When an off-the-shelf platform beats a custom system

We build custom software. The text below is about when it is not worth ordering — including from us.

Not out of politeness. A project that should never have started ends the same way for both sides: somebody pays for a tool they do not use, and we have a deployment we would rather not show.

Four situations where the platform wins

The process is standard and nobody plans to change it. Selling a few dozen products, one delivery method, one payment gateway. A ready-made store does that from day one and costs a fraction of an implementation.

Nobody knows yet whether it will work. A new idea, a new market, no certainty about the model. A platform is then the cheapest way to test the hypothesis — and if the hypothesis fails, you cancel a subscription instead of writing off a build.

There is nobody to run it. A custom system needs somebody on your side who makes decisions and knows the process. Without that person the project stalls, however well it is written.

The scope changes every week. That is not a fault of the company, it is a stage. Until the process settles enough to be described, every week of a developer's work is writing something that will have to change shortly.

The moment it reverses

It does not arrive suddenly. It looks like this: first one plugin is added, then another. Then somebody starts re-typing orders into the warehouse by hand, because the integration "almost" works. Then the stock levels drift, and once a month they have to be levelled manually.

At that point you are already paying for a custom system — only in the form of somebody else's hours and somebody else's mistakes.

There is a signal that can be measured: count the hours somebody spends working around the platform. Not working in it — working around its limits. If that is a dozen-odd hours a month and growing, the bill starts to change. You can count it with the manual work calculator — every number is yours, no averages from our side.

What this text does not settle

We do not know what your platform costs together with commission on turnover, because we do not know your contract. Nor do we know how much of your process is genuinely non-standard — very often it turns out to be less than it seems at the start of a conversation.

Those are two things to count, not to settle with an article.

If it works out that the platform is enough, stay with it. If you are not sure, let us talk and we will go through your numbers. It may well end with us saying "you do not need this".