The question shows up in nearly every e-commerce engagement we take. Sometimes it’s posed directly: should we use Shopify, or should we build something custom? More often it’s posed obliquely — “we’re hitting the limits of our current platform, what should we move to?” — but the underlying question is the same. And the answer most agencies give is reflexive rather than considered: Shopify shops will tell you Shopify; custom-build shops will tell you custom; the platform-agnostic shops give a noncommittal “it depends” without explaining what it depends on.
After thirteen years of these conversations, we’ve developed a framework that produces consistent answers — and more importantly, defensible ones. The framework isn’t sophisticated. It’s just the discipline of asking the right questions in the right order, instead of starting with the answer and reverse-engineering the justification.
The wrong way to make this decision
Most “Shopify versus custom” debates start with the wrong question, which is “what can each platform do?” Both can do almost anything, given enough time and money. Shopify with the right combination of apps, themes, and custom code can replicate 95% of what a custom build offers. A custom build, given enough engineering hours, can replicate 100% of what Shopify offers. So feature comparison alone is rarely decisive.
The right starting question is different: what does the business actually need to do, and what’s the cost of getting that capability versus the value it produces? Once you frame it as a build-versus-buy problem rather than a platform-versus-platform debate, the decision becomes considerably clearer.
The questions that actually matter
In order, these are what we work through with every e-commerce client deciding between Shopify and custom infrastructure.
1. What’s your gross merchandise volume, and what’s your trajectory?
If you’re under $1M GMV annually, the answer is almost always Shopify. The math doesn’t work for custom infrastructure at that scale: the cost of building, hosting, and maintaining a custom platform exceeds the cost of Shopify Plus until you’re well past $5M GMV. Below the threshold, you’re paying for engineering complexity that doesn’t earn its keep.
Above $20M, the calculus genuinely shifts. Shopify Plus pricing is volume-based, and at the high end you start paying enough that custom infrastructure becomes economically viable. But “viable” doesn’t mean “preferable” — most $20M brands stay on Shopify because the operational cost of running a custom stack exceeds the platform fees they’d save.
Between $1M and $20M is the genuinely contested zone. The right answer here depends on the rest of the questions below.
2. What does your inventory model look like?
Shopify is brilliant at standard SKU-and-variant inventory. T-shirts in three colours and four sizes? Handled cleanly. Subscription boxes with monthly variants? Doable with apps. But certain inventory models break Shopify’s mental model in ways that fight the platform constantly: complex product configurators with dozens of dependent options, made-to-order goods with variable lead times, bundled products where pricing depends on other inventory, B2B catalogs with customer-specific pricing tiers and minimum order quantities.
If your inventory model fits Shopify’s assumptions, the platform is enormously productive. If it fights those assumptions, you’ll spend more time and money working around Shopify’s constraints than you would building custom infrastructure that matches your actual operations.
3. How many integrations do you need, and how exotic are they?
Shopify has the most-developed e-commerce integration ecosystem in existence. Your accounting software, ERP, warehouse management system, marketing automation, customer service platform — almost all have first-class Shopify connectors that work out of the box. If your stack is mainstream, this is a massive advantage.
But “first-class” gets thinner the further you go from the mainstream. Custom enterprise systems, niche international payment processors, specialised B2B portals, regional logistics providers in non-North American markets — these often require custom integration work whether you’re on Shopify or custom infrastructure. The advantage of Shopify diminishes when half your integration list is bespoke anyway.
4. What’s your team’s actual capability?
This is the question agencies rarely ask, and it’s the most important. A custom-built e-commerce platform is, in most respects, a piece of software that needs ongoing maintenance, security patching, performance tuning, and feature development. If you don’t have a permanent engineering team — or aren’t willing to retain an agency to function as one — you should not be on a custom platform regardless of what your other answers look like.
We’ve seen this fail repeatedly. A founder builds custom infrastructure during a high-growth period, the agency hands off, and within eighteen months the codebase has rotted, security patches have been skipped, the original engineers are gone, and the business is paying more to maintain dysfunctional custom infrastructure than they would to migrate back to Shopify. The platform decision must match the operational reality of who will run it.
5. What’s your competitive edge in your category?
This is the question that surfaces whether custom is genuinely worth it, when it might be. The answer is rarely “the technology” — most successful e-commerce brands win on product, brand, customer service, or operational excellence, not on storefront technology. If your competitive moat is a unique product, a great brand, or operational efficiency, your storefront should optimise for reliability and cost-efficiency, which means Shopify.
But occasionally — and we mean occasionally — the storefront experience itself is the moat. Configurator-driven products where the configuration is the brand experience. Subscription services where the unboxing experience is core to retention. Marketplace-style platforms where matching buyers and sellers is the primary product. In these cases, custom infrastructure that lets you build exactly the experience you want, without fighting platform constraints, can be genuinely worth the cost.
What good Shopify looks like
There’s a quiet phenomenon worth naming: most “we need to leave Shopify” conversations are actually “we have bad Shopify, and we’re blaming the platform.” Shopify done well is a remarkably flexible platform. Shopify done badly — too many apps, performance-killing themes, customisations bolted on rather than built in — is exactly as constrained as the migrations consultants will tell you.
The best Shopify storefronts we’ve seen share certain qualities. They use a small number of well-chosen apps rather than dozens of marginal ones. Their themes are custom-built or heavily modified, not stock themes with surface customisation. They use Shopify’s APIs strategically to build the bespoke experiences that matter, while leaning on the platform’s defaults for everything that doesn’t. They treat performance as a non-negotiable, hitting Lighthouse scores in the 90s by being disciplined about what gets loaded when.
If your Shopify storefront is failing you, the question to ask first is whether you’re failing Shopify. Often the answer is yes, and the fix is not migration; it’s better Shopify.
What good custom looks like
Custom e-commerce done well — the rare cases where it genuinely pays off — has a specific shape too. It’s built on a modern framework with strong separation of concerns: storefront layer, commerce layer, integration layer, all decoupled. It uses headless architecture so the storefront can evolve independently of the back-end. It has comprehensive automated testing because it has to; nobody else is going to catch the regressions. It’s hosted on infrastructure (Vercel, AWS, Cloudflare) that handles the operational toil that Shopify would otherwise handle for free.
And critically: it has a permanent engineering presence behind it. Either an in-house team or a long-term agency relationship that functions like one. The cases where custom e-commerce works are cases where the business has accepted that running a piece of bespoke software is now part of their operating model, and they’re staffed and budgeted accordingly.
The decision in practice
When we work through this framework with clients, the answer is usually clearer than they expected — and surprisingly often, it’s “stay on Shopify but rebuild it properly.” The cases where custom is genuinely the right call are the minority, and they’re the ones with very specific business shapes: complex configuration as a brand experience, marketplace-style platforms, or scale that justifies the engineering cost.
For everyone else, the question isn’t really “Shopify or custom?” It’s “are we running Shopify well, or are we paying for the platform without using it?” The answer to that question is the answer to the rebuild conversation.
Most agencies pitching you a custom e-commerce build are not solving for your business. They’re solving for their own preference for greenfield engineering work, which is more interesting than fixing existing Shopify implementations. That’s a real bias, and it costs clients real money. Asking the questions above, in the order above, surfaces it before the proposal gets signed.