Should we test a transformation phase before funding it?
Yes, and the reason is specific to transformation rather than a general principle about testing.
When you build a new product the risk is demand: will anyone want it. When you change how an existing organisation works, demand is not usually in doubt. The work exists, the process exists, people are doing it today. The risk is adoption: will the people who have to change how they work actually do it?
That risk does not show up in the evidence transformation programmes normally collect. Interviews, workshops and benchmarking tell you what people say they will do. What they do is a different thing, and the gap between the two is where programmes disappear, into workarounds, shadow spreadsheets, and a system everyone has technically adopted and nobody actually uses.
The only evidence that closes that gap is watching the people whose jobs change use something real, for a task they recognise.
The usual objection is that you cannot test it until it is built, and building it is the expensive part. That stopped being true. A working prototype of the changed process can be built in days rather than months (from £750 in five days), which against a phase measured in six figures makes it the cheapest gate you can put in front of a funding decision.
What that buys you is not reassurance. It is the ability to be wrong about something early, while changing your mind is still cheap and still politically survivable. Every programme of that size turns out to be wrong about something; the only variable is when you find out.
So a funding gate worth having asks for three things before the next phase is released:
- The question that phase was meant to answer, written before it started
- Evidence of what people did, not what they said in a workshop
- A genuine option to stop, agreed at the outset while stopping is still a legitimate outcome