Shopify vs Custom E-commerce: When to Pick Which

Almost every founder building an online store hits the same fork in the road within their first month. Do you launch on Shopify and accept the guardrails, or do you build something custom and own every pixel of the funnel? The honest answer is that most teams pick wrong in both directions. They either over-engineer a bespoke stack for a store doing three orders a day, or they cram a complex B2B catalog with tiered pricing and negotiated freight into Shopify and then quietly bleed conversion for two years. The right choice is not about taste, it is about matching your business model, order volume, and margin structure to the actual constraints of each platform. This post walks through the decision the way I do it with founders who have real revenue on the line: no platform loyalty, no hype, just the specific signals that tell you which side of the line you are on and what the switching cost looks like if you get it wrong the first time.

The real question is not tech, it is control over the checkout

Shopify is not a website builder, it is a checkout with a store attached. That is its superpower and its cage. You get PCI compliance, a fast global CDN, fraud tools, and a checkout that has been A/B tested against billions of sessions. You do not have to think about card vaulting, 3DS flows, or refund state machines. In return, you agree to work within Liquid, the App Store, and Shopify's data model. For 80 percent of consumer brands doing under 20 million in annual revenue, that trade is a bargain. You are not going to out-engineer Shopify's checkout with a two-person team, and you should not try.

Custom becomes interesting the moment your business does something the checkout cannot express cleanly. Subscription boxes with variable cadence, marketplace splits across multiple sellers, quote-based B2B pricing, complex bundles that need to be atomic in inventory, or region-specific tax logic that Shopify Tax cannot handle. If any of those describe your model, you will fight Shopify every quarter. Custom is not automatically better here, but it stops being a luxury and starts being a requirement. The signal to watch is how many apps you need to duct-tape together to make a single order flow work. Once you cross five apps in the checkout path, you are already paying the custom tax without getting any of the benefits, because every one of those apps is a webhook that can fail, a data model you do not control, and a monthly bill that scales with GMV.

There is a second signal that matters as much as the first: how much of your revenue depends on merchandising nuance. If your product pages are the product, if a bundle configurator or a personalized recommendation engine is the reason people buy, Shopify themes can feel like painting a mural through a mail slot. You can hire agencies to twist Liquid into shape, but you will always be one platform update away from a broken build.

Total cost of ownership over 24 months

Founders love to compare the Shopify monthly fee to a developer estimate and declare custom cheaper. It is not. A realistic Shopify Plus store with three or four premium apps, a custom theme, and a mid-tier agency on retainer runs between 4,000 and 12,000 dollars a month all-in. A custom Next.js storefront with a headless commerce engine, a payments integration, an admin panel, and observability starts at 80,000 to 150,000 dollars to build and then costs 3,000 to 8,000 a month to keep running once you add hosting, error tracking, and at least one part-time engineer.

The break-even is usually somewhere between month 14 and month 20, and only if you actually need the custom capabilities. If you are building custom to save money, you will lose. If you are building custom because a specific feature is worth ten percent of revenue and Shopify cannot deliver it, the math flips fast. Run the numbers with real conversion assumptions before you decide, and be honest about the engineering headcount you will need to keep the lights on. A store that generates 5 million in annual revenue but requires a two-engineer team just for maintenance is often less profitable than a Shopify Plus equivalent generating the same revenue with a single freelance developer on retainer.

Hidden costs founders forget

• Fraud tooling: Signifyd, Kount, or Riskified add 0.5 to 1.5 percent of GMV, and you need it the moment chargebacks start.
• Tax calculation: Avalara or TaxJar run 2,400 to 12,000 dollars a year for custom stacks that Shopify Tax handles inline.
• PCI audits: Level 1 merchants need annual audits, budget 15,000 to 40,000 dollars.
• Search: Algolia or Typesense hosting, 100 to 2,000 dollars a month depending on catalog size.
• CMS: Sanity, Contentful, or Storyblok, another 100 to 1,500 dollars a month.
• Email and SMS: Klaviyo integrations that are free on Shopify often require custom event pipelines on a bespoke stack.
• Returns portal: Loop, Aftership, and Narvar are Shopify-native and cost 3x to replicate cleanly on custom.

When Shopify is the obvious answer

If you are a DTC brand under 10 million in revenue, selling physical goods, shipping from one or two warehouses, and your checkout is a straightforward cart-to-payment flow, launch on Shopify and do not look back. You will get to revenue faster, you will not need a full-time engineer, and you can spend your energy on product, brand, and customer acquisition, which are the three things that actually determine whether your store survives its second year. The theme ecosystem, the app marketplace, and Shopify Payments will cover 90 percent of what you need out of the box.

Shopify also wins when you need to launch quickly and validate a category. Our own startup and MVP work often starts on Shopify precisely because we can get a founder from idea to first sale in under three weeks. You can always re-platform later if the business justifies it, but you cannot get back the six months you spend building a custom checkout that nobody uses because you never got the product-market fit right. The dirty secret of e-commerce is that most stores fail because of demand, not technology, and no amount of custom code fixes a product people do not want.

Two more scenarios where Shopify is the obvious call: you are running a physical retail location and need POS parity, or you are selling internationally and want Markets, geolocation, and multi-currency without building your own currency conversion pipeline. Both are hard problems that Shopify has already solved for you.

When custom is worth the pain

Custom e-commerce earns its place when your business model breaks Shopify's assumptions. A few concrete examples: a rental marketplace where inventory is time-based, not quantity-based. A B2B distributor with 40,000 SKUs, customer-specific pricing, and net-30 terms. A subscription platform where users can swap items mid-cycle and pause without canceling. A regulated category like alcohol, CBD, or firearms where you need age verification, shipping restrictions, and payment routing that changes by state. In each of those cases, the workarounds on Shopify add up to more complexity than a purpose-built system.

The other trigger is checkout conversion. If you have real data showing that Shopify's checkout is costing you two or more points of conversion because you cannot restructure the flow, upsells, or address collection the way your customer research demands, custom starts to pay for itself quickly. On a 20 million dollar business, two points of conversion is 400,000 dollars a year, which more than funds a small engineering team. Just be sure you have the data before you spend the money, because most conversion problems are actually product page problems, pricing problems, or trust problems, none of which get fixed by rewriting the checkout.

A third trigger is data ownership. If you are running a media-plus-commerce play, or you want to fuse first-party analytics with editorial content in a way that Shopify's event model does not support, custom lets you own the data pipeline end to end. That matters more every year as third-party tracking dies and first-party data becomes the currency of paid acquisition.

The middle path: headless Shopify

There is a third option that founders often miss. You can keep Shopify as the checkout and commerce engine, and build a custom storefront on top using the Storefront API or Hydrogen. This gives you full control over the product pages, cart UX, and content, while Shopify still handles payments, inventory, and fulfillment. It is the sweet spot for brands that need marketing flexibility and performance but do not want to rebuild the boring parts of commerce.

Headless Shopify is not free either. You are now maintaining a Next.js app, a CMS, a search layer, and the integration glue between them. Expect an extra 2,000 to 5,000 dollars a month in engineering and hosting compared to a themed Shopify store. But if your brand lives or dies on content, editorial pages, or a specific product page experience, it is often the right compromise. Our web development team ships headless Shopify projects in six to eight weeks when the scope is clean.

The right test for headless: would you happily rebuild the storefront every 18 months to keep pace with design trends and Core Web Vitals? If yes, headless makes sense. If you would rather set the store once and iterate on merchandising for years, stay themed. Headless demands ongoing engineering attention that a themed store simply does not.

The QwiklyLaunch 45-day approach to picking

When a founder comes to us undecided, we spend the first week not writing code. We map the order flow end to end, list every checkout requirement, model unit economics at three revenue levels, and stress-test each option against the roadmap for the next 18 months. Only then do we commit to a stack. That week saves months of rework later.

Inside our 45-day launch program, we default to Shopify or headless Shopify for consumer stores because it is the fastest path to a real transacting business. For B2B, marketplaces, or regulated categories, we default to custom on Next.js with Medusa, Saleor, or a bespoke commerce core. Either way, the goal is the same: a store live and taking real orders in six weeks, not six months. If you want us to run that discovery week for your project, get in touch and we will walk you through it.

The 45-day scope forces a discipline that founders otherwise skip. You cannot build every feature, so you must rank them. You cannot try three payment providers, so you must pick one. That constraint is the reason our launches ship on time, and it is also the reason a Shopify default wins most of the time. Complexity is the enemy of a first launch.

Team, workflow, and who owns what

Platform choice determines the team you need to run the business, and founders forget this until they are three months in and drowning in tickets. A themed Shopify store can be operated by a merchandiser, a customer service rep, and a part-time freelance developer. That is it. Product uploads, promotions, theme tweaks, and app configuration all happen inside the admin, and non-technical people can move fast without an engineering bottleneck.

A headless or custom stack changes the org chart. Now you need at least one full-time engineer or a retained agency because every content change, every promotion rule, every payment method addition potentially touches code. That is fine if you have the revenue to fund it, but it turns a lean commerce team into a small software company. Budget accordingly. The founders who succeed with custom stacks treat their storefront the way a SaaS company treats its product: continuous deploys, on-call rotation, and a real engineering backlog. The founders who fail try to run custom the way they ran their Shopify store, and everything grinds to a halt within a quarter.

One workflow trap worth naming: on Shopify, marketing owns the store. On custom, engineering owns the store, and marketing files tickets. That single shift is often the biggest culture change in a re-platform, and it kills more projects than any technical issue. Plan for it in advance by giving marketing a real CMS with preview, scheduling, and rollback, or you will end up with a beautiful stack that nobody wants to use.

How to switch platforms without losing revenue

If you picked wrong the first time, re-platforming is survivable but not cheap. Plan for 90 to 180 days depending on catalog size, historical order data, and integrations. Migrate customers and order history first, then products, then content, then checkout. Run both stores in parallel for at least two weeks with a canary of traffic on the new stack. Redirect every legacy URL, and monitor organic search rankings daily for the first month. We have seen brands lose 30 percent of organic traffic overnight from a botched migration, and it takes six months to recover. Look at existing case studies on our projects page for examples of migrations we have run without a traffic dip.

Two rules that save you from disaster during a migration. First, do not change your URL structure and your platform in the same release, ever. Change one, stabilize, then change the other. Second, keep the old store online in read-only mode for 60 days after cutover so support agents can pull historical orders while the new system's order search is still shaking out bugs.

The right platform choice is the one that matches your business today and does not paint you into a corner 18 months out. If you want a second opinion before you commit, talk to us and we will give you an honest read in a 30-minute call.