The choice is rarely between a package and a custom build. It is between three options, and the middle one is where most budgets die. This page sets out the four questions that settle it and what each route actually costs once you count the years after go-live.
By Tom Joseph · Last updated: 27 August 2026
Buy the parts of your business that look like everyone else's, and build only the part that is the reason customers choose you. Accounting, payroll and CRM are solved problems: a package will be less expensive and better than anything built for you. The exception is the process that is genuinely yours, the one no vendor models properly, where a package forces you to work in a way that costs you your advantage. Build that, buy the rest, and connect them.
Three routes, one choice
Package, configuration or custom: the least expensive route depends on your process, not your budget.

Package, configuration and custom. The trap sits between them: rebuilding a package until you inherit custom maintenance without the ownership.

How unusual is the process, who does the work in year two, what is the exit, and where does your edge sit.

Licences continue and custom needs maintenance. The comparison is only honest once year two counts.

Configuration is faster than custom and more flexible than a package, until the day you want to leave the platform.
Most comparisons offer two choices. There are three, and the middle one is the expensive surprise.
Run these before anyone quotes. Three "no" answers means buy a package and stop reading.
Package licensing looks low-cost in year one and compounds; custom looks expensive in year one and then mostly stops. The crossover depends on seat count, and with enough users a package can quietly cost more than the build it replaced.
The number nobody quotes is the cost of the workaround. A package that fits 85% leaves 15% living in spreadsheets, re-keying and a person who knows the trick. That cost is real, recurring, and invisible in the business case because it never appears on an invoice.
Our own prices are published, which makes at least one side of this comparison easy to run: audit €2,500, proof of concept €20,000, production from €50,000.
To start, yes, almost always. Over five years it depends on seat count and how much workaround the package leaves behind. Compare total cost including the manual work a package forces, not licence against build.
Often the best route. Buy the package, run the business, and let the friction show you which part is genuinely yours. Then build only that and connect it. It is also the least expensive way to be wrong.
Good for internal tools with a handful of users and simple rules. It becomes expensive at exactly the point custom would have been the answer: many users, real integrations, or logic that needs testing. Judge it as configuration, with the same upgrade and exit questions.
Own the source, own the data, and insist the stack is one you could hire for. If a build can only be maintained by the firm that wrote it, that is not a technical property of the software, it is a commercial choice someone made.
The €2,500 audit runs the four questions against your actual processes and gives you a written answer, including the case for buying a package and not hiring us.
Book a conversation