Custom E-commerce Build vs Platform: When Each Makes Sense
Key takeaway
Use a hosted platform unless a specific requirement makes it impossible — complex B2B pricing, deep ERP integration, unusual fulfilment logic, or a catalogue model the platform can't express. Custom builds win on fit and lose on total cost of ownership, and most stores don't need the fit.
The honest default for online retail is a hosted platform. It handles payments, tax, fraud, PCI scope, and checkout conversion — all of which are expensive to build and unglamorous to maintain. The interesting question is what legitimately forces you off that default.
What a hosted platform gives you
- A checkout optimised by a vendor with vastly more conversion data than you have.
- PCI scope reduction, tax calculation, and fraud screening included.
- An app ecosystem that covers most operational needs without engineering.
- Patching, uptime, and scaling as somebody else's problem.
Reasons that genuinely justify a custom build
- B2B pricing complexity — customer-specific price lists, contract pricing, quantity breaks, approval workflows.
- Configurable or made-to-order products where the pricing depends on a calculation, not a variant list.
- Deep, real-time ERP or inventory integration where the platform's data model can't represent your reality.
- Fulfilment or subscription logic that the platform's order model fights at every step.
- Regulatory or data-residency requirements the platform can't satisfy.
The headless middle ground
Headless commerce keeps a platform for the commerce engine — catalogue, cart, checkout, payments — while you build a custom front end against its APIs. You get design and experience freedom without rebuilding the parts that carry the most risk. It costs more than a themed storefront and far less than a full custom build, and it's where most ambitious retailers land.
Compare total cost, not build cost
A platform's monthly fee and transaction percentage are visible and easy to resent. The custom build's equivalents — hosting, PCI compliance effort, security patching, payment integration maintenance, and the engineering time to keep checkout converting — are invisible and larger. Model three years, not three months, and include the cost of every feature the platform gives you for free.
A simple decision test
Write down the requirements the platform cannot meet. If the list is short, solve them with apps, an integration layer, or a headless front end. If the list includes your core commercial model, you have a real case for building — and the budget conversation should reflect a multi-year commitment, not a project.