← SORT

Is custom software worth building?

Usually not — and the useful question is which cases are the exception. Here is how I decide, including when I’ve told a client not to build.

The short answer is that it usually isn’t, and the useful part is knowing which cases are the exception. Below is how I decide, including the cases where I’ve told a client not to build.

Is custom software worth building?

Often, no. Custom software is worth building when the work it replaces is repetitive, measurable and already costing somebody real hours — and when no existing tool covers it without being bent out of shape. If an off-the-shelf product does 80% of the job, the honest answer is usually to buy it and live with the other 20%.

When is off-the-shelf the better buy?

When the problem is common. Accounting, payroll, CRM and inventory are solved problems, and a product used by thousands of businesses will be more reliable and cheaper than anything built once for you. The right move is usually to configure one of those properly rather than to build a worse version of it.

When is custom actually worth it?

When the work is specific to how your business runs, and that specificity is the point rather than an accident. A scoring rule only your sales desk uses. A document only your practice produces. A question that spans two systems no product was designed to join. These are the cases where bending a product costs more than building the missing piece.

What does custom software cost to run?

More than people expect, and the running cost matters more than the build. Hosting, backups, the services it calls, and somebody to change it when the business changes. A tool that costs a little to build and a lot to keep alive is a worse deal than the reverse. Ask for the running cost before you agree to the build.

How do you decide?

Start with one task rather than a system. Work out where the information actually lives, what a correct result looks like, and who owns the decision. By the time those three are written down, it is usually obvious whether anything needs building — and sometimes the answer is that it doesn’t.