Before You Buy Software, Fix the Process

There is a purchase order being signed somewhere in Northwest Arkansas right now for software that will not fix the problem it was bought to fix. The demo was beautiful. The sales engineer was sharp. The monthly fee seemed reasonable against the pain. And eighteen months from now the team will be running the old workarounds inside the new system, paying a subscription for the privilege. We've watched this movie inside Fortune 1 supply chains and twelve-person clinics, and it always has the same plot hole: nobody fixed the process first.

Software multiplies; it doesn't repair

Technology is a multiplier on whatever it touches. Point it at a clean process and you multiply throughput. Point it at a mess and you multiply the mess — now with license fees, migration costs, and a vendor relationship. The research on technology adoption is consistent: systems succeed when they fit how work and people actually operate, and fail — expensively — when organizations expect the tool to impose order that management never established.

A faster broken process is still broken. It just produces rejects at a more modern pace.

The tell-tale signs you're about to buy a mess multiplier

• Nobody can draw the current process on a whiteboard without an argument breaking out.
• The phrase “the system will force people to…” appears in the justification. (It won't. They'll work around it, heroically.)
• Different people give different answers to “what, exactly, will be better ninety days after go-live — and how will we measure it?”
• The loudest pain belongs to one department, but the workflow crosses four.

The order of operations that works

First, walk the job. Follow one order, one patient, one project from first contact to cash — physically, not from memory. Write down every wait, every duplicate entry, every workaround. This takes days, not months, and it is where the six-figure savings hide.

Second, fix what costs nothing. Most process repairs are subtraction: an approval nobody reads, a form keyed twice, a handoff that exists because of an org chart from 2019. Lean research has shown for decades that practice changes — not capital — drive most of the performance gap between firms. Do this before any demo, because it changes what you need to buy. Often it shrinks the requirement from a $40,000 platform to a $400 configuration.

Third, buy for the process you now have. Requirements come from your walked, fixed, written-down process — not the vendor's feature tour. Score demos against your scenarios. Call references whose operation looks like yours, not the logo customers.

Fourth, implement adoption, not installation. Go-live is a people project wearing a technology costume: champions trained first, data migrated clean, the first bad week supported in person. The system your team actually uses at month six is the only one you bought.

The honest exception

Sometimes the software is the fix — when the process is sound but drowning in volume a spreadsheet can't hold. You'll know you're in that case because the whiteboard exercise is boring: everyone agrees how work flows, and the only problem is scale. If the whiteboard session is boring, buy the software. If it's a fight, you just saved yourself the subscription — spend the savings fixing the fight.

Research: Venkatesh, Morris, Davis & Davis (2003) — “User Acceptance of Information Technology: Toward a Unified View,” MIS Quarterly, 27(3). doi.org/10.2307/30036540

Research: Shah & Ward (2003) — “Lean manufacturing: context, practice bundles, and performance,” Journal of Operations Management, 21(2). doi.org/10.1016/S0272-6963(02)00108-0