headless CMS. Choice depends on the team, content complexity, and infrastructure.

Strapi — open-source, self-hosted, Node.js. Suits teams that need control over their data and API customization capabilities. Plugin architecture allows adding custom routes, middleware, lifecycle hooks. REST and GraphQL out of the box. Weak point — Strapi v4 and v5 are incompatible, migration is painful.

Directus — also open-source, but different approach: doesn't generate schema, but wraps an existing database (PostgreSQL, MySQL, SQLite) into REST/GraphQL API. If the database already exists — Directus connects to it without migrations. Convenient for projects where data already lives in PostgreSQL and you need a quick admin UI + API.

Sanity — cloud-based CMS with real-time editor. Distinctive feature — GROQ (Graph-Relational Object Queries), its own query language, more powerful than REST for complex relationships between documents. Portable Text for structured content. Suits media, publishing, marketing sites with non-standard editorial workflows.

Contentful — enterprise cloud CMS. Strong point — localization (up to 1000 locales), rich SDK for all platforms, Contentful Apps for custom UI. Weak point — price when scaling and limited flexibility of data models compared to open-source alternatives.

Drupal — not headless in pure form, but with JSON:API and GraphQL modules turns into a powerful API-first backend. Strong point — maturity (Drupal 10), granular access control, enterprise clients (NASA, weather.com). Entry barrier is high, but for complex government or corporate portals there are few alternatives.

How we build projects on headless CMS

Content modeling — critical stage. A mistake here is expensive. Typical mistake: a single body field of type rich text for everything. After half a year, the content manager wants to insert video between paragraphs, add a pull quote with custom styling, embed an interactive table. Rich text doesn't allow this. Solution — Portable Text (Sanity) or custom components in Strapi/Directus through Dynamic Zone.

Frontend. Headless CMS almost always comes with Next.js or Nuxt. For Contentful and Sanity — App Router with ISR: pages are statically generated at build, updated via revalidatePath() when content changes through webhook. For Strapi/Directus with frequent data updates — SSR with cache: 'no-store' or SWR on client.

Case study: corporate site redesign with Strapi + Next.js. Client — manufacturing company, previous site on WordPress with ACF, 200+ pages, 4 languages. Problems: TTFB 3.8s, editors complained about slow admin.

Switched to Strapi 4 (self-hosted on VPS, PostgreSQL), Next.js 14 App Router. Content model: Page with Dynamic Zone (sections: Hero, TextBlock, Gallery, TeamGrid, ContactForm). Localization via Strapi i18n plugin + next-intl on frontend. Frontend deployment on Vercel with ISR, revalidation via Strapi webhook on entry.publish.

TTFB fell from 3.8s to 180ms (static from CDN). Editors got a clean interface without 47 plugins.

Process and timelines

Standard path: audit content needs → data schema → CMS and API setup → frontend development → content migration (if any) → API endpoints testing → deployment.

Migration from WordPress to headless CMS takes as much time as the project itself — often more. Especially if WordPress accumulated custom fields through ACF with non-standard structure.

Cost — after analyzing requirements and auditing existing infrastructure.