Something has gone wrong with the program. Executive patience is thinning, milestones are slipping, and the usual recovery moves — more governance, a new internal communications push, a tightened implementation plan — haven’t changed the trajectory. The question now is whether this is a recovery situation or a reset situation, and that question cannot be answered without first being able to name what is actually failing.
In Brief
- Operating model transitions fail in four structural ways, and the fix for each one is structurally different.
- Applying the correct fix to the wrong failure mode is as damaging as doing nothing — and harder to detect.
- The most expensive failure mode occurs when a transition redesigns the right operating model at the wrong altitude.
- Correctly identifying the failure mode tells the sponsor not just what to fix, but whether they hold the authority to fix it.
- Three diagnostic checks — on decision pathways, team operating rhythms, and program scope — reveal which mode is in play.
Most operating model transitions that stall don’t fail for a hundred different reasons. They fail in one of four structural ways. Research tracking major program outcomes across the 2020–2024 cycle finds that close to one in five programs are cancelled outright and more than half remain challenged on scope, time, or cost — a pattern stable over a decade despite continuous methodological evolution (Arcidiacono, 2026). The structure of the failure determines the structure of the fix. That matters because the most common recovery instinct, recommitting to the existing plan with renewed vigour, is the correct response to one of the four failure modes and actively harmful when applied to the other three. Getting the diagnosis wrong costs more than the original problem.
The fix must match the failure
The four structural failure modes ZXM observes across operating model transitions share a surface presentation: slow progress, disengaged teams, eroding sponsor confidence. Their structural causes, though, sit at different layers of the organisation and require different interventions. Treating the right presentation with the wrong intervention extends the program. In some cases, it forecloses the possibility of a genuine fix.
Failure mode #1: Governance didn't follow the structure
The most common failure mode is also the easiest to miss from inside it. The organisation redesigned its structure — teams were reorganised, roles renamed, reporting lines redrawn; the decision-making authorities were never touched. Investment committees, approval thresholds, and escalation pathways still route through the old hierarchy. People are being asked to work in the new model while the authority to make decisions within it is held by governance bodies designed for the predecessor state.
The diagnostic tell is precise: teams in the new structure cannot complete meaningful work without escalating decisions upward into processes the transition nominally replaced. The new model has the appearance of authority without the substance of it.
Fixing this requires explicitly redesigning decision rights, specifying who approves what at which threshold in the new model, and amending the terms of reference of standing committees accordingly. This is not a training intervention or a change management campaign. It requires executive-level authority to change how the organisation’s governance bodies are constituted and what they are empowered to decide. If the program sponsor does not hold that authority directly, the fix requires escalation before it can begin.
Failure mode #2: Skills landed above the daily workflow
Training ran. Awareness was high. Six months after the transition was declared complete, teams had reverted. This failure mode is routinely attributed to change resistance, but the structural cause is more specific: the skills required to operate in the new model were delivered through workshops, courses, and facilitated sessions that exist above the daily workflow, not embedded in it. Managers were not brought into the new model’s decision logic. The team’s actual work was never redesigned to require the new capabilities in practice.
The fix is not more training. It is redesigning how work is structured at the team level so the new operating model’s logic appears in what people do every day, not in what they once attended. This is typically a shorter intervention than a full re-transition, but it requires a diagnostic to distinguish which teams reverted from which ones were never genuinely transitioned, a distinction standard reporting rarely makes visible.
Failure mode #3: The redesign targeted the wrong level
Of the four failure modes, this one takes the longest to name, because the work looked right while it was happening. Research on large-scale project delivery documents this same phenomenon — programs that invest substantially in executing a well-framed but incorrectly scoped intervention produce credible outputs against the wrong outcome (Flyvbjerg & Gardner, 2023). Considerable investment went into redesigning the operating model at the organisational level: its structure, governance frameworks, role definitions, and process design. But the unit of the organisation that actually needed to change was different. Sometimes it was the portfolio funding model. Sometimes it was the contractual arrangement with external vendors. The real constraint, in more cases than program leaders tend to acknowledge, was a product team’s authority to make technical decisions without escalating to an architecture committee. The redesign happened at a layer above the actual problem.
The programs that take longest to correct are rarely the ones that were poorly executed. They are the ones that executed the wrong thing well.
The tell is most visible in retrospect: when the redesigned operating model is in place and outcomes have not improved, and nobody in the program can clearly explain why a team operating correctly in the new model still cannot do the thing the transition was intended to enable. The answer is usually that the thing they cannot do was never in scope.
Going back to redesign at the right level is the only viable path. This often means acknowledging that a substantial investment addressed a genuine but secondary problem. The organisation then faces a second intervention while managing the overhead of the first. Whether this is recoverable within the current program’s runway depends entirely on how far the original work has progressed and whether the budget cycle allows reorientation. In some cases, a controlled reset is structurally preferable to attempting a mid-program pivot at the wrong level.
Failure mode #4: Deadline pressure delivered form over substance
The fourth failure mode is the one most likely to close cleanly on paper. Milestones were met. The governance dashboard showed nothing wrong when the transition was declared complete. Twelve months later, the organisation had complied with the form of the new operating model; meetings ran in the new format, reports used the new language, but decisions were still being made the old way. The substance of the transition had not occurred.
The difference is structural: a program reporting to completion is not the same as a program that changed how the organisation operates, and that distinction is rarely visible in standard progress reporting. This happens when implementation pace is driven by stakeholder timelines or political pressure rather than the organisation’s actual capacity to absorb the change. The pressure to show progress creates a compliance signal that substitutes for an operational change signal. By the time the difference becomes visible, the program has formally closed and the team has moved on.
Correcting this mode starts with a separation exercise: review each unit that reported transition-complete and distinguish those that changed how they operate from those that changed what they report. This is an activation of teams that were counted but not converted, not a full re-transition, and it is the quickest of the four fixes, provided it is attempted before the operational consequences of the compliance deficit become visible.
Naming the problem reveals your options
Correctly identifying which failure mode is present does two things simultaneously. It tells the program sponsor what kind of fix to pursue, and whether the current program has the authority, the budget, and the runway to pursue it.
The second of these is the more valuable finding. An executive sponsor who diagnoses failure mode three, wrong altitude, and recognises that the fix requires a mandate they do not currently hold is in a substantially better position than a sponsor who keeps applying the wrong intervention because the correct diagnosis felt too difficult to name. The former can make the case for a controlled reset or an escalation of authority. The latter watches the program drift until someone above them makes the decision they were avoiding.
Three checks to run this fortnight
The diagnostic sequence for a stalled operating model transition does not require an external review to begin. Three checks can be run against the program in the next two weeks.
The first: identify one team nominally operating in the new model and ask them to walk through the last substantive decision they made. Trace where the decision was escalated, who approved it, and whether that approval pathway exists inside the new model or outside it. If the approval came from a body the transition was designed to replace, failure mode one is in play.
The second: select three teams marked transition-complete six or more months ago and review their current operating rhythm: not the reports, but what a standard week actually looks like. If the cadence, the decision points, and the escalation paths are not materially different from what they were before the transition, failure modes two or four deserve focused attention.
The third: ask the program team to name specifically what the transition will enable that currently cannot be done. If the answer is vague, or if it describes a capability that sits outside the program’s current scope, the wrong-level question from failure mode three is worth examining directly.
These three checks will not complete the diagnosis, but they will identify which failure mode deserves the most attention. An executive sponsor who can name the structural failure with precision, and articulate what fixing it actually requires, is better positioned to make the case for the right next step than one who is still managing the appearance of progress.
References
- Arcidiacono, G. (2026). Comparative research on IT project failure rates: A 2025 longitudinal update. PM World Journal, XV(I). https://pmworldjournal.com/wp-content/uploads/2026/01/pmwj160-Jan2026-Arcidiacono-research-on-IT-project-failure-rate-2025-update.pdf
- Flyvbjerg, B., & Gardner, D. (2023). How big things get done: The surprising factors that determine the fate of every project, from home renovations to space exploration. Currency.
