"We'll just build it ourselves" is the most expensive sentence in operations. It is sometimes also correct. The difference is whether the workflow you are automating is genuinely distinctive, or merely under-configured in the tool you already pay for.
Buy when the process is standard
Accounting, payroll, email, helpdesk, HR records. These are solved problems with regulatory depth you do not want to reimplement. Buy, configure properly, and spend the saved budget on integration.
Build when the process is your advantage
If the way you route deliveries, schedule technicians or grade produce is why customers choose you, a generic tool will sand that advantage down to the vendor's median customer. That is where custom software earns its cost.
Three questions before you commit
- Can we describe the rule in one paragraph? If not, discovery is the project — not development.
- Who owns the process day to day? A build without an internal owner becomes orphaned software.
- What is the cost of the current workaround? Hours multiplied by rate, plus error cost. If the number is small, buy a better spreadsheet and move on.
Deliver in slices
Do not fund an eighteen-month monolith. Fund a six-week first release that replaces one painful step, measure adoption, then decide. Teams that ship monthly finish; teams that ship once usually ship late.
Own the exit
Insist on source ownership, documented APIs, standard frameworks and a deployment you can run yourself. The mark of a healthy build is that you could leave your supplier without shutting down the business.
Have a question about this?
Our engineers answer directly — no gatekeeping.
