In short
- Custom is only the better choice when your business model no longer fits in a standard platform, not because it sounds fancier.
- The signals: subscriptions and contracts you imitate with plug-ins, integrations that run through a separate layer, and changes you have to make in several places.
- Custom takes time, a careful migration and a build partner for the long term. At Homie, the build took six months.
In this article (6)
Most businesses that ask me for a custom platform don’t need one. A well-set-up Shopify or WooCommerce shop does what they want, faster and with less hassle. But sometimes your business model simply no longer fits in a box. Then custom is not a luxury, but the only way to keep growing.
In this article I show you where that tipping point is. Using Homie, a client we moved from WordPress to a single platform of their own, together with our build partner Hoort.
Custom is not for everyone
Let me start with the honest part. A standard platform is an excellent choice for almost everyone who sells products. You get a proven checkout, an ecosystem of apps and themes, and updates that someone else builds for you.
Custom sounds tougher. That is exactly the problem. I see businesses that want a platform of their own because it feels like growing up, while their actual problem is a messy product page. Six months of development won’t fix that.
So the question is not: what is the best platform? The question is: does the way I sell still fit in this platform?
Three signals your platform no longer fits
You rarely notice it all at once. It creeps in, plug-in by plug-in.
1. You sell something the shop doesn’t know. An online shop is built for a simple rhythm: product in the basket, pay, ship, done. If you sell subscriptions, contracts, instalments or a separate route for business customers, you start imitating that. With a plug-in here and an extension there.
2. Your integrations run through a separate layer. Your ERP, your payment provider and your logistics have to talk to your shop. If they can’t do that directly, a middle layer appears to translate. It works, until something changes.
3. Every change has to be made in several places. This is the signal that hurts the most. Imagine you want to change the term of a subscription. That has to happen in the shop, in the middle layer and in the system behind it. Forget one place and a customer gets the wrong invoice. Your team becomes cautious. And a cautious team never builds anything new again.
Standard still fits
Fixed products, a regular checkout, a few integrations that work directly. You change something in one place and it is right everywhere.
Standard pinches
Subscriptions and contracts through stacked plug-ins, separate middleware to your systems, and fear of every update.
What happened at Homie
Homie rents out washing machines, tumble dryers and fridges on subscription, to consumers and to businesses. Sounds simple. But behind every subscription sits an order, a contract, recurring payments, a delivery, a collection date and the appliance itself.
The old platform was WordPress with WooCommerce. According to the case study by our build partner Hoort, a separate middleware layer was attached to it, which in turn talked to three external systems: an ERP, a payment provider and a logistics platform for route and warehouse management. That middleware had to be maintained separately, so changes often had to be made in several places. And WordPress with WooCommerce no longer carried a subscription service with logistics flows, instalments and business routes very well.
All three signals, then. At the same time.
We took care of the strategy, the design and the SEO safeguarding: what does the new platform need to do, for consumers and for the business channel, and how do we keep findability intact during the switch. In six months, Hoort built a new platform in Laravel and Vue. The logic of the middleware now lives in the platform itself, the three systems are connected again and the background tasks run through Laravel Horizon.
According to Hoort, customers barely noticed the switch. Same accounts, same contracts. Only now it runs on one platform, with its own CMS and in Dutch and English. You can read the whole switch in the case study of Homie’s custom platform.
What you take from this for your own business: the tipping point is rarely one big outage. It is the sum. Every new proposition that needs an extra plug-in. Every change that has to be made in three places. Every meeting that starts with “if we change this, what breaks?” If you hear that question more often than you’d like, that is your signal.
What you gain
Your business model becomes the foundation, not an add-on. At Homie, subscriptions, contracts, payments and appliances are fully fledged parts of the system. Not plug-ins stacked on top of a shop.
One place for everything. From order flow to blog, it runs on one codebase and one database. You make a change once.
Less that quietly goes wrong. If a background task fails, such as a payment or a status update from the warehouse, Laravel Horizon automatically retries it.
Speed. The Hoort case study reports a load time of 0.9 seconds for the new platform.
Room to grow. An extra language, a business channel, a new proposition: you build it in instead of around it.
Want to know what such a project involves? The page on custom ecommerce development explains how we approach it. And if you want to know why Laravel is often the foundation, read on at Laravel website development.
What it takes
Now the part that is missing from most sales pitches.
Time. At Homie, six months of building. A standard shop is up much faster.
A migration you don’t underestimate. Moving 40,000 records is not copy-and-paste work. Hoort had to build its own decryption layer for the encrypted data and first ran extensive test migrations on a staging environment, before the real switch came.
A build partner for the long term. With a standard platform, someone else builds the updates. With custom, that is your partner. So don’t choose on the first project alone, but on who will still be sitting next to you in three years.
Clear choices up front. Custom means anything is possible. That is also the danger. Without a sharp goal, you build a platform full of features nobody uses.
One lesson applies to every custom project: take your existing data and your old URLs along from the first sketch, not in the last month. How to do that is in ecommerce migration without losing SEO.
Goal first, platform second
We never start with the platform. Goal first: what does this site or shop need to deliver, and for whom? Then we bring the branding and the design into line. Only after that do we choose the platform that fits that goal. Sometimes that is custom. Often it isn’t.
The goal
What does the platform need to deliver, and how do you really sell today? Subscription, quote, individual products, business or consumer.
Design and branding
What it looks like and how it works for your customer, separate from the technology.
The platform that fits
Shopify, WooCommerce, headless or custom. The choice follows from the first two steps.
SEO safeguarding
At Homie, the category pages are still at the same addresses as in the old WooCommerce shop. That is how you take your findability with you.
Strategy, design and SEO safeguarding are part of every website and online shop we deliver. No extras. That is how we keep quality under control, whichever platform you land on. Still unsure whether you need custom at all? Start with the comparison on ecommerce website design.
Custom is not a status symbol. It is what you build when your business no longer fits in a box.
Rubin is the founder of Online Marketing Bakery in Roermond. You take the stage, we fill the room. And we make sure it buys.
[email protected] · More about the bakers
Further reading: Custom ecommerce development · Laravel website development · Ecommerce website design · Headless ecommerce development
How we work: first the offer, then the room. Want to get going with a brief like this? We do that in the Sold-Out Session: your offer and your event structure in one half-day.