Buy-versus-build is usually framed as a single, permanent choice. In practice a workflow is layered: some of it is genuinely commodity, some of it is standard behaviour you can configure, and a thin, valuable slice is specific to your business.
Deciding at the layer level is what stops both classic mistakes — buying a platform that cannot express how you work, and building software that competes with a twenty-dollar subscription.
Five questions that settle it
Run these against one workflow, not your whole business.
- Would a competitor doing this the same way as everyone else lose anything? If no, rent it.
- Does the workflow decide price, margin, risk or customer experience? If yes, ownership is on the table.
- How often does the process change? Frequent change favours software you control; stable process favours a licence.
- How much of the work today is people moving data between tools? That is integration, not a new platform.
- If your best operator left tomorrow, would the process survive? If not, the rules belong in software.
The three lanes
Rent means take the market standard and stop customising it: accounting, payroll, email, e-signature, helpdesk basics. Paying full price and using it plainly is the cheapest outcome available.
Extend means keep the platform of record and add the behaviour it will not express — configuration, integrations, a workflow layer over the top. This is the right answer far more often than vendors like to admit.
Own means build and hold the software behind a differentiating workflow, with the source, data and IP in your name. Narrow scope, high leverage.
The trap in the middle
The expensive failure mode is not building too much. It is bending a rented platform until it behaves like custom software — layers of workarounds, plugins and consultants, all of which you pay for and none of which you own.
The symptom is easy to spot: an annual bill that keeps rising while the team still keeps a spreadsheet on the side. At that point you are already paying custom-software prices without any of the ownership.
What 'buy' should always include
Before signing, ask three questions: can we get our data out in a usable form, on demand? What does this cost at three times our current volume? And what happens to our process if the roadmap changes direction?
If any answer is uncomfortable, you are not buying software — you are renting a dependency. That can still be the right call, as long as it is priced honestly.
Sequencing beats deciding
Almost every business we work with ends up in all three lanes at once, and the value comes from the order: rent the commodity immediately, extend what already works, then own the one workflow that decides your economics — after it has been measured, not before.