The Thirty-Day Window
Every large migration programme has a moment , usually around the third month, when the people running it privately know it is in trouble. Publicly, everything is still on track, but behind the scenes is another story.
What happens in the gap between those two versions is where migrations are actually won or lost. And in our experience, across more than forty large enterprise change programmes over the past decade, that gap opens in the first thirty days - before a single asset has moved, before the first milestone has been hit, before anyone has used the word ‘problem’ out loud.
In the majority of cases, the pattern was already in motion before execution formally began - embedded in a planning assumption, a governance decision, or an ownership structure that nobody had thought to question.
That is what makes migration failure so costly and so persistent. It is not visible when it starts. By the time it is visible, it has been compounding for months.
The organisations that navigate migrations well are not the ones with the best plans. They are the ones that learn the fastest, and have the structures in place to act on what they learn.
This paper names the failure modes directly, explains the single error that connects all of them, and describes a framework we developed to break the pattern before it becomes structural.
38%
of migration exceed their original budget, with average overruns of 23% above planned costs
Bloor Research
The Same mistake, four different ways
Before naming the patterns individually, there is one observation worth making about all of them together.
Every one of the four patterns described in this paper is a version of the same error: the attempt to manage uncertainty through control rather than through learning.
Lift and Shift Migration controls uncertainty by not changing anything - if the old environment worked well enough, replicating it keeps the risk bounded.
Fragmented Ownership emerges because teams control their own domain rather than exposing it to the uncertainty of shared decision-making.
All-or-Nothing Planning tries to resolve uncertainty on paper before it can surface in reality.
Control Instead of Enablement is the pattern named directly for the error: governance that responds to uncertainty by restricting people rather than improving systems.
None of these responses to uncertainty is irrational. Each makes sense in the moment, given the pressures any migration programme creates. The problem is that they share a structural flaw: they are all static responses to a dynamic situation. A complex migration does not stay still long enough for control to work. It requires continuous adaptation. The organisations that understand that (early, and structurally) are the ones that come out the other side with what they went in for.
The conventional response to any account of migration failure is to plan more carefully. That response is wrong. More planning does not prevent these patterns - in many cases it accelerates them. What prevents them is the combination of structured learning and the agility to act on it. Learning without agility produces insight that goes nowhere - the team knows what went wrong but the plan cannot accommodate the correction. Agility without learning produces speed without direction. The SMA is designed to create both simultaneously: each wave generates real evidence, and the programme has the structural capacity to respond to it before the next wave begins.
