Custom e-commerce and B2B, when Shopify is not enough.

When custom-built becomes necessary

Shopify covers most e-commerce projects, but some constraints go beyond it. A B2B distributor with 50,000 SKUs, customer-specific pricing, and CSV-imported orders from an ERP has needs that Shopify Plus B2B does not cover natively. A marketplace with third-party sellers, an order approval workflow, or a complex product configurator falls into the same category.

The practical rule: if you are working around the platform to make your sales model function, that is the signal to consider a custom build. Workarounds carry a hidden cost in maintenance, bugs, and technical debt that compounds over time.

The three custom architectures

Adobe Commerce (formerly Magento 2)

Adobe Commerce is the reference for retailers and brands with complex catalogs, sophisticated pricing rules, and multi-warehouse workflows. The platform offers near-total flexibility on the product data model, pricing rules, customer segments, and checkout customizations. The trade-off: the ecosystem demands significant maintenance effort. A mid-market Adobe Commerce project typically requires a team of 2 to 4 certified developers over 20 to 40 weeks.

Djust

Djust is a B2B e-commerce platform built in France, designed for distributors and manufacturers. It natively covers advanced B2B features: customer-specific catalogs, purchase rights management, order approval workflows, and deep ERP integrations. For a B2B pure-player project with specific business logic, Djust significantly reduces custom development compared to a Shopify Plus or Adobe Commerce implementation.

Headless architecture

Headless decouples the front end (React, Next.js) from the e-commerce back end (Shopify, Commercetools, BigCommerce). This choice makes sense when performance is a hard constraint (LCP scores under 1.5 s), when the design exceeds what a standard theme can deliver, or when the same content must feed multiple channels (web, mobile app, kiosks). Headless adds development and infrastructure complexity; it is not justified for a standard project.

B2B e-commerce: features that make the difference

67% of B2B buyers prefer to purchase without a sales rep in 2026 (Gartner). That figure gained six points in under a year. What those buyers expect is DTC-level fluidity combined with professional features.

• Company accounts with user and rights management (buyer, approver, view-only)
• Price lists by account or segment, applied automatically at login
• Deferred payment terms (net 30, net 60) integrated into the checkout
• Quick order by SKU or CSV file upload for reorders
• Self-serve portal: invoices, purchase orders, order history, returns without human contact
• Real-time ERP integration for stock levels, prices, and order statuses

Migration from an existing custom platform

Custom e-commerce projects often start from an existing situation: a Magento 1 instance at end of life, an in-house build that has become impossible to maintain, or a SaaS platform that no longer covers business needs. These migrations add a layer of complexity: auditing the existing code, mapping custom features, drafting the data migration plan, and building the SEO redirect plan.

The classic trap: underestimating the volume of business logic accumulated in the existing platform. A 'simple e-commerce site' can hide hundreds of pricing rules, approval workflows, and tacit integrations with internal tools. The upfront functional audit is not optional. It is what allows accurate quoting without surprises.

Headless SEO: key watch points

A poorly configured headless architecture can hurt SEO. Topics to address systematically: server-side rendering (SSR or SSG for catalog pages and product listings), dynamic XML sitemap management, hreflang for multilingual sites, and Core Web Vitals (headless helps LCP but can hurt CLS if image dimensions are not reserved).