E-commerce tools

The shop is rarely the problem — the work around it is

Most companies do not need a new shop. They need less manual work around the existing one: orders re-keyed between systems, labels printed one at a time, exports to Excel and production data someone assembles by hand in a graphics program.

That is exactly where I start. And if it turns out that the platform itself is the limit, I build the shop too — headless on Medusa.js.

Tools facing inwards

• Warehouse and shipping — barcode capture by smartphone or handheld terminal, batch printing of labels and delivery notes, goods receipt, stocktaking.
• Order processing — batch handling, status, notifications, returns. Without re-keying between systems.
• Integration with ERP and accounting — through an API, not through overnight CSV exports.
• Production — straight from the order to the print or production file in the right format, bleed and colour space.

Tools for customers on the website

• Product configurators — customers assemble a variant and see the price and the result immediately.
• Print designers and personalisation — upload a motif, position it, view it on the product. Production receives a finished print file.
• Calculators and enquiry forms — for products that cannot be sold with a single click.

Headless commerce on Medusa.js

Medusa.js is an open-source commerce backend written in TypeScript. It turns the shop into a service you operate yourself: your own data models, your own pricing and discount logic, B2B roles and approvals, several sales channels on one inventory — and a frontend that is not tied to a theme.

I work in Medusa at source-code level, not in a configuration screen. For Medusa v2 I wrote a payment provider and published it under the MIT licence: medusa-plugin-comgate — authorisation and capture, verification of incoming webhooks, refunds and the reconciliation between order and payment. Comgate is a Czech payment service provider, but the provider interface behind it is the same one that Stripe, Adyen or Mollie implement. The work transfers: whoever has built that layer cleanly once will integrate your provider too — including purchase on invoice, direct debit or instalments.

The honest decision rule

Building your own is more expensive to set up and to maintain. It only pays off once the standard platform costs money instead of saving it. That line is usually easy to recognise:

• Stay with Shopify or Shopware as long as your processes can be represented with apps and standard functions and the effort lies mainly in the warehouse, in production or in the connection to the ERP. Then I build the tools, not the shop.
• Headless becomes sensible when product logic, pricing, B2B roles or production data only work through workarounds, when app fees and revenue shares grow to an order of magnitude at which they would finance the development, or when the shop is only one part of a larger application.

Which of the two cases applies I will tell you after the initial call — including when the answer is that you should stay with your platform.