Off-the-shelf or custom software?

The decision is not settled by which one is cheaper. The right question is this: does the company compete because its processes are different, or because it does the same work for less?

Accounting, payroll and statutory invoicing are defined by regulation and are the same for everyone. Building custom software there means solving a solved problem. Production planning, pricing logic, dealer relationships and order flow, by contrast, diverge from company to company — and that divergence is not accidental, it is accumulated.

A package asks you to fit that second group into its mould. The company adapts to the software. For some companies that is good news and their scattered practices come into order. For others it means giving up a difference that took years to earn.

In practice the pure choice is rare. A package is bought, then customised to fit the company. That is where the trouble starts.

Every customisation makes compatibility with the next release an open question. Over the years a company arrives here: it cannot upgrade because the customisations will break, and if it does not upgrade it falls outside support. That is customisation debt, and the interest compounds.

The question to ask at decision time is how many points of the package you will have to change. If the answer is “a few screens”, the package is right. If it is “order flow, pricing and production planning”, you are buying custom software after all — inside a package, and at a higher price.

Compared on licence and build cost alone, the package always looks cheaper. A multi-year total also contains:

• Annual per-user licensing and its yearly increase
• Customisation and consulting days
• The cost of version upgrades and the downtime they carry
• Per-integration module fees
• The cost of extracting your data and changing supplier

With custom software the build cost sits at the front, followed by maintenance and development, and there is no licence tied to headcount. Over five years those two curves cross for most mid-sized and large companies, and the crossing point arrives earlier as user numbers rise.

A software decision is also a decision about what it will cost to reverse it. Whether the data is yours and portable, who owns the source code, and whether the system runs on your own servers matter more than price across a ten-year relationship.

That is why the systems we build belong to the client in source and run on the client’s own servers. We treat that as a contract term protecting the buyer’s negotiating position rather than as a marketing promise.

Buy the package if all three are true: your processes are close to how the industry generally works, you can use the package largely unmodified, and your user count keeps licensing reasonable.

Take custom software seriously if even one of these is true: your advantage comes from how your processes differ, running the package would require breaking how you work, or your data and continuity would depend on a supplier’s commercial decisions.

If you are unsure, measure before deciding. Once a process map, the state of the current system and a multi-year cost comparison exist, the answer usually becomes visible on its own.