“Should we just buy something, or have this built?” We hear the question weekly, and unlike most people you could ask, we don’t have a horse in the race — we build custom software, but we talk clients out of it as often as into it. Here’s the framework we actually use.

Rule one: buy commodity, build differentiator

Every business process is one of two kinds.

Commodity processes are ones you do the same way as everyone else, because there’s no advantage in doing them differently: payroll, accounting, email, calendars. Buying off-the-shelf here is correct, full stop. QuickBooks exists; do not commission a bespoke ledger. Any money spent customizing a commodity process is money spent making yourself weirder, not better.

Differentiating processes are the ones that are your competitive edge — how you quote jobs faster than competitors, how you schedule crews, the workflow behind your unusually good customer experience. Here the logic flips. Off-the-shelf software imposes its process, which by definition is everyone’s process. Forcing your differentiator into a generic tool sands off exactly the edge that wins you business.

Most companies buy everything, then express their actual competitive process in the gaps — which brings us to the diagnostic.

The spreadsheet test

Find the spreadsheet your business cannot operate without. Nearly every company has one: the monster workbook that schedules the crews or prices the quotes, maintained by one person everybody prays never leaves, emailed around in versions titled FINAL_v7_USE_THIS_ONE.

That spreadsheet is a confession. It marks the exact spot where off-the-shelf tools failed to fit your business, and where someone built custom software anyway — in the most fragile medium available. No access control, no history, no validation, one mis-sorted column from catastrophe.

You’re not deciding whether to have custom software. You already have it. You’re deciding whether it should be load-bearing duct tape or something engineered.

The five-year math

Off-the-shelf pricing is per-seat subscription. The demo price is one seat for one month; the real price is every seat, every month, compounding with headcount and the vendor’s annual increases. Fifteen people on a $40/seat tool is $36,000 over five years — for software that half-fits, plus the payroll hours spent working around the half that doesn’t.

Custom software inverts the curve: a build cost up front (typically comparable to one or two years of the subscription it replaces, when scoped to one workflow), then modest hosting and maintenance. No per-seat meter. Employee sixteen costs nothing. And it’s an asset you own, not a rental that evaporates the day you stop paying.

Neither curve wins universally — that’s the point of doing the math instead of guessing. Subscriptions win for commodity tools and small teams; ownership wins for core workflows and growing ones.

How to de-risk the build

The failure mode of custom software is the two-year everything-system that ships nothing. The fix is scope discipline: build the one workflow that hurts most — usually wherever the spreadsheet lives — get it into daily use within weeks, then expand only what proves itself. Working software beats comprehensive software, every time.

If you’ve got a FINAL_v7 spreadsheet running some corner of your business, we’d genuinely enjoy seeing it. Worst case, we’ll tell you the subscription you already pay for can do the job — it happens, and we’ll say so.