Advanced strategy · Module 6

Strategy execution, change and crisis

Why do sound strategies fail in execution, how can organizations change without breaking, and what separates a crisis handled well from one that becomes a catastrophe?

Beyond the IB course UndergraduateMBAProfessional About 30 minutes · 3,129 words · checked 16 November 2026

From the IB course to this module

The IB course covers resistance to change and the human resource strategies that reduce it (2.1), cultural clashes when organizations change (2.5, HL), and crisis management and contingency planning, with transparency, communication, speed and control as the factors in an effective response (5.7, HL); the toolkit adds force field analysis. This module connects the three: it explains why strategies fail between decision and result, compares theories of change and the evidence for them, and treats crises as products of the systems and decisions that precede them.

Section 01: Conceptual foundation

Strategy is realized through thousands of decisions taken after the plan is approved: what to fund, whom to hire, which projects to stop, how to measure progress. A gap between intended and realized strategy is normal. Mintzberg and Waters (1985) showed that realized strategy combines the deliberate part of what was intended with emergent patterns that no one planned. Execution is the discipline of closing the gap where the plan was right and learning from it where it was not.

Organizational change is the hardest kind of execution, because it asks people to stop doing what has made them successful. Approaches differ on whether change should be driven from the top by economic targets or built through the organization's own capabilities, and on whether it can be planned in stages at all.

A crisis is a low-probability, high-impact event that threatens an organization's viability, with ambiguous causes and pressure to decide quickly (Pearson and Clair, 1998). Crises are rarely as sudden as they look. Most grow from weaknesses in systems, decisions and culture that existed long before, which is why crisis management begins with execution and governance, not with press releases.

Section 02: Advanced explanation

Where execution breaks

Studies of execution point to three recurring failures. The first is resource allocation: budgets, people and management attention keep following the old business, because they are allocated by routine rather than by the strategy. The second is coordination: Sull, Homkes and Sull (2015) found that most managers could rely on their bosses and direct reports, but far fewer on colleagues in other functions and units, so execution failed at the boundaries. The third is commitment without learning: once a plan is announced, deadlines become promises, and evidence that the plan is failing is discounted.

The planning fallacy makes the third worse. Lovallo and Kahneman (2003) argued that managers forecast from the inside view, building up from the specifics of their own project, and so underestimate time and cost. The remedy is the outside view: a forecast anchored on what happened to a reference class of similar projects. If comparable programmes overran their schedules by a median of 40%, an 18-month plan should be expected to take about 25 months, and a public commitment to 18 months should be treated as a risk in itself.

Measurement can help or hide. A programme reported as on track while its components slip, sometimes called a watermelon report (green outside, red inside), is the usual sign of incentives that reward reporting progress over revealing problems.

Approaches to change

Lewin's (1951) model of unfreezing, changing and refreezing underlies most planned-change thinking: the forces that hold behaviour in place must be weakened before new behaviour can be learned and stabilized. Kotter (1996) expanded it into eight stages, from establishing urgency and forming a guiding coalition to anchoring new approaches in the culture, and argued that skipping stages is the commonest cause of failure.

Beer and Nohria (2000) contrasted two theories of change. Theory E pursues economic value through top-down decisions, restructuring and incentives, and measures success by shareholder returns. Theory O builds organizational capability through participation, culture and learning, and measures success by the organization's capacity to change again. Theory E produces fast financial results and often erodes commitment; Theory O builds commitment slowly. They argued that the most successful changes combined the two, typically by confronting the financial facts first while involving people in designing the response.

Under threat, organizations tend to centralize control and narrow the information they consider, a pattern Staw, Sandelands and Dutton (1981) called threat rigidity. It helps when the threat is familiar and fast action is needed, and harms when the threat is new and demands learning. This is the central tension of turnarounds: urgency is needed to unfreeze, but fear makes people defend the routines that must change.

The anatomy of a crisis

Perrow (1984) argued that in systems with interactive complexity, where components interact in unexpected ways, and tight coupling, where there is little slack or time between a failure and its consequences, serious accidents are normal: no amount of care removes them entirely. Large IT systems, power grids and financial networks share these properties. The strategic implication is to reduce coupling where possible, through phased change, buffers and fallbacks, and to plan for failure where it is not.

Research on high-reliability organizations (Weick and Sutcliffe, 2001) is more optimistic: some organizations that run hazardous systems achieve very low failure rates by attending to small failures, resisting simplified explanations, staying close to front-line operations, building resilience and deferring to expertise rather than rank. Both views treat reliability as a property of the whole system, not of individual vigilance.

When a crisis breaks, the quality of the response depends on the first hours. The IB course's four factors (transparency, communication, speed and control) conflict in practice: speaking fast risks speaking inaccurately, and full transparency can create legal exposure. Coombs's (2007) situational crisis communication theory recommends matching the response to the responsibility stakeholders will attribute: when the organization is seen as responsible, accepting responsibility, compensating and correcting protect its reputation better than denial or diversion.

Governing large change programmes

Boards approve strategies and transformations but rarely see their execution directly. Three practices separate programmes that fail safely from those that fail in front of customers. Stage gates with explicit criteria: a migration, launch or integration proceeds when defined tests are passed, not when a date arrives. Independent assurance: someone without a stake in the timetable reports to the board on readiness. Pre-agreed fallbacks: the conditions for stopping, rolling back or phasing the change are set before the pressure of the moment.

Each practice costs time and money, and feels like caution in a programme sold on urgency. The cost should be weighed against what follows when a tightly coupled change fails in public: redress, regulatory penalties, management distraction and reputational damage that can last for years.

Section 03: Models and frameworks

Where each comes from, how to use it, and where it stops being useful.

Mintzberg and Waters (1985)

Deliberate and emergent strategy

Realized strategy combines deliberate strategy, the part of the intended plan that is carried out, with emergent strategy, patterns that form from actions no one planned. Unrealized intentions are the remainder.

How to use it

Compare the announced strategy with the pattern of decisions actually taken, using budgets, hires, launches and exits as evidence. Decide which emergent patterns reflect learning worth adopting and which reflect drift, and which parts of the plan need firm commitment and which should stay open to adjustment.

Where it fails

The distinction is easier to draw afterwards than in the moment, and it can be used to excuse poor execution as emergence. It says little about how much commitment a strategy needs to be credible to employees, investors and rivals.

Henry Mintzberg and James A. Waters (1985). Of strategies, deliberate and emergent. Strategic Management Journal 6(3).

Kaplan and Norton (1996)

The balanced scorecard

The balanced scorecard translates strategy into linked measures across four perspectives (financial, customer, internal process, and learning and growth), on the logic that capabilities drive processes, processes drive customer outcomes and customer outcomes drive financial results.

How to use it

Start from the strategy, not from the data available: state the causal chain the strategy assumes, choose a few measures and targets for each link, and assign owners. Review the measures in strategy meetings and test whether the links hold; process measures that improve while customer measures do not are evidence that the strategy's theory is wrong.

Where it fails

Scorecards often become long lists of indicators without causal logic, and measures can displace the goals they were meant to track. The links between perspectives are hypotheses, and the evidence that scorecards improve performance is mixed.

Robert S. Kaplan and David P. Norton (1996). The Balanced Scorecard: Translating Strategy into Action. Harvard Business School Press.

Kotter (1996)

Kotter's eight stages

Kotter's stages run from establishing urgency, forming a guiding coalition and creating a vision, through communicating it, empowering action and generating short-term wins, to consolidating gains and anchoring the change in the culture. They elaborate Lewin's unfreezing, changing and refreezing.

How to use it

Use the stages as a checklist for a change plan and as a diagnostic for one that is stalling: which stage was skipped or rushed? Combine them with a force field analysis of what holds current behaviour in place, and plan short-term wins that are visible, unambiguous and clearly linked to the change.

Where it fails

The model is linear, while real change is iterative and political. It rests mainly on its author's experience rather than systematic tests, and its emphasis on urgency can provoke the threat rigidity it is meant to overcome.

John P. Kotter (1996). Leading Change. Harvard Business School Press. Kurt Lewin (1951). Field Theory in Social Science.

Beer and Nohria (2000)

Theory E and Theory O

Theory E changes an organization to raise economic value quickly, through top-down restructuring and incentives; Theory O changes it to build capability, through participation and learning. They differ in goals, leadership, focus, process, rewards and the use of consultants.

How to use it

Make the choice explicit for each part of a programme. Use Theory E where the financial facts are urgent and the required actions clear, and Theory O where change depends on new behaviour and judgement. Sequence them deliberately, for example by confronting the economics first and then involving people in redesigning the work.

Where it fails

The theories are ideal types, and combining them is easier to recommend than to do, since the trust that Theory O needs can be destroyed by the cuts Theory E demands. The evidence for combined approaches comes largely from case studies.

Michael Beer and Nitin Nohria (2000). Cracking the code of change. Harvard Business Review 78(3).

Perrow (1984)

Normal accidents and crisis response

Systems that are interactively complex and tightly coupled will produce serious failures despite everyone's efforts. When one occurs, the response should match the responsibility stakeholders attribute to the organization.

How to use it

Before a major change, map where the system is tightly coupled, with no slack and no time to intervene, and reduce the coupling through phasing, buffers and fallbacks, or invest in detection and recovery where it cannot be reduced. Prepare crisis roles, decision rights and holding statements in advance, and judge in the moment how much responsibility customers will attribute.

Where it fails

Perrow's theory explains why accidents happen better than how to prevent them, and high-reliability research suggests some organizations do far better than it predicts. Crisis communication theory is tested mainly in experiments with hypothetical crises.

Charles Perrow (1984). Normal Accidents: Living with High-Risk Technologies. Basic Books. W. Timothy Coombs (2007). Protecting organization reputations during a crisis: the development and application of situational crisis communication theory. Corporate Reputation Review 10(3).

Section 04: Real business applications

Real organizations appear only with dated, cited facts; illustrative examples are labelled.

Real organization, 2022-12

TSB: a migration timetable that outran its testing

What the regulators found. Sabadell acquired TSB in July 2015. The FCA's final notice records that the rationale included the financial returns expected from moving TSB's IT services off the Lloyds Banking Group platform it still used onto a new UK version of Sabadell's own, with the aim of migrating by the end of 2017. The programme fell behind from the start. On 20 September 2017 TSB decided it had to re-plan; nine days later, before the re-plan was complete, it announced publicly that it would migrate in the first quarter of 2018. Testing phases ran late and in parallel, and testing concluded the day before the decision to migrate. TSB relied on a Sabadell subsidiary to build and run the platform without sufficiently understanding its capability to operate it. The migration took place over 20 to 22 April 2018, and failures in digital, telephone and branch services followed. TSB did not return to normal operation until 10 December 2018, received 225,492 complaints in the following year and paid £32.7 million in redress. In December 2022 the FCA and PRA fined it £48.65 million in total.

What it suggests. A synergy promised at acquisition became a date, the date became a public commitment, and the commitment compressed the testing that would have exposed the problems. The system was tightly coupled, so failures cascaded across channels, and, as the regulators noted, the board did not sufficiently challenge the timetable.

Real organization, 2026-05

BP: resetting a strategy, and the test of delivery

What the company reported. In February 2025 BP announced that it had fundamentally reset its strategy: oil and gas investment of about $10 billion a year; investment in transition businesses of $1.5 to 2 billion a year, more than $5 billion a year below previous guidance; $20 billion of divestments announced by the end of 2027; and the retirement of every aim first announced on 4 August 2020. In December 2025 it agreed to sell 65% of its Castrol lubricants business to Stonepeak at an enterprise value of $10.1 billion, bringing completed or announced divestments to about $11 billion; its net debt was $26.1 billion at the end of the third quarter of 2025, against a target of $14 to 18 billion by the end of 2027. The same month its chief executive stepped down, with a successor from outside the company appointed from April 2026. In May 2026 the board removed its chair, citing concerns about governance standards, oversight and conduct, and reaffirmed the strategy.

What it suggests. The aims of 2020 were a deliberate strategy that the company itself later retired and replaced. After a reversal, credibility becomes part of execution: investors and employees judge the new plan by delivery against hard targets, such as divestment proceeds and net debt, and turnover at the top raises the stakes of each milestone.

Section 05: Case studies

Fictional case, simulated figures

Meridian Mutual's go-live decision

Meridian Mutual is a fictional insurer with 1.8 million customers. Two years ago its board approved replacing its thirty-year-old policy and claims systems, with go-live at the end of month 12 of the final phase, a date announced to customers and to its regulator. The business case rests on switching off the old systems: each month of delay costs about €3 million in savings forgone and in running both systems. If go-live slips past month 12, the next suitable window is five months later, when renewal volumes fall.

The programme director reports the programme as amber and on track, and the supplier says it is ready. A non-executive director has asked for the evidence. The exhibit shows the status, at month 10, of the five workstreams that must be complete before go-live.

Meridian Mutual: readiness of the migration workstreams at month 10
Workstream Planned completion (month) Forecast completion (month) Tests passed / planned Pass rate
Data migration rehearsals9114 / 580%
Performance testing at peak load10136 / 1060%
Branch and call-centre training111238 / 4095%
Supplier operational readiness10127 / 978%
Rollback rehearsal11Not scheduled0 / 10%

Simulated data. Go-live is planned for the end of month 12.

Questions

  1. Which workstreams put go-live at the end of month 12 at risk, and which failures would be hardest to recover from once customers are on the new system?
  2. What does a delay to the next window cost? How large, and how likely, would the losses from a failed go-live need to be to justify that delay?
  3. What stage-gate criteria should the board set for the go-live decision, and who should report against them?
  4. Draft the first public statement Meridian would make if go-live failed. Which elements of Coombs's guidance apply?

Section 06: Critical evaluation

What the evidence supports, where the debate is, and what the models leave out.

How often does change fail? The claim that 70% of change programmes fail is repeated widely, but Hughes (2011) traced it and found no valid and reliable empirical basis for it. Change is difficult and many programmes miss their goals, but the precise figure is folklore. The distinction matters, because a belief that most change fails can justify either fatalism or excessive force.

Do the models work? Kotter's stages and Lewin's sequence are widely used and persuasive, but they rest mainly on experience and case studies; systematic tests are few, and practitioners report that change rarely moves in neat stages. Beer and Nohria's contrast is a useful diagnostic rather than a tested theory. The balanced scorecard is the best studied of the three, and the evidence that adopting it improves performance is mixed, depending on whether it is used to test strategy or only to report on it.

Alignment or adaptability? One view holds that execution fails for lack of alignment; another that too much alignment suppresses the learning emergent strategy needs. Sull and colleagues found coordination across units to be the more common failure. Interpretation: commit firmly to a few priorities and leave room to learn how to deliver them.

Can crises be prevented? Perrow and the high-reliability researchers disagree about how far design and culture can prevent failure in complex systems. Both support the reading of the TSB case: the decisive choices were made long before the crisis, in timetables, testing and governance. What is established in this module's cases is what the regulators and companies documented; the causal readings offered here are interpretation consistent with those documents, not findings.

Section 07: Decision exercise

A fictional situation. Decide before you read the guidance.

Calder Foods: a possible contamination on a Friday evening

Calder Foods is a fictional maker of chilled ready meals, sold under its own brand and supermarkets' labels. At 6 p.m. on a Friday its quality team finds a fragment of metal in a sample from one production line; a sensor that should have stopped the line was bypassed during maintenance three days ago. About 120,000 meals from that line, under four brands, are in shops or customers' homes. No illness or injury has been reported. Supermarkets expect to hear within hours of any safety concern, and the regulator must be told of any unsafe food on sale.

  1. Recall everything from the line tonight

    Notify the regulator and supermarkets tonight and recall all 120,000 meals with a public notice. This protects customers fastest and accepts responsibility clearly, at a cost of about €1.5 million and headlines naming Calder.

  2. Withdraw first, recall if confirmed

    Ask supermarkets to remove the products from sale tonight and notify the regulator, but delay a public recall until tests on retained samples confirm the scale, within 36 hours. This disrupts less if the fragment was isolated, but meals already in homes stay there longer.

  3. Recall the affected batches only

    Use production records to identify the output since the sensor was bypassed, about half the volume, and recall only those batches. This is proportionate if the records are reliable, and damaging if contamination is later found outside them.

  4. Recall and investigate in public

    Recall all affected products tonight, publish what is known and unknown, and commit to publishing the investigation into the bypassed sensor within 30 days. This is the most transparent response, but it exposes the maintenance failure and may affect liability.

Judge the options on

  • Speed of protecting customers
  • Accuracy of what is said and done
  • Responsibility that customers and supermarkets will attribute
  • Cost, including relationships with supermarkets
  • What the response does to prevent a recurrence
Read the guidance

Start from the attribution of responsibility. The cause is internal and preventable, a bypassed safety sensor, so customers, supermarkets and the regulator will attribute high responsibility, and situational crisis communication theory then favours accepting responsibility and acting decisively over minimizing. Waiting for test results protects against overreaction, but the risk lies in meals already in homes, and a delay that later looks like concealment would damage trust far more than an over-broad recall.

The batch-only recall is defensible if the production records are reliable, and its narrowness is its risk: if the records are wrong, Calder faces a second recall and loses its credibility. Recalling everything costs more but removes that risk, and the cost is small against the brands and supermarket contracts at stake.

The strongest answers also treat the crisis as a failure of execution: a safety control was bypassed during maintenance, which points to procedures, supervision and the pressure to keep lines running. Publishing the investigation is uncomfortable, but it is the response most likely to rebuild trust, and it forces the fix.

Section 08: Further reading

  • Henry Mintzberg and James A. Waters (1985). Of strategies, deliberate and emergent. Strategic Management Journal 6(3).

    The distinction between deliberate and emergent strategy, essential for thinking about execution.

  • John P. Kotter (1996). Leading Change. Harvard Business School Press.

    The eight-stage model in full, with the errors that defeat transformations. Read it critically.

  • Michael Beer and Nitin Nohria (2000). Cracking the code of change. Harvard Business Review 78(3).

    A short article contrasting economic and organizational theories of change.

  • Robert S. Kaplan and David P. Norton (1996). The Balanced Scorecard: Translating Strategy into Action. Harvard Business School Press.

    How the balanced scorecard was meant to work as a management system rather than a report.

  • Charles Perrow (1984). Normal Accidents: Living with High-Risk Technologies. Basic Books.

    Why some systems are prone to serious failure however carefully they are run.

  • Christine M. Pearson and Judith A. Clair (1998). Reframing crisis management. Academy of Management Review 23(1).

    A definition of organizational crisis and an integrated model of crisis management.

  • Dan Lovallo and Daniel Kahneman (2003). Delusions of success: how optimism undermines executives' decisions. Harvard Business Review 81(7).

    The planning fallacy in executive decisions, and the outside view as its remedy.

Every work cited here is listed in the platform's sources register. Real organizations appear only with facts from the sources shown; everything else is labelled illustrative or fictional.