Most large transformation programmes share a recognisable trajectory. They launch with genuine ambition, visible executive sponsorship and a portfolio of workstreams. Early momentum is real. Progress is reported. And then, quietly, the pace drops. Outputs accumulate. Steering committee slides multiply. And the operating reality of the organisation remains, in most of the ways that matter, largely unchanged.
This is not a story about weak strategy or insufficient investment. The strategies are usually coherent. The investment is often substantial. What is missing is the disciplined translation of strategy into structural reality: the deliberate redesign of how the organisation works, how it makes decisions, how it deploys capability, and how it holds itself accountable for building new ones. Bain & Company research published in 2024 found that 88 per cent of business transformations fail to achieve their original ambitions.1McKinsey reports that 70 per cent of large-scale transformation efforts do not meet their objectives.2 The failure rate is not an accident. It is structural.
This article argues that the discipline required to make transformation work goes beyond programme management, leadership alignment or change communication, though all of these matter. It is the discipline of structural translation: the deliberate conversion of strategic intent into operating model design, capability architecture, decision authorities, performance systems, and a sequenced change engine that builds them with precision over time. Without that discipline, strategy remains a narrative. Execution fragments. Energy dissipates.
The gap between ambition and outcome in transformation is almost always, at root, a structural gap. And structural gaps are designable, which means they are fixable.
What 'structure' means
In the context of transformation, the word 'structure' is often misread as organisational hierarchy. Although hierarchy might be one component of structure, the broader concept is the deliberate configuration of an organisation's operating reality: how work is done, how decisions are made, where capability sits, how performance is measured and how authority is distributed across the enterprise. Changing strategy without changing structure leaves the organisation with a new direction and an old engine.
Strategic intent becomes operational reality only when it is translated across five interdependent dimensions. The operating model defines which activities are done where, by whom, and to what standard. The capability map identifies what skills, tools, and institutional knowledge must be built, acquired, or sourced. Governance and decision authorities determine who has authority to act, spend, and commit across functions and geographies. The performance system defines what gets measured, what gets rewarded, and what commands leadership attention. And the change architecture is the deliberate design of how the organisation mobilises, sequences, and sustains the transition from current to future state.
Most transformation programmes engage with two or three of these dimensions and leave the others implicit. The operating model redesign is skipped or deferred. Capability requirements are assumed rather than mapped. Decision authorities are unchanged. Performance systems continue to reward current-year delivery. The change architecture is a project plan dressed in programme governance clothing. The result is a programme that is technically active and structurally inert, producing outputs without producing change.
The missing translation layer
The most common structural failure in transformation is the shortcut between strategy and execution. Strategy is declared. A programme office is stood up. Workstreams are populated and milestones are set. The step that is systematically skipped, quietly and without anyone making an explicit decision to skip it, is the operating model redesign that should sit between strategic intent and initiative portfolio.
An operating model redesign asks a fundamentally different set of questions from a programme plan. Not "what initiatives will we run?" but "how does this organisation need to work when the strategy is achieved?" What processes look different? What decisions move where? Which capabilities are core versus sourced? What does the governance model look like at the other end? These questions are harder and more politically exposed than populating a workstream. They are also the questions on which everything downstream depends.
Strategy translated directly into a programme portfolio is strategy deferred, not strategy delivered.
Microsoft's transformation under Satya Nadella, which began in 2014, is one of the more instructive examples of structural discipline in practice. Nadella inherited an organisation structured around competing product lines, each with its own P&L and incentive system: internally fragmented, collectively slow, and rewarding behaviour, through stack ranking, that made cross-unit collaboration structurally irrational. His response was not a transformation programme. It was an operating model redesign.
He eliminated the Windows division as an autonomous strategic unit, reorganised around shared platform capabilities (cloud, AI, developer tools), changed the organisation's decision-making architecture, and dismantled the stack ranking system that had been corrupting incentive alignment for a decade. These were structural choices, not cultural ones. Culture shifted as a consequence of structural change, not as its precondition. Microsoft's market value grew from approximately $300 billion in 2014 to more than $2.5 trillion by 2023. The operating model redesign preceded the growth.
The translation layer is the operating model design. Programmes that skip it are not transformation programmes. They are improvement programmes with transformation ambitions, and the gap between the two is where most of the failure lives.
The two systems problem
Every organisation undergoing transformation faces a structural contradiction. It must continue to deliver today's business, serving customers, hitting targets and managing risk, while simultaneously building tomorrow's: redesigning processes, developing new capabilities, and embedding new ways of operating. These are not the same activity. They require different time horizons, risk tolerances, skills, and leadership behaviours. Treating them as one system is one of the most reliable ways to fail at both.
John Kotter, writing in Harvard Business Review, called this the dual operating system: the need to run a hierarchy optimised for today's delivery while simultaneously operating a more agile, network-based structure oriented toward strategic build.3 Kotter's key insight was that the two systems cannot be separated or sequenced. They must operate simultaneously, with clear boundaries, clear roles, and constant information flow between them.
In practice, most organisations manage this poorly. The change system is either absorbed into the run system, with transformation leads carrying operational responsibilities alongside their change mandates and the latter consistently displaced by the former, or it is isolated from it, operating as a programme structure with no real authority over the operating business it is supposed to change. In either case, the transformation lacks the structural authority it needs to make decisions stick.
The structural requirement is explicit and binary. People need to know whether they are delivering today's business, building tomorrow's, or both, and if both, in what proportion, and with what protection for their change-system time. Leaders need to actively protect the attention and authority of those in the change system from being colonised by the run system. Without that clarity and protection, the run system wins by default. It always does: its rewards are immediate, its accountability is visible, and its demands are relentless.
Structural failure modes
When structural discipline is absent, transformation fails in predictable patterns. Three are consistent enough to name.
Conceptual misalignment. The future state has been described as a strategy or a vision but never translated into an operating design. There is no explicit picture of how the organisation will actually work when the transformation is complete: what decisions will be made where, what the governance model looks like, what capabilities exist and where they sit. Without that design, every initiative team makes its own assumptions about the destination. The initiatives may individually succeed and collectively fail to add up to anything coherent.
Accountability diffusion is not shared ownership. It is dissolved ownership, and the transformation pays the price.
Accountability diffusion. Transformation that matters, whether building a new capability, redesigning an end-to-end process or shifting a revenue model, requires a single identifiable owner with the authority, resources, and mandate to deliver it. In most programmes, this accountability is distributed across workstream leads, steering committees, functional heads, and sponsors who each hold a partial view. Bain research shows that 90 per cent of the value in a transformation comes from fewer than five per cent of the roles within it. Diffusing accountability across the remaining 95 per cent does not distribute ownership. It dissolves it.
Weak change architecture. A programme governance structure and a change architecture are not the same thing. Governance manages milestones and escalation. A change architecture designs the sequence, pace, and logic of the transition: what needs to change first because everything else depends on it, what capabilities must be in place before the next wave of change can land, what signals in the run system will indicate that change is taking hold. GE's transformation attempts under Jeff Immelt illustrate the failure mode: acquisitions were made without the integration architecture to absorb them, governance was insufficiently independent to challenge strategic assumptions, and accountability for outcomes was distributed across a leadership layer that could not function as a coherent decision-making unit. Over Immelt's tenure, GE lost approximately 75 per cent of its market value. Programme activity was never the problem. Structural accountability was.
What structural discipline requires
Structural discipline is not a methodology or a framework. It is a set of deliberate choices that most leadership teams are reluctant to make, not because they do not understand them, but because they require surfacing the underperformance the current structure has been obscuring, and committing to a future state design before the organisation feels ready to be held to it. The choices are demanding. They are also the ones that determine whether transformation delivers.
An explicit future-state operating design. Not a vision statement or a capability aspiration, but a designed picture of how the organisation will work: which decisions will be made where, what the governance model looks like, how performance will be managed, what capabilities will sit where. This design does not need to be exhaustive at the outset. It needs to be explicit enough to discipline the initiative portfolio and give the change architecture a coherent destination.
Decision authorities clarity. Transformation involves decisions that cut across function, geography, and hierarchy. Without explicit authority (who can commit resources, who can override local priorities and who can design process across organisational boundaries), every significant decision becomes a negotiation and every negotiation introduces delay. Organisations that implement structured decision governance consistently outperform those that do not: Bain research on the RAPID decision-making model shows that organisations with clear decision authorities grow revenues and earnings at more than three times the rate of peers who rely on informal authority.4
Incentive realignment. Performance systems that reward current-year delivery, functional efficiency, or activity volume will consistently undermine transformation investment. Leaders measured on the run system will protect the run system, rationally, individually and at collective cost. Incentive realignment does not require dismantling performance management. It requires redesigning it to reward the build, not just the delivery: milestone-based recognition for capability build, visibility for change-system contributions, and long-horizon measures for transformation leaders.
A dedicated change authority with real power. This is not a programme management office. It is a structural role with direct access to the CEO or board, authority to make or recommend design decisions across functions, control over resources allocated to transformation, and accountability for outcomes, not outputs. Bain research indicates that disciplined management of internal risk, specifically the kind of structural ownership this role enables, accounts for 80 per cent of transformation success, and that leadership teams who exercise it are nearly twice as likely to achieve their ambitions and three times as likely to sustain the change they produce.5
Sequenced capability build. Transformation cannot build all required capabilities simultaneously. The change architecture must identify what is foundational, the capabilities on which everything else depends, and sequence the build accordingly. Data infrastructure before analytics deployment. Governance design before AI workflow integration. Process ownership before automation. Attempting to build everything in parallel produces a long list of partially developed capabilities and no structural momentum.
Why leaders avoid structural discipline
If structural discipline is this important, why is it so consistently absent? The answer is not ignorance. Most senior leaders understand the argument. The answer is that structural discipline is politically and personally costly in ways that programme management is not. Standing up a programme signals ambition without redistributing power. Designing the operating model does both.
Designing the future-state operating model requires making decisions about authority: who will have it, who will lose it, which functions gain scope and which are diminished. It requires identifying the underperformance that the current structure has been concealing. It requires releasing the organisation's best people from run-system roles, where their contribution is visible, measured and rewarded, into change roles where success is harder to attribute and the personal cost of failure is real. It requires binary decisions (this capability is core, that one is not; this decision sits here, not there) in environments where ambiguity is often preferred because it preserves existing power arrangements and avoids the discomfort of explicit accountability.
Most programmes are designed to signal ambition without making the structural choices that would make ambition achievable.
The consequence is programmes that are large enough to demonstrate seriousness and small enough to avoid the structural choices that would make them matter. BCG analysis of more than 850 companies found that only 35 per cent of transformation initiatives reach their stated goals.6 Structural avoidance is not incidental to that failure rate. It is one of its primary causes.
The strategic consequence
The organisations that get this right share a common pattern. They design before they mobilise. They make the future-state operating model explicit before they stand up the initiative portfolio. They define the decision authorities before appointing workstream leads. They realign incentives before they ask leaders to invest in the build. They establish a change authority with genuine power before asking the organisation to change direction. And they manage the boundary between the run system and the change system with deliberate, sustained attention throughout the transformation.
This matters beyond transformation for its own sake. As AI becomes embedded in enterprise operations, automating decisions, orchestrating workflows and redesigning processes at scale, the governance and structural discipline required becomes more demanding, not less. Intelligent systems operating at enterprise scale require exactly what structural transformation requires: clear decision authority, explicit accountability, designed governance, and a sequenced approach to capability build. The organisations that have developed structural discipline as a genuine competence will govern intelligent systems more effectively than those that have not. Transformation capability and AI governance capability are, at their structural core, the same discipline. This is the work of Transformation Leadership and Delivery.
The gap between ambition and outcome in transformation is real and persistent. But it is not inevitable. It is structural, which means it is designable. The discipline required is not complexity. It is clarity: about what the future state looks like, about who owns what, about how the organisation will be built toward it, and about the authority needed to make those choices stick. That clarity is available to every leadership teamwilling to choose it over the comfort of a well-populated programme plan.
References
- Bain & Company (2024). 88% of Business Transformations Fail to Achieve Their Original Ambitions. www.bain.com/about/media-center/press-releases/2024/88-of-business-transformations-fail-to-achieve-their-original-ambitions-those-that-succeed-avoid-overloading-top-talent/
- McKinsey & Company (2024). Common Pitfalls in Transformations. www.mckinsey.com/capabilities/transformation/our-insights/common-pitfalls-in-transformations-a-conversation-with-jon-garcia
- Kotter, J. (2012). Accelerate! Harvard Business Review, November 2012. hbr.org/2012/11/accelerate
- Bain & Company (2023). RAPID: Bain's Tool to Clarify Decision Accountability. www.bain.com/insights/rapid-decision-making/
- Bain & Company (2024). Build an Organisation That Decides and Delivers. www.bain.com/insights/decision-insights-4-build-an-organization-that-decides-and-delivers/
- BCG (2024). How CEOs Can Beat the Transformation Odds. www.bcg.com/publications/2024/how-ceos-can-beat-the-transformation-odds