The no-nonsense guide to composable commerce
By Steven Van ·
Vercel's guide tells teams to pick one outcome, keep the existing monolith as an API backend, and usually skip middleware.
Composable commerce projects should start with one business outcome and the smallest change that delivers it. That is the argument in a guide from Vercel, which sets out five "truths" for teams moving off a monolithic commerce platform.
The guide names three possible goals: a frontend focus (faster personalised experiences), an integration focus (specialised tools for metrics like acquisition, retention and returns), or a backend focus (fulfilment and inventory). It advises picking one and doing only the work it needs.
- Keep the monolith as the backend. For frontend or integration goals, use the existing platform as an API layer behind a new headless frontend. It stays the source of truth, and marketing and merchandising teams keep their current systems.
- Keep frontend work low effort. For integration or backend goals, avoid redesigning the UX during migration. Vercel suggests AI-assisted tooling to reproduce the existing experience, and says even substantial migrations with unchanged UX can typically finish within weeks.
- Skip the middleware. Frameworks like Next.js can often act as the integration layer. Vercel says middleware may still be warranted in complex, multi-backend setups.
- Simplify. The guide says AI-driven development has lowered the cost of custom solutions, and recommends vendors with open APIs and flexible SDKs.
For a backend-only goal, the guide suggests fully decoupling with a bare-bones frontend first, then replacing the monolith step by step.