Custom Shopify app development

When you do not need one

When an existing app does the job. The App Store is enormous and most requirements are not unusual. If a £30 a month app covers what you need, paying four figures to rebuild it is money set on fire and I will say so.

When the problem is configuration. A surprising number of build enquiries turn out to be a setting nobody found, or two apps fighting each other. That is an afternoon, not a project.

When it is theme work. Changing how something looks or behaves on the storefront is usually a theme change. It does not need an app and an app would make it harder to maintain.

When it is genuinely the right answer

Your process is the unusual part. Delivery windows, cut off times, production schedules, anything where the rented app has a settings page that almost fits. Almost fitting is the expensive case, because you end up running half the process in a spreadsheet beside the app you are paying for.

The percentage has become real money. Apps that take a cut of revenue are cheap at the start and one of your larger bills later, and the software does not improve as the number goes up.

You need two systems to agree. Most integration pain is one system not knowing what another has done. That is bespoke by definition.

What one developer means in practice

There is no team. Nobody is assigned to you after you sign, nothing is passed to someone junior, and nothing is outsourced. You talk to the person writing it, which is the only arrangement where the answer to a question arrives in minutes rather than after a round trip through an account manager.

The honest trade is capacity. One person cannot run four builds at once, so I take work in sequence and give you a date in writing. An agency can start tomorrow, and you will be one of nine projects a project manager is tracking.

How a build runs

Everything is built against a development store. Your live store is untouched until you have seen the thing working and said so. Where a build replaces something already running, the old app stays switched on until the new one reconciles against it, and if the numbers do not match we roll back and nothing has happened to your customers.

That is the part worth insisting on from anyone you hire, not just me. A migration that expects you to find the problems in production is not a migration, it is a gamble with your revenue.

What I have actually built

A subscription engine that replaced a rented subscription app for a UK meal box business, running on Shopify's native subscription contracts. It decides who gets charged twice a week and it has to be right again the following week. An inventory and fulfilment integration covering purchase orders through to dispatch. An AI support desk that reads inbound messages and either answers them from order data, drafts a reply, or escalates.

All three are still running and I still maintain them, which is a different claim from having shipped something once.