Anatomy of a Failed IT Project
Darryl Carlton
AI Governance Thought Leader | Digital Transformation Expert | Bestselling Author in Cybersecurity & AI Governance | Dispute Resolution and Expert Witness
Listen to the Podcast HERE
Bram Stoker’s character, Professor Van Helsing in Dracula, said “we learn from failure, not from success!”.
This is a favourite quotation of mine, and with forty years of managing IT projects, I have spent the last ten years studying IT project failure. And I have developed some ‘minority opinions’.
My interest in why projects fail was sparked by a question attributed to the former Chief Information Officer at the Treasury Board of Canada, who raised his hand at a Standish Group presentation in 1995. The question has come to be known as Cobb’s Paradox:
‘We know why projects fail, we know how to prevent their failure – so why do they still fail?’
IT projects don’t just fail a little bit. It’s not an occasional thing which we can fix by saying, “Oh, just implement Agile or Prince, and it will be fine”.
Professor Bent Flyvbjerg, currently the world’s leading authority on project performance has highlighted in his research that not only do IT projects have a much higher failure rate than most other project types, the so-called “fat tail” in IT is incredibly bad, the average cost overrun on failed IT projects is 447% “meaning that many IT projects have even higher overruns than this”.
Big projects fail big. Where an IT project has a projected cost that exceeds USD$100 million the likelihood of successful completion is just 2%. And by successful we mean on-time, on-budget with required features. 51% of these large projects will be ‘challenged’ and 47% will fail . We really do not see these figures reversed until we look at projects of less than USD$3m, where 39% are considered successful, or less than USD$1m where the number is a respectable 70% successful.
In many cases, small agile projects of less than USD$1m in total project expenditures have created a somewhat false impression that agile is the silver bullet that will finally address IT project failure rates. The literature today is full of discussions about the challenges of scaling agile to mega projects.
1. Why Do Projects Fail?
Understanding why projects fail, and correcting for those failures, is critical. Even more as we enter the age of Artificial Intelligence where project failure may have very consequential outcomes.
There is a history of IT projects failing and taking companies down with them. But most IT projects simply fail, and the costs and consequences are somehow ‘absorbed’. No-one appears ever to be held to account, not the executives who had oversight or the vendors who got paid.
The Queensland Health Payroll project was conceived as a transformative initiative aimed at modernising payroll processing for Queensland Health’s extensive workforce. With over 80,000 employees reliant on accurate and timely payment, the stakes for this project were immense. Despite the ambitious vision and substantial resources allocated to the initiative, the project culminated in one of the most significant public sector IT failures in Australian history. Costs spiralled to over AUD$1 billion, more than ten times the original budget, and the project’s implementation was plagued by defects, delays, and operational disruptions that left employees unpaid and the government’s credibility severely damaged.
The Honourable Richard Chesterman QC headed up a Commission of Inquiry into the failure of this project, and in his final report he stated :
“Its failure, attended by enormous cost, damage to government and impact on workforce, may be the most spectacular example of all the unsuccessful attempts to impose a uniform solution on a highly complicated and individualised agency”
Sometimes, IT project failures have an impact beyond financial and cause serious harm to individuals. In Australia we had the Royal Commission into the Robodebt Scheme highlighting serious issues with algorithmic decision making. Harrowing stories of harm, including suicide of impacted individuals were raised.
In the UK what was once the largest IT project in Europe has become the UK’s most appalling IT project failure. Closing Statements have just been entered on the 16th and 17th of December 2024 to what has been a 25 year saga. The UK Post Office Horizon project , in which 236 subpostmasters were jailed, 900 convicted of fraud or theft, and at least four suicides.
In this report, I want to revisit and delve into the multifaceted causes of the failure presented by the Queensland Health Payroll project. I have examined the public documents and witness statements made available through the Chesterman Commission of Inquiry which was established by the Queensland Government. All witness statements were taken under Oath and were admitted into public record.
The conclusions are mine, and I have examined these documents with the intent of trying to understand (a) what were the drivers of this projects failure, and (b) who if anyone, could be seen to be accountable for the failure.
In examining these public documents, I have endeavoured to identify key systemic issues such as governance breakdowns, vendor mismanagement, or insufficient risk mitigation. It also offers forward-looking recommendations to prevent similar outcomes in future projects. These insights are intended to guide both public and private sector leaders in navigating the complexities of large-scale IT initiatives.
2. Background and Context
Scope, Objectives, and Budget
The Queensland Health Payroll project was envisioned as a solution to replace the ageing and unsupported LATTICE payroll system. The LATTICE system, nearing the end of its lifecycle, posed significant risks to payroll stability, especially given Queensland Health’s complex award structures and compliance requirements. The project aimed to deliver a modern, integrated payroll system capable of meeting the organisation’s unique needs while adhering to tight budgetary constraints and a fixed timeline. With over 80,000 employees reliant on accurate and timely payment, the stakes for this project were immense. Despite the ambitious vision and substantial resources allocated to the initiative, the project culminated in one of Australian history’s most significant public sector IT failures. Costs spiralled to over AUD$1.2 billion, against an original budget of AUD$98 million, and the project’s implementation was plagued by defects, delays, and operational disruptions that left employees unpaid and the government’s credibility severely damaged. The project was expected to go live by mid-2008, and it should be noted that this system is now operational, and as of 2024 still being rolled out across the State.
Timeline and Stakeholders
The project’s timeline was ambitious from the outset, reflecting both the urgency of replacing the LATTICE system and the government’s desire to demonstrate swift progress. Stakeholders included Queensland Health as the end user, CorpTech as the government’s technology service provider, and IBM as the prime contractor responsible for delivering the solution. The governance structure was further complicated by overlapping roles and responsibilities among these parties, which became a recurring source of confusion and conflict throughout the project’s lifecycle.
Organisational and Policy Environment
At the time, the Queensland government was pursuing the Shared Services Initiative, an overarching effort to centralise and streamline corporate functions across multiple agencies. While theoretically sound, this policy created significant challenges when applied to Queensland Health’s unique payroll needs. The initiative’s focus on uniformity often clashed with the complexity and specificity of Queensland Health’s requirements, creating a foundational misalignment that hindered the project’s success.
3. Analysis of Budget Overruns
Causes of Budget Escalation
The Queensland Health Payroll project experienced staggering cost overruns, escalating from an initial budget of AUD$98 million to over AUD$1.2 billion. These overruns can be attributed to several interconnected factors:
Governance Failures
The project’s governance framework was ill-defined and poorly executed, resulting in a lack of accountability and delayed decision-making. For instance, CorpTech, responsible for contract management, often failed to enforce penalties or escalate issues effectively, leaving Queensland Health unable to assert control over critical aspects of the project. As one witness noted in their sworn statement:
"I said in the review presentation that the project lacked a person who held the single point of delivery accountability. I said, ‘You haven’t got one. It’s missing. You’ve got a committee of people all doing a number of things, HR, finance, procurement, all away doing their thing, each of them reporting up to a gentleman called Geoff Waite, who fundamentally came from a Treasury finance background and seemed to be a really powerful guy out of Treasury; a very capable man, but reporting to him was done in basically financial form’ .
Scope Creep
Frequent changes to project scope, often driven by evolving requirements and stakeholder pressures, significantly inflated costs. These changes were poorly managed, with limited assessment of their impact on the budget or timeline. The introduction of new features and workarounds further compounded delays and expenses. It was stated, as just one example, that "… this activity identified two new concerns for the program. Firstly, the delivery horizon now crossed the end date for support for the version of SAP being used. This means an upgrade would be required mid program, resulting in a further year delay."
Vendor Mismanagement
It was reported throughout the Inquiry that IBM, the project’s prime contractor, struggled to deliver on its commitments. Issues such as insufficient staffing, high turnover among project managers, and inadequate adherence to project management methodologies undermined progress. It was observed, by the Commission for Inquiry, that "IBM’s inability to provide skilled resources and consistent leadership disrupted project continuity".
"The lack of project methodology has set the project up for a number of critical failures that have impacted the successful delivery of a fit for purpose product."
One witness observed that “IBM has an insufficient number of appropriately skilled staff” and that a significant cause of issues on the project was due to a “lack of adequate skill and resources provided by IBM”
Technical Integration Challenges
The integration of SAP and Workbrain, two complex systems with distinct functionalities, proved far more challenging than anticipated. Poor documentation and insufficient testing have been cited as having exacerbated these difficulties, leading to costly rework and delays.
4. Analysis of Delivery Failures
Technology and System Design Flaws
The project’s technical architecture failed to account for Queensland Health’s unique payroll requirements, resulting in a fundamentally unfit system. Key issues included inadequate configuration of award structures and unresolved defects at the time of the system’s go-live. "The system’s inability to handle basic payroll scenarios caused widespread payment errors, leaving employees unpaid or overpaid"
Vendor Performance
IBM’s performance was marred by poor execution and a lack of accountability. Frequent turnover among key personnel disrupted continuity and eroded trust between stakeholders. Moreover, IBM’s failure to adhere to agreed timelines and quality standards necessitated costly interventions.
Stakeholder Engagement Gaps
A significant communication gap between Queensland Health, CorpTech, and IBM hindered effective decision-making. Queensland Health, the primary end user, was often excluded from governance processes, leading to misaligned priorities and unresolved issues. "Stakeholder engagement was limited, resulting in critical disconnects between technical delivery and operational needs" .
Organisational Resistance
Resistance to change within Queensland Health further compounded the project’s challenges. Internal silos and conflicting priorities among departments created obstacles to adoption and implementation.
5. Root Cause Analysis
Applying the "5 Whys" methodology unveils a layered narrative of missteps and mismanagement that culminated in the Queensland Health Payroll project’s failure. Each ‘why’ exposes a deeper systemic issue that, left unresolved, cascaded into the next.
Why did the project fail? Operational failures were compounded by fragmented and inadequate governance frameworks. The governance structure lacked clear accountability, decision-making authority, and effective coordination between key stakeholders, including the Queensland Government, vendor IBM, and subcontractors. The absence of a cohesive framework meant that critical issues, such as identifying and addressing system defects, were not managed with the urgency or strategic oversight required for a project of this scale and importance. Furthermore, the lines of responsibility were blurred, leading to indecision and delays in implementing corrective measures.
The inquiry also highlighted systemic failures in risk management, change management, and stakeholder engagement. There was insufficient testing prior to deployment, with several unresolved issues either underestimated or ignored entirely. Communication breakdowns between project teams and end-users exacerbated the problem, as frontline staff and managers struggled to navigate a flawed system without adequate training or support. Overall, the inquiry painted a damning picture of poor governance, combined with technical and procedural inadequacies.
Why was the system defective? The Inquiry underscored the severe shortcomings in the requirements gathering and testing processes, which were pivotal to the system's failure. These processes, foundational to ensuring that the system met organisational needs and functioned effectively, were compromised by significant gaps in planning and execution. The reliance on incomplete documentation meant that the requirements were neither comprehensive nor reflective of the operational complexities of Queensland Health. Critical nuances, such as the intricate award structures and varied employee classifications, were either inadequately captured or entirely overlooked, leading to a system ill-equipped to handle its intended scope.
Compounding this issue was a rushed project timeline, driven by pressures to meet arbitrary deadlines rather than ensuring quality outcomes. This urgency truncated essential stages of development, including rigorous validation and quality assurance processes. The accelerated pace left little opportunity for iterative testing or meaningful stakeholder feedback, both of which are vital to identifying and addressing potential flaws prior to deployment. As a result, defects that could have been rectified during development persisted into the post-implementation phase.
A glaring example of this failure was the inadequacy of end-to-end testing. While testing should have simulated real-world scenarios and stress conditions to ensure system robustness, the inquiry revealed that testing protocols were superficial and incomplete. According to the report, "Deficiencies in end-to-end testing allowed critical flaws to persist until post-implementation” . These deficiencies allowed technical issues to go unnoticed and failed to account for user interactions and operational contingencies, leaving the system unprepared for actual use.
Why were requirements poorly defined? A lack of domain expertise among key stakeholders, combined with ineffective collaboration, hindered the project’s ability to capture the nuanced requirements of Queensland Health’s operating environment. Without deep knowledge of healthcare workflows and their inherent complexities, project leaders and technical teams struggled to translate real-world needs into the system’s specifications. This challenge was further amplified by communication gaps that emerged when clinical and non-clinical parties did not have a unified platform for reviewing ongoing development. The result was a series of design decisions that did not fully reflect the day-to-day tasks and strategic objectives of the organisation.
An important observation, highlighted by several witness statements, is that the absence of regular stakeholder engagement led to persistent gaps between technical delivery and user expectations. Overlooking continuous feedback loops meant that users’ insights were never fully incorporated into design iterations, ultimately creating a disconnect between what was built and what the health system needed.
Why was expertise lacking? The Inquiry revealed how excessive outsourcing of the payroll project created a critical knowledge deficit within the organisation. By delegating substantial responsibilities to external vendors, primarily IBM, without developing internal expertise, Queensland Health found itself unable to effectively oversee the project or challenge vendor decisions.
领英推荐
This dependency manifested in mounting costs and operational vulnerabilities. Critical decision-making defaulted to contractors with minimal organisational oversight, while frequent personnel changes among vendor teams disrupted project continuity and institutional memory. The inquiry noted that each transition required extensive onboarding periods, leading to delays and fragmented project management.
Most critically, Queensland Health's failure to build internal capabilities alongside external partnerships left them ill-equipped to manage the system post-deployment. The absence of in-house expertise not only weakened their negotiating position with vendors but also compromised their ability to maintain and adapt the system independently. This case demonstrates how the balance between external support and internal capacity-building can determine a project's long-term viability.
Why was oversight absent? The Inquiry revealed how inadequate governance and accountability structures fundamentally undermined the payroll project. The inquiry exposed a fragmented oversight framework that failed to establish clear lines of responsibility among Queensland Health, CorpTech, and external vendors.
The governance failures were most evident in CorpTech's role as the oversight agency. Despite being responsible for vendor management and contract compliance, CorpTech lacked the authority and mechanisms to enforce standards or hold contractors accountable. This created an environment where technical defects and project issues went unaddressed while vendors operated with minimal oversight.
Most significantly, the absence of unified leadership and clear decision-making protocols left the project vulnerable to bureaucratic inertia. When critical issues emerged during the testing phases, the fragmented governance structure prevented timely interventions and were often mischaracterised as ‘personality conflicts’. This systemic failure in oversight not only compromised the project's execution but also eroded stakeholder trust, highlighting how essential robust governance frameworks are to public sector initiatives. In earlier writings, I referred to this as Situational Incompetence.
Through this methodical deconstruction, it becomes evident that the project’s failure was not the result of a singular misstep but a confluence of interconnected failures at every level.
6. Lessons Learned
Governance and Oversight
Robust governance structures are critical to the success of complex IT projects. Clear delineation of roles and responsibilities, combined with independent project reviews, could have mitigated many of the challenges faced by the Queensland Health Payroll project.
Risk Management
Comprehensive risk assessments and phased rollouts are essential for identifying and mitigating potential issues before they escalate. For example, independent testing phases could have highlighted integration challenges earlier.
Vendor Selection and Management
Future projects should establish stricter criteria for vendor selection and implement performance-based contracts to ensure accountability.
7. Recommendations
The failures of the Queensland Health payroll project, as documented extensively in the Chesterman Commission of Inquiry, offer a cautionary tale of governance, oversight, and strategic execution deficiencies. These lessons are critical not only for Queensland Health but for all public sector agencies undertaking large-scale IT projects. To prevent similar failures in the future, a multi-phased approach comprising immediate, medium-term, and long-term actions is necessary. Below is an expanded framework that aligns with best practices and the insights derived from the inquiry.
Immediate Actions
1. Conduct Independent Audits
All ongoing and planned IT projects across government agencies should undergo independent audits to identify potential risks, governance gaps, and misaligned objectives. Drawing on lessons from the Queensland Health payroll project, audits should prioritise requirements gathering, vendor contract compliance, and end-to-end testing protocols. Neutral parties must conduct these audits to ensure an unbiased assessment of project risks and vulnerabilities. The objective is to provide actionable recommendations before further resources are committed.
2. Establish Clear Governance Frameworks
Fragmentation in governance, as evidenced in the payroll project, allows systemic issues to persist unchecked. Governance frameworks for IT projects should define roles, responsibilities, and accountability mechanisms for all stakeholders, including government agencies, contractors, and consultants. For example, CorpTech’s inability to enforce vendor contract terms during the payroll project demonstrates the need for a governance structure that includes the authority to make binding decisions, enforce penalties, and manage cross-agency coordination effectively.
Medium-Term Actions
3. Develop a Robust Vendor Management Framework
The Queensland Health payroll project revealed critical flaws in vendor relationships, including a lack of accountability and poor contract enforcement. To address this, government agencies must develop comprehensive vendor management frameworks.
These frameworks should incorporate:
Ensuring that vendors are held accountable for deliverables and outcomes will mitigate risks and prevent reliance on goodwill rather than enforceable agreements.
4. Enhance Training Programs for Key Stakeholders
The project suffered from inadequate internal capabilities to manage complex IT implementations, resulting in an overreliance on external vendors. To rectify this, tailored training programs should be developed for project managers, government leaders, and technical stakeholders. These programs should include:
By equipping internal teams with the skills necessary to manage and oversee IT projects, agencies can reduce reliance on external contractors and improve outcomes.
Long-Term Actions
5. Institutionalise Lessons Learned Through Policy Reforms
The failures identified in the Chesterman Commission of Inquiry should serve as a foundation for comprehensive reforms in project governance and risk management policies. Governments must mandate that all large-scale IT projects adhere to a standardised set of guidelines that include:
Embedding these lessons into policy frameworks ensures that they inform future projects, avoiding repetition of past mistakes.
6. Invest in Building Internal IT Capabilities
Overdependence on external vendors was a critical flaw in the payroll project, leading to cost overruns and operational risks. To address this, long-term investments must focus on building robust in-house IT capabilities. Governments should prioritise recruiting and retaining skilled IT professionals, particularly in areas such as:
By developing internal expertise, public agencies can take greater ownership of their IT projects and reduce their vulnerability to vendor shortcomings.
Conclusion & Ongoing Research
The Queensland Health payroll project is a stark reminder of the consequences of fragmented governance, poor planning, and insufficient oversight. The immediate, medium-term, and long-term actions outlined above represent a structured approach to addressing these failures and improving the execution of IT projects in the public sector. By implementing these reforms, governments can strengthen project outcomes, enhance accountability, and rebuild trust in their ability to deliver critical infrastructure for public benefit.
My ongoing research examines how courts assign accountability in IT project failures, analysing key legal precedents and frameworks.
Contract breaches are emerging as the primary basis for liability, particularly regarding unmet deliverables and timeline violations. Courts also heavily weigh misrepresentation claims and deceptive conduct during project execution. The analysis reveals that establishing causation and foreseeability is crucial for plaintiffs, while courts frequently consider implied terms and professional duty of care in their determinations.
Notably, defendants often invoke performance defence strategies, citing external factors beyond their control. The research also highlights how courts evaluate claims of unjust enrichment, particularly in cases of partial fulfilment. Project governance failures and contractual ambiguities significantly influence judicial decisions on accountability and remediation measures.
The legal framework analysis aims to inform better risk management strategies and accountability structures in IT project governance. Hopefully, it will lead to better-prepared projects and more accountable management, reducing the number and consequence of IT project failures.
AI GOVERNANCE PODCAST
PODBEAN: https://doctordarryl.podbean.com
GET MY BOOKS HERE
Governing AI in Australia - https://amzn.asia/d/i5MFgwN
AI Governance - https://amzn.asia/d/07DeET2v
Cybersecurity Governance - https://amzn.asia/d/0edKXaav
AI Digest Volume 1 - https://amzn.asia/d/0ekqTUH0
AI Digest Volume 2 - https://amzn.asia/d/06syVuaJ
#EUAI #AIRegulation #TechPolicy #DigitalTransformation #AIGovernance #RegulatoryCompliance #AI #ArtificialIntelligence #AIGovernance #AIRegulation #AIRegulations #AIPolicy #AIEducation #EdTech #HigherEdAI #ResponsibleAI