Transformation programmes rarely fail at the top. The board is aligned, the business case is credible, the budget is approved. They rarely fail at the bottom either; the software usually works.
They fail in the middle: at the layer of managers and specialists who are asked to change how their work is done while remaining fully accountable for this quarter's numbers. That group has every incentive to keep the old process running quietly alongside the new one, and that is exactly what happens.
The parallel process is the failure mode
You will know it by the spreadsheet. The new system is in place, adoption reports look reasonable, and someone in operations is still maintaining a workbook because the platform does not handle the exception that occurs eight times a month. Two sources of truth, and the shadow one is more accurate.
The instinct is to mandate compliance. The better response is to go and look at the exception. Nine times out of ten it is a real part of the business that the design ignored, and the spreadsheet is doing legitimate work.
Shadow processes are not resistance. They are unpriced requirements.
What we do differently
- Discovery includes the exceptions, deliberately. We ask what happens on the worst Tuesday of the month, not the standard one.
- The people who will use the system every day are in the room when the flow is designed, not shown it at training.
- We run old and new in parallel until the outputs reconcile, then cut over with evidence rather than an announcement.
- Middle managers get their reporting rebuilt first, because that is the work the change threatens to make harder.
That last point matters more than it sounds. A manager whose weekly report becomes harder to produce will resist a platform indefinitely, and they will be right to. Build their reporting into the first increment and you convert your most expensive opponent into your most useful advocate.
Sequence beats scope
The programmes that succeed are the ones sequenced so that something real is in production within a quarter. Not a pilot in a sandbox, but a genuine slice of the operation running on the new platform, with the old one retired for that slice. Confidence compounds from there.
Eighteen-month programmes with a single cutover at the end fail for a structural reason: by the time anything reaches users, the business that commissioned it has changed. Deliver in slices and each one gets tested against reality while it is still cheap to be wrong.
About the author
Bonolo Mathabela
Bonolo co-founded Bonang Technologies to prove a specific point: that a small, senior team with clear standards ships better software than a large one with none. As COO they own how the studio runs: how work is estimated, how it is staffed and reviewed, and what is allowed to reach production.
