Eighteen months into an operating model transition, the milestone tracker shows green across every workstream. Governance forums have met on schedule. Roles have been restructured and reporting lines redrawn. Process redesign workshops are complete. The program board reports confidence. None of it shows in the numbers. Case resolution rates, processing times, the speed at which policy decisions reach implementation: whatever the transition was designed to improve has stayed flat.
In Brief
- Operating model transitions stall for four structural reasons, each requiring a different intervention — redesign, resequencing, re-sponsorship, or wind-back.
- Milestone governance measures activity rather than outcomes, which hides the structural cause of the stall from the program board.
- Applying the wrong intervention consumes the organisation's appetite for change while leaving the actual structural condition intact.
- A fixed diagnostic sequence — design logic, then sequencing, then sponsorship, then premise — identifies which mode is operating.
McKinsey & Company (2021) found that less than one-third of organisational transformations succeed at both improving performance and sustaining those improvements over time. The instinct at the eighteen-month mark is to search for an execution problem: a team underperforming, a vendor falling short, a change management workstream that needed more funding. That instinct treats every stall as the same stall, and it is almost always wrong. Execution failure is a symptom, not a diagnosis. It is produced by one of four structural conditions, and each condition requires a fundamentally different intervention.
Milestones hide why outcomes stall
Execution looks like the explanation because the reporting system was built to track activities, not outcomes. Milestone governance measures whether the program’s planned steps have been completed: role changes made, processes redesigned, technology deployed. When those milestones report green, the natural conclusion is that the transition is working and something else must be preventing results. The structural cause sits below that reporting layer.
Operating model transitions fail for four structurally distinct reasons, and each one produces a pattern that looks like an execution gap from above. All four are serious; what separates them is the type of intervention each requires: redesign, resequencing, re-sponsorship, or an honest wind-back that most program boards are unwilling to name. Applying the wrong intervention does not produce a partial fix. It deepens the stall, because the structural condition that produced the symptoms keeps operating while resources and attention get diverted to a problem that does not exist.
Way #1: The design itself is wrong
Most operating model failures begin here, and it is the hardest one for program boards to accept. The model that was designed and approved does not produce the intended outcomes. The design was wrong. The logic connecting the new structure to the intended service improvements was never validated before the transition began.
Milestones are green because they measure whether the new structure has been stood up, but the relationship between that structure and the outcomes it was supposed to improve was assumed rather than tested. In most cases, the design was built from an analysis of what was wrong with the old model rather than from a clear theory of how the new model would produce different results. The intervention is to redesign the model itself, grounded this time in the causal logic that connects structure to outcome.
Way #2: Right design, wrong sequence
Sometimes the model is sound and the problem is sequencing. Dependencies between changes were not mapped, so one part of the new model requires another part to be in place before it can function, and the transition plan sequenced them independently. Individual workstreams complete on time while the system they are meant to compose remains incoherent.
The tell is visible in the joins. Parts of the new model work well in isolation, but the connections between them produce friction, workarounds, or outright failure. Teams restructured under the new model are still reporting into governance structures designed for the old one. The fix is resequencing: identifying which changes need to land before others can take effect, and reordering the remaining transition work accordingly.
Way #3: The sponsor left mid-transition
Losing a sponsor mid-transition works differently from the first two. Operating model transitions require structural decisions that only a senior sponsor with sufficient authority can make: decisions about which functions lose headcount, which reporting lines are permanently altered, which legacy processes are discontinued rather than reformed. These decisions tend to sit in the second half of most transitions because the first half is consumed by design and early implementation. When the sponsor who approved the transition moves to a new role, retires, or is reassigned before those decisions are made, the authority to make them does not automatically transfer.
What follows is distinctive. The program continues, milestones keep being met, but the structural decisions that would make the new model real are deferred, diluted, or redesigned to avoid the conflict they were meant to resolve. The decisions most likely to be dropped are the ones that require someone to say no to an existing power structure. McKinsey & Company’s (2021) finding that 78% of transformations succeed when all critical actions are implemented, compared with 31% overall, reflects this pattern directly.
Re-sponsorship means something more specific than a handover at a steering committee. It requires identifying the structural decisions that remain unmade and explicitly assigning a single, accountable owner with the authority to make them.
Way #4: The preconditions were absent
Program boards are least willing to name this one. The transition was designed for an organisation that does not exist: one with a level of digital maturity, workforce flexibility, or process standardisation that was assumed in the business case but was never actually present. Design, sequencing, and sponsorship can all be intact, and the organisation still cannot absorb the change because the conditions the transition depends on were never there. The business case projected outcomes from a maturity baseline the organisation had not reached.
Resistance in this mode does not respond to change management because the problem is structural capacity, not behavioural adoption. The honest intervention is a wind-back: acknowledging that the preconditions need to be built before the transition can proceed, scaling back to the elements the organisation can absorb, and redesigning the transition timeline around the organisation that actually exists.
The wrong fix deepens the stall
Naming the right mode changes what happens next. A program board that identifies a design failure and responds with resequencing will produce a more efficiently ordered implementation of a model that does not work. Conversely, treating a sponsorship gap as a design problem leads to discarding a model that was sound and replacing it with one that is politically safer but structurally weaker. The intervention must match the mode, and that match depends on correctly identifying the structural cause.
The cost of misdiagnosis is not stagnation; it is active damage, because every mismatched intervention consumes the organisation’s finite appetite for change while leaving the structural condition intact.
When the wrong fix is applied, the organisation’s willingness to sustain the transition erodes with each failed attempt. Senior leaders lose confidence in the program. The next round of structural decisions becomes harder to approve because the previous round appeared to produce nothing. The distance between the executive’s intent and the organisation’s operating reality grows, and by the time the correct diagnosis is reached, the political and operational conditions for acting on it may no longer exist.
What this means for senior leaders
Zen Ex Machina applies a diagnostic sequence to stalled operating model transitions that runs in a fixed order:
- Test the design logic first. Whether the new model’s connection to intended outcomes was assumed or validated. A design gap means the model itself needs rework before anything else changes.
- If the design holds, test sequencing. Whether the dependencies between changes are mapped and ordered, or whether completed workstreams are waiting on elements that have not yet landed.
- If both hold, test sponsorship. Whether the person with authority to make the remaining structural decisions is still in the role, still engaged, and still willing to make the calls that require saying no.
- If design, sequence, and sponsorship are all intact, test the premise. When outcomes remain flat despite all three being sound, the remaining explanation is the one most program boards will not name voluntarily — the organisation the transition was designed for does not yet exist.
The executive who identifies which of the four modes is operating is already in a different position to one who is still searching for an execution failure that is not there.
References
McKinsey & Company. (2021). Losing from day one: Why even successful transformations fall short. https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/successful-transformations
