"Should we buy a ready system or build our own?" is the right question asked the wrong way. The real question is: where does your way of working differ from the standard one, and what is that difference worth? Here is the framework we walk clients through.

Start with the workflow inventory

List your core processes and mark each one: standard (everyone does it roughly this way, invoicing, basic bookkeeping), customised (we do it differently and that difference matters), or painful (we do it with spreadsheets, WhatsApp and heroics).

Buy software for the standard processes. Consider custom software where the customised and painful lists overlap: that intersection is where money leaks daily.

When buying wins

  • The process is genuinely standard and a mature product exists.
  • You need it working this month, not this quarter.
  • The vendor's update cycle benefits you (tax rules, compliance changes).
  • Your team would rather learn a proven tool than specify a new one.

When building wins

  • Your pricing, approvals or operations follow rules no package supports, and workarounds cost hours daily.
  • The process itself is your advantage, competitors cannot buy it because you would all look the same.
  • Data must flow between systems no vendor connects.
  • Licences scale worse than a one-time build: per-user fees across many staff and years quietly exceed the build cost.

The middle path most businesses miss

Between "buy the suite" and "build everything" sits the pragmatic option: a focused custom system that replaces the painful 20%, integrated with bought tools for the standard 80%. A shop keeps its accounting package but builds its branch-stock console. A school buys finance software but builds its gate attendance. This is usually the best value per shilling, and it is where a good software partner earns their keep by telling you what not to build.

Questions to ask any custom-software vendor

  1. What happens if we outgrow this, is the data exportable?
  2. Who owns the code and the database?
  3. What does year-three maintenance cost, and what does it include?
  4. Which parts will be boring, proven technology?
  5. Can we start with one workflow and expand?

If your answers point toward building, we would rather prove it with a scoped pilot than a grand project. Bring us the workflow inventory: we will tell you honestly which column each item belongs in, even when the answer is "don't build this".