E-commerce website: off-the-shelf platform or custom build?
25 juillet 2026
The badly framed question
"Should we go with Shopify or build custom?" isn't the right starting question. The right one is: what, in your business, doesn't fit a standard store?
If the answer is "nothing", the off-the-shelf platform wins and the subject is closed. Custom development is only justified by a real gap between your operation and the standard model.
What fits the standard
A product catalogue, variants, a cart, card payment, shipping rates by zone, a promo code, order tracking, an invoice. Every platform handles all of that properly, with no development.
If your store fits that list, a custom site will cost you more for a result that's equivalent at best.
What doesn't fit
The cases where custom becomes relevant are fairly identifiable:
A calculated price. Priced by linear metre, by actual weight, by combined options, or negotiated per client. Platforms handle variants, not formulas.
Stock shared with a physical shop. If an item sold in store has to disappear from the site within the minute, syncing with the till becomes the heart of the project, not a plugin.
Specific local logistics. In Reunion Island in particular, carriers, lead times and import constraints don't match the default modules.
A non-standard funnel. Quote before order, business account approval, recurring orders, minimum order per category.
The cost of extensions. Past a certain number of paid modules and workarounds, the platform costs more than a targeted build, and gets more fragile with every update.
The middle option people forget
Between the two sits an option that's often the right one: keep a platform for the catalogue, cart and payment, and build custom only the part that falls outside the standard, connected by API.
You keep payment compliance, security updates and the ecosystem, while handling the business specificity properly. It's almost always cheaper than rebuilding everything.
The cost items discovered too late
A realistic e-commerce budget includes what isn't the site:
- Product records. Photos, descriptions, attributes. For 500 SKUs that's a project in its own right.
- Transaction fees. A few percent per order, which weigh heavily at volume.
- Logistics. Picking, packing, returns, disputes. The site only triggers the work.
- Maintenance. Security updates, module compatibility, certificate renewals.
- Acquisition. A site with no traffic sells nothing, whatever its technical quality.
On a first project, development rarely accounts for more than half the real first-year cost.
The technical points that decide revenue
Mobile speed. Most traffic arrives on mobile, often on an average network. Every second of wait on the product page is paid for.
The checkout funnel. Guest payment, a minimum of fields, shipping costs shown before the final step. Most abandonment happens right there.
Internal search. Past a hundred SKUs, a search that doesn't tolerate typos loses sales silently.
Product structured data. Price, availability and reviews in Product markup determine rich display in search results.
How to decide in a week
List the ten operations you perform most often in your business. For each, check whether the platform under consideration covers it natively, via an extension, or not at all.
If three or more lines fall into "not at all", a custom build, at least partial, is worth costing. Below that, take the platform and put the difference into photography and acquisition.
Want an outside view on that choice? Write to us.