Package or custom: when is off-the-shelf simply better?

I build custom software. So I have an interest in you choosing custom. Keep that in mind for what follows — which is exactly why I start with when you should not.

Take a package when your process is not special. Accounting, invoicing, payroll, stock in an ordinary webshop: packages exist for those, tested by thousands of companies, kept in step with legislation, and cheaper than anything I could ever build. Having custom work made for that is paying for a problem that is already solved.

Take a package too when you do not yet know how you want to work. Custom software sets your way of working in code. If you are still figuring that out, you pay twice: once to build it and once to change it.

Custom starts to make sense the moment you begin working around the package. You recognise it by the spreadsheets. A package for the admin, and next to it a sheet where someone tracks what the package cannot handle. That sheet is proof your process does not fit the model.

The second signal is per-user pricing. Packages charge per employee per month. While there are five of you, that is fine. Grow to twenty and that invoice grows with you while the package does no more than before. At that point a one-off investment can work out cheaper than three years of subscription.

The third signal is your customer. Almost every package is built for your people, not for your customers. If you want a customer to configure their own price, follow their order or find their documents, the package usually stops. That is why most portals we build sit alongside rather than replace.

Which is also the most honest answer to "we already have software": replace nothing. Look at where people work around your package, and build only that piece. Smaller project, less risk, and you keep what already works.

Not sure? Simply write down which spreadsheets circulate in your company and what is in them. That list usually contains the answer already.