Headless commerce is not free — a freelancer's honest ledger
Every headless commerce migration I have run trades one class of pain for another. Here is the ledger I walk clients through before we commit.
Every three months, a founder emails me some variation of the same question: should we go headless? The answer is always the same — it depends on what you are optimising for. Below is the honest ledger I walk clients through before we commit to a migration.
The wins are real
A well-executed headless commerce build usually delivers three things legacy monoliths struggle with: sub-second Time to First Byte on product pages, editorial changes that ship without a backend deploy, and the freedom to plug best-of-breed vendors into search, reviews, and checkout without waiting on a plugin release.
On the last Next.js migration I ran, the Largest Contentful Paint on the top thousand product URLs dropped from 3.9s to 1.4s at the p75, organic traffic recovered inside a quarter, and the merchandising team stopped filing engineering tickets for hero-banner swaps.
The costs are also real
You inherit operational surface area you did not have before. That Next.js app now needs its own CI, its own hosting, its own observability, and its own release process. The checkout is now a contract you are consuming, which means you own the error-handling story when the PSP has a bad day.
Editors lose the WYSIWYG they were used to. Even the best headless CMS (Directus, Sanity, Payload) requires a couple of weeks of onboarding before the marketing team stops missing the old admin.
The three questions I ask before we start
- Is your storefront the bottleneck? If checkout conversion is being killed by page load or a stale front-end, going headless is a genuine lever. If the drop-off is on the payment step, spend that budget on the checkout instead.
- Does your merchandising team ship weekly? If they publish a promotion every Monday and every deploy is a Slack negotiation with engineering, the deploy-decoupling win pays for the whole migration inside a year.
- Do you have on-call capacity for two apps? Because you will have two apps to page — the storefront and the back-office. Plan for that before you sign the vendor contract.
Where I usually land
Most B2C brands under about $10M annual gross merchandise value are better off tuning the monolith they already have. Above that, the merchandising velocity and performance wins usually justify the operational cost — but only if you are ready to run two apps in production, not one.
If you are somewhere in the middle and want a written second opinion, that is exactly what my consultation and audit engagement produces.