Outgrowing Shopify: When a Custom Commerce Platform Is Genuinely the Right Call

We turn down more custom commerce projects than we take. Most brands that ask for one do not need one — they need a better-built store on the platform they already have. But a minority genuinely are constrained by the platform, and for them staying put costs more every quarter.

When the Platform Is Genuinely the Constraint

Multi-vendor marketplaces are the clearest case. If independent sellers need their own dashboards, their own inventory, their own payouts and their own commission structures, you are building a marketplace, and marketplace logic fights every single-merchant platform. Complex B2B is the second: customer-specific catalogues, negotiated pricing, credit terms, approval workflows and purchase orders. Some of this can be forced into Shopify B2B, but past a certain complexity the workarounds cost more than a purpose-built system.

The third is deep operational integration — where the storefront is one surface on a larger system that also runs manufacturing, field service, or a physical logistics network, and the commerce data must live in the same model as everything else.

When It Is Not the Platform

If the complaint is "our store is slow", that is almost always the theme and the app stack, not the platform. If it is "we cannot get the design we want", that is a theme development problem. If it is "the apps cost too much", a custom build will cost more than several years of app subscriptions. And if it is "we want to own our technology", be honest about whether that means owning it or maintaining it — those are different commitments.

What Custom Actually Buys You

Control over the data model, control over the checkout, no per-transaction platform fee, and no dependency on a third party deprecating a feature you rely on. That last one matters more than people expect — every hosted platform eventually changes something that breaks a workflow you built a business process around.

What It Costs You

Everything the platform did for free: PCI compliance, fraud tooling, uptime, payment gateway maintenance, security patching, and the long tail of ecommerce features you did not realise you were using until they were gone — abandoned cart flows, tax tables, address validation, refund handling. A custom platform is not a project with an end date. It is an ongoing engineering commitment, and it needs a budget line every year, not just in year one.

The Honest Test

Write down the three things your platform prevents you from doing. If all three are business-model constraints — not design, not speed, not cost — build custom. If any of them is really a build-quality problem, fix the build first. It is dramatically cheaper, and you will still have the option to go custom next year with far better information about what you actually need.