NEW

Why Projects Fail: The Target Moves First, the Delivery Slips Later

Take your project management to the next level
November 2026
Why Projects Fail: The Target Moves First, the Delivery Slips Later

Why do projects fail? Most often because the target moves, not because the team stops working. When the Project Management Institute (PMI) surveyed practitioners worldwide on why projects fail, the three most cited answers were a change in the organization’s priorities, a change in project objectives, and inaccurate requirements. Each one means people delivered against something that stopped being right. Slipped dates and overspent budgets usually show up later, as the visible result.

The project failure rate depends on who is counting. PMI’s 2025 global study reports that 50% of projects succeed, 37% partly deliver and 13% fail outright. Bent Flyvbjerg’s database of more than 16,000 large projects shows that only 8.5% land on budget and on time. This guide explains why those numbers differ and ranks the causes of project failure. It then shows how the warning signs appear in your hours long before a status report turns red. It closes with the steps that prevent project failure and four public cases from four countries.

What is project failure?

Project failure means the project did not deliver enough value to justify what it cost. That definition is wider than missing a date. PMI now describes success as value that exceeds the effort and expense involved, as perceived by all key stakeholders, which is why a project can finish on time and still count as a failure.

It helps to separate three kinds of project failure, because each needs a different response:

Type What happened Where the cause usually sits
Delivery failure The work finished late or over budget Estimates, capacity, change control
Value failure The work finished, but few people use it or the benefit never arrives Objectives, requirements, user involvement
Abandonment The project stopped before it finished Priorities, sponsorship, funding

A project cancelled early because the business case changed may be a good decision. The real project failure in that case is the money spent before anyone asked the question, not the cancellation itself.

How often do projects fail?

Anywhere from 0.5% to 50% of projects count as successful, depending on the definition and the sample. The studies below measure different things, so their results should not be averaged.

Study Sample Result
PMI, 2025, global More than 5,800 project professionals, stakeholders and knowledge workers 50% succeed, 37% partly deliver, 13% fail outright
PMI, 2024, global More than 10,000 project professionals 48% successful, 40% mixed, 12% failure
Flyvbjerg, book published 2023, worldwide More than 16,000 big projects 8.5% on budget and on time; 0.5% also deliver the promised benefits
McKinsey and University of Oxford, 2012 More than 5,400 IT projects; figures for projects above $15 million 45% over budget and 7% over time on average, with 56% less value than predicted
Flyvbjerg and Budzier, 2011 IT projects 1 in 6 had an average cost overrun of 200% and a schedule overrun of almost 70%

Why do projects fail at such different rates across these studies? The gap comes from two choices. PMI’s figures rest on how people rate projects, so a late project that delivers value can still pass. Flyvbjerg measures hard targets on big projects, where a single miss counts. Neither is wrong. Outright failure is the minority result, and full delivery of cost, time and benefits is rare.

Note:
You will see the claim that 70% of projects fail. It is usually unsourced. Its closest academic relative concerns organizational change, not projects, and a 2011 review by Mark Hughes found no valid and reliable empirical evidence for it even there.

Why do projects fail? What the survey data ranks

The clearest ranked answer to why projects fail comes from PMI’s Pulse of the Profession 2018, a global survey in which respondents named the primary causes of failed projects. We grouped the 16 named causes into four families. The percentages are PMI’s, and the grouping is ours.

Family Causes (share of respondents)
The target moved or was never clear Change in organization’s priorities 39%; change in project objectives 37%; inaccurate requirements gathering 35%; inadequate vision or goal 29%
Control and people Poor communication 29%; opportunities and risks not defined 29%; poor change management 28%; inadequate sponsor support 26%; inexperienced project manager 22%; team member procrastination 13%
The numbers were wrong Inaccurate cost estimates 28%; inaccurate task time estimate 25%
Capacity ran out Resource dependency 26%; limited or taxed resources 21%; inadequate resource forecasting 18%; task dependency 12%

The causes of project failure add up to well over 100% because respondents could name several causes, which fits what most project managers see: failures stack. Three more limits apply. The survey dates from 2018, it records opinion rather than audited root causes, and it shows association, not proof of cause. Still, the shape is useful. The top of the list sits above the project team, with sponsors and portfolio owners. A team cannot fix a changed priority with better task management.

The same report found that 26% of organizations named inadequate sponsor support as a primary cause, and the share reached 41% among the weakest performers against 17% among the strongest. PMI’s 2025 press statement points in the same direction when it says projects fall short because they are misaligned, under resourced or measured in the wrong way.

9 reasons for project failure and what stops each one

These nine reasons for project failure show why projects fail in daily practice. They cover the survey families above and the planning mistakes that feed them. For a wider list of day to day problems, see our guide to project management challenges.

1. Priorities change above the project

A reorganization, a new executive or a budget cut quietly moves the project down the list. People get pulled away, decisions slow down, and nobody says the project has lost its place. What stops it: ask the sponsor to rank the project against other active work once a quarter, and treat a lower rank as a formal change to dates and budget.

2. Objectives are vague or keep moving

“Improve the customer portal” gives everyone permission to imagine a different result. When the goal shifts mid way, finished work has to be redone. What stops it: state one measurable outcome, who signs it off and what the project will not do. Writing down the exclusions matters as much as the scope, because unspoken limits turn into arguments near the deadline.

3. Requirements are gathered from the wrong people

Requirements written by managers on behalf of users tend to miss how the work really happens. Long documents also hide the gap, since nobody tests them against anything real. What stops it: put a rough version in front of real users early, then revise the requirements from what they do with it.

4. Estimates describe the best case

People forecast from their intentions and forget their own history. In a 1994 study of the planning fallacy, students expected to finish a thesis in 33.9 days on average and took 55.5. What stops it: base estimates on logged hours from similar past tasks, and ask the person doing the work rather than the person selling it. Our article on causes of cost overruns covers the budget side.

5. The deadline comes before the sizing

A date promised to a client or a board turns the schedule into a wish. The team then compresses testing and review, and peer pressure keeps anyone from saying it will not fit. What stops it: size the work first, then negotiate scope, people or the date with the numbers on the table. If the date cannot move, decide openly what gets cut.

6. Capacity is assumed, not checked

A person listed as full time on three projects is not available to any of them. Wrong role assignments make it worse, since a task given to someone without the skill costs more hours and creates rework for others. What stops it: compare assigned hours with real availability, including leave and support work, before committing to dates. Our guide to resource management explains how.

7. Risks are never said out loud

Teams often see the danger, yet nobody wants to sound disloyal at kickoff. Gary Klein’s project premortem gives people permission: imagine the project has already failed and write down why. Klein reports that this kind of prospective hindsight raises the ability to identify reasons for an outcome by 30%. What stops it: run a premortem of 30 to 60 minutes before the work starts and give each serious risk an owner.

8. The sponsor disappears after kickoff

An absent sponsor means slow decisions and no protection when priorities shift. What stops it: agree on a short regular slot for decisions, and bring the sponsor choices with numbers rather than status updates.

9. Changes get in without a price

Each small request sounds harmless, and together they consume the buffer. Poor change management drew 28% in PMI’s survey, and the same report lists avoiding scope creep among its top recommendations. What stops it: give every change an estimate in hours and money, and have someone approve it before work begins.

Early warning signs of project failure

Most articles on why projects fail stop at causes, but the practical question is when you can see trouble coming. A status report relies on someone choosing to raise a hand. Time and task data do not, which is why they often reveal project failure first. We explain the reporting side in why projects stay green until they turn red.

The thresholds below are our starting points, not an industry standard. Adjust them to your own history.

Signal Look for What it usually means
Actual vs estimated hours on finished tasks Finished tasks run more than 20% over for two weeks Estimates are wrong across the project, not on one task
Task status Tasks sit in progress with hours rising and no status change Hidden blockers or unclear requirements
Hours logged to the project Weekly hours fall 25% below the expected level People were pulled away, so priorities may have changed
New tasks after kickoff New estimated hours exceed 10% of the original total Scope is growing without a decision
Overtime The same people log overtime three weeks in a row Capacity was overstated, and burnout and rework may follow
Sponsor decisions A decision has been waiting for more than a week The sponsor has moved on

Our guide to tracking project progress covers methods in depth, and the project management KPIs article explains each formula.

Worked example: reading a project at week 4

Here is an illustrative project with invented numbers: 10 weeks, 1,000 estimated hours, a blended cost of $60 an hour. At the end of week 4 the schedule called for 400 hours of work. Finished tasks carry 300 estimated hours, and the team has logged 380.

Schedule performance: 300 / 400 = 0.75
Cost performance: 300 / 380 = 0.79
Forecast hours at this rate: 1,000 / 0.79 = 1,267 hours
Forecast overrun: 267 hours x $60 = $16,020
Forecast length: 10 weeks / 0.75 = 13.3 weeks

The team looks busy, and 38% of the hours are gone, so a percent complete update might say “on track”. The data says 30% of the work is done. If the remaining 700 estimated hours run exactly to estimate, the total is 1,080 hours and the overrun falls to $4,800. So the range at week 4 is $4,800 to $16,020, and the sponsor has six weeks left to act on it.

How far can you trust an early reading? Research on US defense contracts found the cumulative cost index stabilizing from about the 20% mark. A later study by Henderson and Zwikael, using projects from the UK, Israel and Australia, found that the rule cannot be generalized and that stability often came after the 80% mark. Treat an early index as a reason to investigate, not as a final forecast.

Compare estimated and actual hours on every task with actiTIME and spot a slipping project while there is still time to act

Try Free

Project tracking

Project failure examples from four countries

Public audits give the best evidence on project failure, because independent reviewers documented what went wrong. These four cases show why projects fail at national scale, and the same mechanisms show up in a team of ten. The family label on each is our reading of the findings. Product flops such as New Coke or the Edsel are market failures rather than delivery failures, so we left them out.

FBI Virtual Case File, United States

The FBI abandoned the system in 2005 after spending more than $170 million. Requirements filled an 800 page volume and still kept changing, with roughly 400 change requests in a single year, while the contract set no specific completion milestones. The Inspector General listed poorly defined requirements, unrealistic scheduling and a lack of management continuity among the reasons. Family: the target moved.

Phoenix pay system, Canada

The Auditor General of Canada called Phoenix “an incomprehensible failure of project management and oversight” in a 2018 audit, which found that executives could have given deputy ministers accurate information before launch and did not. A Senate committee reported that instead of $70 million in annual savings, the government would incur about $2.2 billion. Family: control and people, since warnings did not reach decision makers.

National Programme for IT, United Kingdom

England’s NHS programme started in 2002 with an initial budget of about £6.2 billion and was officially dismantled in 2011. The Public Accounts Committee reported a forecast cost of £9.8 billion. A 2017 review blamed weak end user engagement, no phased approach to change and an underestimate of the scale. Family: requirements set far from the people who would use the result.

Berlin Brandenburg Airport, Germany

The airport opened in 2020 after 14 years of construction and six delayed openings, with costs at 6 billion euros against an initial budget of 2 billion. Reuters lists a bankrupt planning company and faults in fire doors, cabling and sprinklers. Family: control and people.

How to prevent project failure

Flyvbjerg’s rule is “think slow, act fast”, as his book review in The Independent Review summarizes it. Knowing why projects fail only helps if it changes what you do first. The sequence below follows that logic. Most of the protection comes before delivery starts, and the rest comes from a short weekly routine.

  1. Write the success test first.

    State the outcome, the measure, the owner and the exclusions on one page. Our kickoff meeting guide shows how to get agreement on it.

  2. Estimate from similar past work.

    Pull actual hours from comparable tasks before anyone commits to a date, and record the range rather than one number.

  3. Run a premortem.

    Imagine the failure, collect the reasons and assign an owner to each one that matters.

  4. Check capacity against real availability.

    Include leave, support work and other projects, then fix the gap before the start date rather than after it.

  5. Agree on what stops the project.

    Name in advance the conditions that would trigger a review, such as a forecast overrun above an agreed share of the budget.

  6. Price every change.

    No new work starts without hours, cost and an approver.

  7. Review the signals each week.

    Spend 15 minutes on the table above, and take anything over its threshold to the sponsor as a decision.

  8. Run a lessons review while memory is fresh.

    Compare estimates with actual hours and feed the difference into the next estimate.

If the project is already off course, recovery needs different steps. Start with what to do when your project goes wrong.

Catching trouble early in actiTIME

actiTIME records work at the task level, under a customer and project structure you can rename, so the signals above come from the same hours people already log. Set task estimates and deadlines, then watch the Estimated vs. Actual Time Report for the first sign that estimates are off. Custom workflow statuses and the Kanban view show which tasks have stopped moving.

For money, time, cost and billing budgets can be set at task, project or customer level, and the Cost of Work and Profit/Loss reports show what an overrun means in currency, including overtime. A lessons review then has real numbers behind it, because last quarter’s actual hours are one report away.

Now we can better predict future project requirements

Our company needed a simple way of tracking time used on multiple projects – and actiTIME fit the need. Its interface is simple and easy to maintain. We use the application for time management, task estimation and also to communicate deadline information to our team members. Now having actiTIME we can better predict future project requirements!

The Online version starts from $5 per user per month, and a permanent free version covers teams of up to 3 users with a limited feature set. Budgets and task estimates need the paid version or the 30 day trial, which is full featured. Start a free trial on your next project and compare the first four weeks against your estimate.

FAQ

Is a late project a failed project?

Not by itself. A project that finishes late and still delivers more value than it cost counts as a success under PMI’s current definition. It becomes a failure when the delay destroys the value, for example when a product misses its market or a contract penalty wipes out the margin. Agree on which date matters, and why, before work starts.

Who is responsible for project failure?

The project manager is accountable for delivery, and the sponsor is accountable for the business case and the decisions that protect it. PMI’s survey puts inadequate sponsor support at 26% and an inexperienced project manager at 22%, so blaming one role rarely explains why projects fail. A useful review asks which decision went wrong and what information was missing when it was made.

When should you cancel a failing project?

Cancel when the cost to finish exceeds the value still available. Money already spent should not count, so compare only remaining cost with remaining benefit. In the worked example above, the question at week 4 is whether $16,020 more buys something the business still wants. Decide against criteria set before the project started, since that makes the choice easier to defend.

What are the 5 C’s of project management?

There is no single standard list. One version repeated online names complexity, criticality, compliance, culture and compassion, while a PMI paper uses a 5 C approach for something else entirely: knowledge management. Treat any such list as a memory aid rather than a framework.

Watch the target before you watch the team

So why do projects fail? The evidence keeps pointing the same way. The target moves. The numbers were optimistic. The capacity was imagined. The warnings sat in the data until it was too late. Fix the first three before delivery begins, then review the data every week. You will not avoid every project failure, but you will see it while it is still cheap.

Are you ready to drive your business growth with actiTIME?

Start Using actiTIME