Transformation programmes are almost universally measured against their original business case — did the programme deliver the cost savings, the efficiency gains, the specific metrics used to justify the investment. This is necessary and, in my experience, insufficient for understanding whether a transformation actually succeeded in the way that matters most.
A programme can hit every business case metric while leaving the underlying decision-rights ambiguity that originally motivated it completely unaddressed — the new system is live, the cost target is met, and the same governance confusion that made the old process slow and error-prone has simply relocated into the new structure, waiting to resurface.
Evidence Architecture, done honestly, has to measure this too: not just whether the stated business case metrics were hit, but whether the actual structural problem — the decision that used to take three weeks and shouldn't have — genuinely improved, specifically and measurably, against its own baseline.
This requires a different kind of post-implementation review than most institutions currently run: not a business-case reconciliation exercise, but a genuine re-diagnosis, asking whether the specific decision-rights problem identified at the outset of the engagement has actually been resolved, or whether it's simply been redescribed in the language of the new operating model.
The transformations that hold up over time are the ones willing to measure this uncomfortable question honestly, including in cases where the business case metrics were hit but the underlying problem wasn't — because a programme that meets its numbers while leaving the real issue intact hasn't actually transformed anything. It's relocated the same problem into more expensive infrastructure.