Why IT Projects Fail and How Governance and a PMO Help
Almost every experienced project manager has worked on a project that failed or nearly failed. Industry studies have reported for decades that a large share of IT projects exceed their budget or schedule, or deliver less than promised, although the exact percentages vary by study and by how success is defined. One well-known McKinsey and University of Oxford study of large IT projects, published in 2012, reported that on average they ran far over budget and delivered less value than predicted, and that a notable minority went so badly wrong that they threatened the existence of the company. The encouraging news is that the causes are well understood, and most are preventable.
What Does “Failure” Mean?
Failure comes in degrees. A project can be an outright failure (cancelled or never used), a partial failure (delivered late, over budget or with reduced scope) or a technical success but business failure (works as specified but delivers no value because the specification was wrong or users do not adopt it). Equally, a project can overrun yet prove very worthwhile. This is why success criteria should include business benefits, not only schedule and budget; see Project Charter and Project Initiation.
The Most Common Causes
| Cause | What it looks like | Prevention |
|---|---|---|
| Unclear or changing requirements | Constant rework; “that is not what we meant” at UAT | Strong requirements practice, prototypes, change control (see Requirements Gathering) |
| Weak sponsorship | Decisions delayed, priorities unclear, resources withdrawn | Named, engaged executive sponsor with real authority |
| Unrealistic estimates and deadlines | Schedule set before scope understood; permanent crunch | Range-based estimates, contingency, honest negotiation |
| Poor stakeholder engagement | Users surprised at go-live; hidden resistance | Stakeholder analysis and regular engagement |
| Inadequate risk management | Predictable problems treated as surprises | Active risk register with owners and responses |
| Scope creep | Many small additions, no matching change in time or cost | Baseline scope and a change control process |
| Insufficient or wrong resources | Missing skills; key people overloaded or shared | Realistic resource planning, skills development |
| Communication breakdown | Teams out of sync; bad news hidden | Communication plan, honest status reporting |
| Technology and integration problems | Unproven tools; underestimated interfaces | Early prototypes, architecture reviews, integration testing |
| Weak vendor management | Late or poor deliverables; contract disputes | Clear contracts and active oversight |
| Poor change management and adoption | System works but users avoid it | Training, communication, user involvement |
| No benefits focus | Project finishes but value never realised | Business case, benefit owners, post-implementation review |
Early Warning Signs
Projects rarely fail suddenly; they drift. Watch for these signals:
- Milestones move repeatedly, or status stays green while deliverables slip.
- The team is working long hours with no end in sight, and morale falls.
- Defects keep growing late in testing.
- Key stakeholders stop attending meetings or reviews.
- Requirements and scope keep changing without formal decisions.
- Risks are not discussed, or the risk register is empty.
- The sponsor is hard to reach or decisions take weeks.
- Cost performance index or schedule performance index trends below 0.9 (see Earned Value Management).
- People avoid giving honest estimates or bad news.
Project Governance
Governance is the framework of authority, decision rights, policies and oversight that guides and controls a project. It answers: who decides what, how are decisions escalated, how is performance reviewed and who is accountable for outcomes? Typical governance elements include:
- Project sponsor: owns the business case, secures funding and resolves high-level issues.
- Steering committee: senior stakeholders who guide direction, approve major changes and review progress at set intervals.
- Project board roles: in PRINCE2, an executive, senior user and senior supplier share accountability.
- Stage gates: formal checkpoints with go, hold or stop decisions; see The IT Project Life Cycle.
- Change control board: decides on baseline changes; see Scope Creep and Change Control.
- Assurance and audit: independent reviews of progress, quality, security and compliance.
- Standards and templates: consistent methods across projects.
Good governance is proportionate. Heavy approval layers on a small project add delay and frustration; no oversight on a large one adds risk. Governance exists to support delivery, not to replace thinking.
Running an Effective Steering Committee
- Keep it small and senior, including people who can make decisions and release resources.
- Meet on a fixed cadence, typically monthly, with extra sessions for critical decisions.
- Send a concise pack in advance: status, key risks, decisions required and forecast.
- Bring decisions, not just updates, with options and a recommendation.
- Record decisions and actions and follow them up.
- Be candid. A steering committee that only hears good news cannot steer.
What Is a PMO?
A Project Management Office (PMO) is a group that defines and maintains project management standards and supports the organisation’s projects. PMI describes three broad types, from least to most controlling:
| PMO type | Role | Example activities |
|---|---|---|
| Supportive | Consulting; provides templates, best practices and training | Maintains a template library; coaches new project managers |
| Controlling | Requires compliance with frameworks or standards | Audits projects; mandates reporting formats and gate reviews |
| Directive | Directly manages projects | Assigns and manages project managers; owns outcomes |
PMOs typically provide standard methods and templates, portfolio-level reporting, resource and capacity visibility, training and mentoring, lessons-learned repositories, tool administration, benefit tracking and quality assurance. A good PMO helps teams deliver; a poor one is seen as a bureaucratic tax. The test is whether project managers ask for its help or avoid it. Many organisations also have an Agile or “value delivery” office with a similar supporting role for Agile teams.
Portfolio View
At portfolio level, governance asks whether the organisation is working on the right projects. It balances demand against capacity, stops low-value or failing work, and aligns investment with strategy. Starting too many projects at once is itself a cause of failure, since shared people become bottlenecks. See Business Case and Project Selection.
Recovering a Troubled Project
When a project is in trouble, act decisively:
- Stop and assess honestly: where are we on scope, schedule, cost and quality, and is the business case still valid?
- Engage the sponsor and agree on whether to continue, reshape or stop.
- Re-baseline realistically with a fresh bottom-up estimate rather than a hopeful extension.
- Reduce scope to the essentials and deliver the most valuable parts first.
- Fix the root causes: unclear requirements, wrong skills, missing decisions.
- Strengthen visibility: shorter reporting cycles and more frequent demos.
- Rebuild trust: communicate openly and keep commitments.
Beware of the sunk cost fallacy: past spending is gone regardless of what you decide. The right question is whether the remaining cost is justified by the remaining benefit.
Common Mistakes
- Treating governance as paperwork rather than decision support.
- Letting a project continue because stopping feels like failure.
- Reporting green until the very end.
- Over-engineering PMO processes until teams work around them.
- Ignoring lessons from previous failures.
- IT projects usually fail from unclear requirements, weak sponsorship, unrealistic plans, poor communication and neglected people issues, not purely from technology.
- Watch for early warning signs: repeated slippage, falling morale, silent stakeholders and trending-down performance indices.
- Governance defines decision rights, escalation paths and stage gates; keep it proportionate.
- A PMO should support delivery; stopping a failing project can be the right decision.