Buy versus build is usually argued on licence cost against development cost. That comparison is the least useful one available, because it ignores the two factors that actually determine the outcome: whether the process differentiates you, and what it costs to leave.
Start with differentiation, not cost
Ask a direct question of any process: if a competitor ran this exactly as we do, would we lose anything? For payroll, expenses, email and accounting the answer is no. Buy them, take the vendor’s process, and spend no further thought on it.
For the process that constitutes your actual operating advantage — how you price, schedule, route, underwrite or match — the answer is different. Buying there means adopting the vendor’s view of how the work should be done, and that view is deliberately designed to fit everyone. Fitting everyone is precisely what you do not want in the process that differentiates you.
Count the cost that does not appear in the licence
Off-the-shelf rarely stays off-the-shelf. The figure worth comparing includes:
- Implementation and configuration, frequently a multiple of first-year licence cost
- Integration with every system it needs to exchange data with
- Customisation to close the gap between the vendor’s process and yours
- Per-seat cost escalation as you grow, and at each renewal
- The internal workarounds people build because the tool does not quite fit
- Migration cost when you eventually leave
That last item is the one most consistently omitted. Custom software has a known exit: you own the code and the data. Vendor software has an exit cost that is unknown at purchase and discovered at the worst moment.
The workaround tax
Watch what people do six months after a platform lands. If there are exported spreadsheets, a shared mailbox operating as a queue, or a step everyone knows to do outside the system, that is the fit gap made visible. It is a real recurring cost in hours, and it never appears on the business case that justified the purchase.
Buy the process you want to be ordinary. Build the process you want to be better at than everyone else.
The answer is usually both
The strongest architectures are neither purely bought nor purely built. They buy commodity capability — authentication, payments, messaging, accounting, CRM — and build the thin, high-value layer that connects them in a way specific to the business.
That layer is often far smaller than people expect. A focused platform that orchestrates bought components and encodes your particular operating model is frequently a matter of weeks, not years, and it is the part that competitors cannot buy a copy of.
A test worth applying
- Would we lose advantage if a competitor ran this identically? If no, buy.
- Does credible off-the-shelf software exist for it, without heavy customisation? If yes, buy.
- Will the vendor’s process force us to change how we work in ways that cost us? If yes, build.
- Can we state what leaving this vendor would cost in three years? If not, be cautious.
- Is the gap between what we need and what we can buy small enough to bridge with integration? If yes, buy and integrate.
Applied honestly, this usually produces a clearer answer than a spreadsheet comparing licence fees to day rates — and one that still holds two years later.
Written by the Voxalone team. If this raised a question about your own systems, get in touch — you will get a straight answer from an engineer, with no obligation attached.