Headless commerce: when it's worth it, when it isn't

Headless commerce splits your storefront from your commerce backend. The Shopify or WooCommerce backend keeps managing products, orders, inventory, and payment. A custom frontend (usually Next.js or Astro) renders the actual experience to users.

It sounds clean. Sometimes it is. Often it's overkill.

The honest case for headless

Headless wins on these axes:

Performance. A well-built Next.js storefront easily hits 95+ Lighthouse scores. Shopify's themed Liquid frontend rarely does.
Editorial flexibility. You're not fighting a theme. You're building exactly what the brand needs, pixel by pixel.
Animation and interaction. Real React + Framer Motion, not third-party Shopify section apps.
Multi-channel. One product catalog can power web, native app, kiosk, marketplace. Hard to do with a themed setup.

When clients show up with content-heavy brand stories (luxury, beauty, fashion, lifestyle), headless is often the right call. Stripe Press, Vercel's site, Aesop's web presence: they all live in this lane.

The honest case against headless

The wins above come with costs that scare off most mid-market brands:

Engineering overhead. A themed Shopify store can be edited by a marketing manager with no dev background. A headless Next.js store needs someone who knows React every time you want to ship a new section.
App ecosystem doesn't "just work". Half the Shopify apps assume a Liquid storefront. Headless means you reimplement or skip them.
Three systems to maintain. Shopify config + your Next.js codebase + your CMS (if you use a separate one for content). Versus one Shopify theme.
Cost. Custom headless build is typically 3–5x a themed Shopify build of similar visual quality.

The framework: should you go headless?

We use a simple rubric. Score each axis 1–3, where 1 = "headless is overkill" and 3 = "headless is justified."

Revenue scale

1. Under $2M/yr (headless ROI doesn't materialize)
1. $2M – $10M (depends on the rest)
1. $10M+ (engineering overhead is a rounding error)

Content density

1. Catalog-driven, minimal editorial (think: hardware, supplements, basics)
1. Moderate editorial (lookbooks, blog, product stories)
1. Editorial-first brand (think: magazines, luxury, brand-as-content)

Team capacity

1. Founder + marketing manager, no engineers
1. One in-house or contract dev
1. Real engineering team or agency-on-retainer

Performance sensitivity

1. Slow site doesn't materially hurt (B2B, niche, captive audience)
1. Speed helps but isn't the bottleneck
1. Mobile Core Web Vitals are make-or-break (mass-market DTC)

Total it up:

4–6 points. Stay themed. Headless will sink resources for no return.
7–9 points. Probably themed, possibly headless for specific reasons. Worth a deeper conversation.
10–12 points. Headless is likely the right call. Build it well.

A real example

Recent client. Apparel brand, ~$4M revenue, growing 80% YoY. They came in convinced they needed headless because their current site felt "slow and limiting."

We scored them:

Revenue: 2
Content density: 2 (catalog-heavy, a few campaigns per season)
Team capacity: 1 (no engineers, marketing-led)
Performance sensitivity: 3 (mobile-first audience, paid acquisition)

Total: 8. Probably themed.

What they actually needed:

A faster theme (we rebuilt on a custom Shopify theme, not headless)
Better Klaviyo flows
Image optimization that was being skipped by the old theme

Result: 30% faster mobile, +18% checkout conversion, no headless rebuild required. Build cost: ~25% of what a headless project would have been.

When we recommend headless

Concretely, in the last twelve months, we built headless for:

A beauty brand with deep editorial content + subscription mechanics
A luxury home goods brand that wanted maximum brand control
A SaaS-adjacent product company that needed mobile and web parity

In each case, the rubric was 10+ and the team was equipped to maintain it.

TL;DR

Headless is a power tool. It's the right answer ~10% of the time. The 90% case is "build a fast, well-coded themed Shopify store". Boring, but it works.

If you're trying to decide, we're happy to score it with you. Drop us a line.