The build-or-buy question is usually decided on price. That is the least reliable yardstick, because the relevant costs only become visible later.
How distinctive is the process really?
If a workflow is essentially what thousands of other companies also do, a standard product is almost always the better choice. Building your own pays off where the process is a differentiator and changes regularly.
Who carries operations?
Software is not finished when it ships. Updates, dependencies, security advisories and changing interfaces keep coming. That question should be answered before the first line of code.
What about getting out?
A frequently overlooked point: how do the data get out again? A standard product with a documented export is worth more than a custom build whose data model only one person understands – and vice versa.
A pragmatic middle ground
Often the best answer is not either-or. A standard product for the core, extended by a thin custom layer for the part that is genuinely specific, combines reliability with a precise fit.
Takeaway
Do not decide on licence cost, decide on distinctiveness, operational ownership and exit capability. Those three questions say more about total cost than any quote.
