Shopify, Composable or Headless in 2026: Choosing Without the Religion

Platform debates in commerce are conducted with religious intensity — usually by people selling one of the religions. Having built on most of the options, here is the decision logic we actually use, agency incentives removed.

The 2026 landscape, briefly

Shopify won the mid-market so thoroughly that the honest question for most brands is not "which platform" but "why not Shopify". Checkout extensibility, native markets and B2B, and functions closed most of the historical gaps that used to force ejection. Meanwhile "headless" stopped being a badge of ambition and settled into being an architecture with specific costs and specific justifications. The MACH vendors consolidated; the agencies that sold headless-by-default are quieter now, mostly because their clients' retainers spoke.

Where Shopify-standard is simply correct

Catalogue under ~50k SKUs, one primary market or Shopify Markets-compatible expansion, conversion driven by merchandising and offer rather than exotic UX, team without dedicated engineers. Here, standard Shopify with a well-built theme ships in weeks, iterates without engineers, and its TCO embarrasses every alternative. The premium theme plus disciplined apps beats a custom front-end for these brands on conversion too — because speed of iteration compounds and bespoke React does not.

Where composable earns its keep

Real justifications, from projects that worked: multi-region catalogues with per-market pricing/tax/content complexity beyond Markets' comfort; content-led selling where the site is closer to a magazine with checkout (the feed and PIM layer doing heavy lifting); B2B price-list and approval-flow complexity; unification of commerce across apps, kiosks and web from one API; or an engineering org that genuinely maintains front-ends as products. The pattern: composable pays when the constraint is the platform, not when the constraint is traffic, offer or content — which is most brands.

The math agencies skip

A headless build is not a project; it is a standing engineering commitment. Realistic mid-market numbers: 3–6x the initial build cost of a themed approach, plus a permanent team (in-house or retained) for what a theme editor did for free — every banner, every landing page, every A/B test now passes through a deploy. Budget the second year, not the launch. If the second-year line makes the CFO flinch, the answer was Shopify-standard, and the flinch just saved you.

The hybrid that quietly wins

Most of our commerce builds land here: Shopify as the commerce engine (checkout, payments, orders — problems you never want to own), custom where differentiation lives — a Next.js storefront only if content velocity justifies it, custom apps for operations, and the integration layer treating Shopify as one node among marketplaces rather than the centre of the universe. Platform religion optimises for the vendor; architecture optimises for the P&L. Choose the boring thing everywhere except where the exciting thing pays rent.