Commerce architecture
Shopify vs Custom Website: A 2026 Decision Framework
Shopify, a headless Shopify storefront and a separate custom commerce stack each solve a different problem. Compare them against what your business actually needs and how you intend to run it, not against invented revenue thresholds, app-spend figures or PageSpeed targets.
Reviewed August 8, 2026 against current Shopify and Google documentation.
Hassan Jamal·Mar 30, 2026·5 min read
First: which “custom website” do you mean?
The phrase covers two completely different projects, and most comparisons quietly pick one without telling you which.
- Self-hosted open-source, usually WooCommerce on WordPress. You rent hosting, install free software, and assemble the store from plugins. Cheap to start, and you own the updates, the security patching and the plugin conflicts. If this is what you meant, our WooCommerce comparison is the closer read, and the three-year cost model puts numbers on the maintenance side.
- A commerce experience built for you in code, either a headless storefront in front of Shopify's checkout or a separate custom stack. Higher to build, no plugin marketplace holding it together. This page is about that one.
The distinction matters because the two options fail differently. Self-hosted breaks through plugin conflicts and deferred updates. Custom code breaks through dependency drift and whoever built it becoming unavailable. Comparing either one to Shopify without naming which you mean produces an argument that cannot be checked.
The four real options
Stay on a Shopify theme when the theme and its supported extensions already deliver the customer experience you have agreed to, and the performance problems you can measure are fixable without changing the architecture. Move to a headless storefront when a differentiated experience is impractical inside those theme constraints and you have a team that can own a release process. Replace the commerce back end only when a documented business, regulatory or data requirement makes Shopify unworkable, because that is the largest of the four moves. Most stores land on the first option or on the fourth, which is optimizing what already exists.
- ✓Shopify theme: platform-native storefront, checkout and merchant workflow.
- ✓Headless Shopify: custom storefront with Shopify's commerce back end and checkout.
- ✓Separate custom commerce: custom storefront plus additional ownership of commerce services and operations.
- ✓Optimization or re-theme: improve the current system without changing its architecture.
Native
Fastest operational path when theme capabilities fit.
Headless
More experience control with Shopify still in the core.
Custom
Maximum control plus maximum lifecycle responsibility.
Evidence
Requirements, total cost, risk and measured outcomes.
1. Compare commerce requirements first
Write down what the store has to do before you compare any platform against any other. Six domains cover most of it: what you sell and how it is priced and taxed, how money is taken, how orders are fulfilled and returned, what the customer sees and searches, how the store is measured and fed to ad platforms, and what has to integrate with everything else. A platform comparison run without that inventory compares marketing pages instead of requirements.
- ✓What you sell, its variants, stock levels, the markets you sell into, prices, discounts and tax.
- ✓The cart, checkout, taking payment, fraud checks, customer accounts, subscriptions and selling to businesses.
- ✓Orders, fulfillment, returns, customer service and retail operations.
- ✓The content, search, recommendations, anything personalized, and whether people with disabilities can use it.
- ✓Analytics, the consent banner, ad platforms, product feeds, structured data, and protecting your search traffic.
- ✓Integrations, data ownership, security, recovery and support.
A platform that satisfies these requirements with supported native capability can be lower risk than custom code. A requirement that repeatedly fights the platform may justify a different frontend or stack. If the vocabulary here is unfamiliar, start with what headless commerce means in practice before scoring your own requirements.
Some requirements meet a platform boundary that configuration does not move. Check these early, because they change the shape of the answer rather than the size of the work:
- ✓Checkout is Shopify-controlled. What you can change depends on your plan and is done through the checkout editor and supported checkout extensions, not free template editing. Confirm what your plan allows before promising a bespoke checkout flow.
- ✓Storefront URLs keep Shopify's product and collection path structure, and a product that sits in several collections can be reachable at more than one address. Decide how canonicals and internal links handle that.
- ✓Variants and options per product are capped. Check the current limits against a configurator, made-to-order range or B2B price list before designing around them.
- ✓Taking payment through a gateway other than Shopify Payments adds a platform fee per transaction on top of the processor's own rate. The rate depends on plan and region, so read it from the plan you are on.
- ✓Each installed app can add its own scripts to the pages it touches, so the app list is a performance decision and a cost decision at the same time.
2. Use current commercial terms
Shopify prices, card rates, transaction fees and feature entitlements vary by plan, term and region. App charges can be recurring or usage-based and some apps bill outside Shopify. Use the merchant's contract, invoices and current official pages rather than a generic monthly total. For a method of auditing the app side of that bill line by line, see how to read your real monthly Shopify app bill.
This article gives you no app-spend, revenue or conversion figure that triggers a move off the platform. A threshold like that ignores what your apps actually do, what your team can operate and what the replacement costs to run once it is live. Commercial hosting is not free either, and payback follows your own scope, traffic and internal effort. Build the cost model below from your contract, your invoices and your own delivery plan, then compare it against the work the platform is doing for you today.
3. Build a comparable total-cost model
Five cost lines have to appear on both sides of the comparison for it to mean anything: what you pay providers, what implementation costs, what it costs to operate once it is live, what it costs to change or leave, and what you are carrying in risk. A comparison that prices the build on one side and the monthly plan on the other is not a comparison. Fill the table below from your own contract and invoices rather than from published list prices.
| Cost | Shopify theme | Headless or custom |
|---|---|---|
| Platform and providers | Plan, apps, payment and optional services | Commerce back end, hosting, CMS, search, email, monitoring and other services |
| Implementation | Theme, configuration, content and integrations | Discovery, product design, frontend, APIs, migration and acceptance |
| Operation | Merchandising, app administration and theme releases | Engineering, releases, security, monitoring and incident response |
| Change and exit | App, theme and platform transition limits | Code, data, provider, license and team dependencies |
| Risk | Platform and app constraints | Implementation, integration and maintenance complexity |
The right-hand column is where most estimates go wrong, because the build is only part of it. Our ecommerce development service page sets out what a custom storefront engagement actually contains, which is the figure that belongs in this table rather than a build price alone.
4. Measure performance instead of assigning platform scores
Shopify's Web Performance reports show real-user LCP, INP and CLS by URL, page type, device and time. Use them with repeated lab traces to identify the responsible work. A theme does not have one permanent score ceiling, and a custom storefront is not automatically fast. That applies at every plan level, including Plus, where the entitlements change but the storefront still runs the theme, the apps and the scripts you installed. Our diagnostic for a slow Shopify Plus store works through that case.
- ✓Test a real example of each page type: home, a collection, a product, search, the cart and a campaign page.
- ✓Note the device, the connection, where you tested from, whether the cache was warm, whether you were logged in, and what you had consented to.
- ✓Then trace where the time went: images and video, JavaScript, apps, analytics, data requests, and the theme or frontend code itself.
- ✓Preserve functional, accessibility, SEO, analytics and error guardrails.
Core Web Vitals explained covers the metric definitions and thresholds those reports use. For theme-level diagnosis, see diagnosing a slow Dawn theme and the wider Shopify store speed optimization guide. To connect those measurements to the funnel rather than to a score, see how to measure speed inside the conversion funnel and how to build a revenue-side evidence model.
5. Compare ownership precisely
“Own the website” is too vague for a contract. Identify control of domain, DNS, Shopify organization, payment accounts, repositories, hosting, CMS, analytics and vendor accounts. Separate client content and data, paid custom deliverables, PandaCodeGen's reusable or pre-existing tools and third-party components under their original licenses. We break the same question down further in what it means to own your website.
6. Know when Shopify remains the right choice
Stay on Shopify when the theme and its supported extensions already meet the customer experience you signed off, and when the merchant team would rather keep editing the store natively than operate a separate frontend. The performance test matters here: if the slow pages can be traced to images, apps and theme code, the fix is work inside the platform, not a replatform. Five conditions, all of which usually hold together:
- ✓The theme and supported extensions meet the accepted customer experience.
- ✓The merchant team values platform-native editing and operation.
- ✓Measured performance issues can be fixed without architectural change.
- ✓Required apps and workflows are well supported and economically acceptable.
- ✓The organization does not want to operate a separate frontend and release system.
7. Know when headless deserves evaluation
Headless earns evaluation when the experience you need cannot be built inside the theme constraints you have accepted, or when several content and commerce systems have to appear in one interface you control. The second test is organizational rather than technical: a headless storefront needs a team that can own a release process and a deployment target. If either half is missing, the extra integration work will not pay back.
- ✓A differentiated experience is impractical in the accepted theme constraints.
- ✓Multiple content, commerce or product systems must appear in one controlled interface.
- ✓The business needs a consistent experience across supported channels.
- ✓An established product and engineering operation needs frontend release independence.
- ✓The benefit you agreed on is worth the extra integration work and the maintenance that follows it.
Headless keeps Shopify as the system of record. Admin, product data, inventory, orders, customers and checkout stay where they are, so the storefront changes without a product or order-data migration and the merchant team keeps the interface it already knows. What moves to your side is the storefront code, its hosting and its release process, plus any customer-facing feature that a theme or an app used to provide. The storefront reads commerce data through Shopify's Storefront API, so the integration surface you take on is that API plus whatever else the page needs. Price that transfer before deciding, because it is the part that keeps costing after launch.
Our headless Shopify architecture and migration guide goes through the storefront paths, the integration surface and the operating responsibility each one transfers to your team.
8. Know when a separate custom commerce stack is justified
Replacing the commerce back end is a larger decision than replacing the storefront. Require a documented reason such as unsupported business logic, regulatory or data constraints, or materially different operating requirements. Include payments, tax, fraud, inventory, order management, customer service, security and recovery in scope. Our custom engineering service describes how work at that scope is specified and delivered.
9. Make migration SEO-safe
Search visibility survives a replatform when the URL inventory is captured before anything moves, every changed address is mapped to a permanent redirect, and the new storefront renders navigation and product information in a form a crawler can read. Product structured data and merchant feeds have to keep matching the offers a shopper sees. Everything else in the list below is verification.
- ✓Inventory current URLs, status codes, canonicals, metadata, content, internal links and structured data.
- ✓Preserve URLs where practical and map changed URLs to relevant permanent redirects.
- ✓Render crawlable navigation and product information with accurate index controls.
- ✓Keep product structured data and merchant feeds consistent with visible offers.
- ✓Validate robots, sitemaps and Search Console before and after release.
- ✓Expect possible fluctuations; do not promise ranking retention or a fixed recovery date.
What a migration does to search visibility covers the same controls from the search side, including what to monitor after release.
10. Use project-specific commercial terms
PandaCodeGen's public planning tiers are $1,500 Starter, $3,500 Growth and $5,000 to $10,000 Scale. Those are entry points for a defined scope, not quotes, and final scope may add or remove features. A common payment option is 30 percent at onboarding and 70 percent at the delivery milestone, and another written schedule may be agreed. Refund, ownership, performance, support and change-control rights follow the accepted contract, not a generic platform comparison. The pricing page lists what each tier covers.
Related reading
The same decision framework applied to other platforms: Webflow compared with a custom website, Wix compared with a custom website and Squarespace compared with a custom website.
Get your migration plan
We will compare the current Shopify stack, requirements, costs, dependencies and SEO controls before recommending optimization, headless Shopify or a separate custom build.
Primary sources
- Shopify pricing
- Shopify Plus pricing and features
- Shopify web performance reports
- Shopify app billing
- Shopify headless benefits and tradeoffs
- Shopify Storefront API
Frequently Asked Questions
When should I switch from Shopify to a custom website?
Do not switch from a headline threshold. Compare required experience, checkout, integrations, data, performance, current cost and operating capability. A theme optimization may solve implementation problems; headless can change the storefront while retaining Shopify commerce; fully custom transfers more responsibility.
How much does a custom ecommerce website cost compared to Shopify?
PandaCodeGen's planning tiers start at $1,500 Starter, $3,500 Growth and $5,000 to $10,000 Scale, but commerce scope can exceed those ranges. Compare current Shopify plans, apps, fees and internal work with migration, providers, maintenance, security, support and team time.
Will I lose sales during a Shopify migration?
A parallel build, tested data and integration flows, URL mapping, rollback and monitored cutover can reduce risk, but zero interruption or uninterrupted revenue cannot be guaranteed. Define acceptable interruption, rollback triggers and owners in the signed launch plan.
Can a custom website do everything Shopify does?
Only if each required catalog, checkout, payment, tax, fraud, inventory, order, account, discount, subscription, localization and support workflow is implemented or assigned to a provider. Custom code removes neither third-party dependencies nor ongoing operating work.
What is headless Shopify and is it worth it?
Headless Shopify keeps Shopify for commerce services while replacing the storefront. It can be worthwhile when measured storefront requirements justify the added search, analytics, consent, account, localization, integration, deployment and monitoring responsibility. There is no universal traffic or payback threshold.
Related Articles
Shopify Store Speed Optimization: What Actually Works (2026)
A field-first Shopify speed optimization guide covering Core Web Vitals, PageSpeed lab diagnostics, themes, apps, images, scripts, and the decision between theme work and headless commerce.
Shopify Stocky Sunset: August 31, 2026 Migration Guide
Shopify says Stocky cannot manage inventory after August 31, 2026. What to export, why suppliers require manual recreation, which APIs stop, and how to compare replacement workflows.
Shopify App Costs in 2026: Audit the Real Bill
Reconcile Shopify-billed charges, external subscriptions, usage, workflows, data and performance before you keep, replace or remove an app.