galapagos.

Research base

Six cases, 238 evidence observations

6external cases
238evidence observations
17failure mechanisms
2pre-registered success tests

Four troubled or failed transformations and two successful transformations, selected in advance to test whether the model can distinguish between them.

The six-case matrix

CaseOutcomeChange LoadCouplingExecution CapacityEvidence and response
CQC Troubled / failed Very High Very High, poorly coordinated Insufficient and deteriorating Warnings and feedback did not change decisions fast enough.
Birmingham Failed; recovery open Very High Very High Insufficient at original go-live Late validation and weak response; recovery improves evidence-to-action.
ESN Troubled; recovery open Extreme Extreme Insufficient historically Repeated warnings, resets and supplier dependencies; recovery materially strengthens controls.
Target Canada Failed / exited Extreme Very High Insufficient for launch pace Learning and remediation arrived after a large live footprint had amplified problems.
Netflix pre-registered test Successful Very High High inherent, actively contained High and growing Rapid real-scale validation and response; long absorption horizon.
Capital One pre-registered test Successful overall Very High High inherent, actively simplified Generally high, uneven by domain Strong application learning; slower risk / control remediation in one critical domain.

Ratings show the strength of evidence that each condition was materially present. They are not probabilities of failure or causal weights. The wording is reproduced from the research paper, including "Extreme", which the paper uses for the two most heavily loaded cases.

The discriminating pattern. Scope Breadth and Scope Depth were high in both the failed and the successful transformations. The stronger differentiation appeared in concurrency and dependency overload, organisational readiness, validation, response to evidence, integration and the wider sufficiency of Execution Capacity. Netflix and Capital One both carried very high breadth and depth while succeeding overall, which is why size alone does not separate the two groups.

What each case was

CQC

A national regulatory transformation combining a new assessment framework, organisational redesign and technology change.

Birmingham Oracle

A council-wide ERP and operating-model transformation spanning finance, payroll, procurement, HR, income and controls.

ESN

A national mission-critical communications programme serving more than 300,000 frontline emergency-service users.

Target Canada

A rapid national retail launch that built stores, systems, distribution, workforce and supply-chain capability concurrently.

Netflix

A seven-year rebuild of virtually the entire streaming technology environment while the company scaled globally.

Capital One

An eight-year regulated-enterprise transformation covering cloud, applications, data, software delivery and technology capability.

Evidence boundary

What this research establishes

The cases provide cross-sector evidence that the model can distinguish meaningfully between different transformation outcomes.

What it does not establish

Six purposively selected cases do not provide statistical validation, universal effect sizes or a calibrated probability of failure. The paper states that the cases can sharpen and challenge the theory but cannot establish population-level validity.

Sources and references (29)

Primary sources include independent reviews, public interest and value-for-money reports, National Audit Office and parliamentary committee material, regulatory enforcement actions and first-party engineering publications. The research workbook preserves observation-level provenance, evidence stance, confidence and alternative explanations for all 238 observations.

  1. Harkin, B. (2024). Evolving from Digital Transformation to Digital Acceleration Using The Galapagos Framework. CRC Press / Taylor & Francis. DOI: 10.1201/9781003404217
  2. Harkin, B. (2026). Galapagos Baseline & Trajectory Assessment Protocol v1.0. Working research protocol.
  3. Dash, P. (2024). Review into the operational effectiveness of the Care Quality Commission: full report. UK Government.
  4. Richards review (2024). Review of CQC's single assessment framework and its implementation. Care Quality Commission.
  5. Gill review (2025). Independent IT review. Care Quality Commission.
  6. Grant Thornton (2025). Public Interest Report / Value for Money report on Birmingham City Council Oracle implementation.
  7. Birmingham City Council (2023). Plan to stabilise and optimise council IT systems.
  8. Birmingham City Council Commissioners (2025). Third report. UK Government.
  9. National Audit Office (2016). Upgrading emergency service communications: the Emergency Services Network.
  10. UK Parliament, Public Accounts Committee (2019). Emergency Services Network: further progress review.
  11. National Audit Office (2023). Progress with delivering the Emergency Services Network.
  12. Home Office (2026). ESMCP Accounting Officer Memorandum, March 2026.
  13. Target Corporation (2013). FY2012 Form 10-K. U.S. Securities and Exchange Commission.
  14. Target Corporation (2014). FY2013 Form 10-K. U.S. Securities and Exchange Commission.
  15. Target Corporation (2014). Response to SEC comment letter.
  16. Target Corporation (2015). Q&A: Brian Cornell on Target's exit from Canada.
  17. Alvarez & Marsal (2015). Proposed Monitor pre-filing report, Target Canada Co.
  18. Netflix (2016). Completing the Netflix Cloud Migration.
  19. Netflix TechBlog (2010). 5 Lessons We've Learned Using AWS.
  20. Netflix TechBlog (2012). Fault Tolerance in a High Volume, Distributed System.
  21. Netflix TechBlog (2015). Chaos Engineering Upgraded.
  22. Netflix, Inc. (2017). FY2016 Form 10-K. U.S. Securities and Exchange Commission.
  23. Capital One (2018). Digital transformation begins with audacious goals.
  24. Capital One (2017). How we migrated a customer-facing mobile app to AWS.
  25. Capital One Financial Corporation (2019). FY2018 Form 10-K. U.S. Securities and Exchange Commission.
  26. Office of the Comptroller of the Currency (2020). OCC assesses $80 million civil money penalty against Capital One.
  27. Federal Reserve Board (2020). Enforcement action against Capital One Financial Corporation.
  28. Capital One (2020). Using machine learning to build resilient mobile apps.
  29. Galapagos Transformation Failure Model Research Workbook v1.0 (2026). Unpublished working research artefact containing 238 evidence observations across Cases 001-006.

Claims and limitations

Six cases sharpen and challenge the theory; they cannot establish population-level validity. Ratings measure strength of evidence that a mechanism or condition was materially present, not probabilities. The conceptual relationship expresses interaction, not calibration. Material contributors outside the model are recorded as Not Explained by Current Model rather than forced into a hypothesis.