Anyone who wants to know why IT projects fail will first come across percentages and rarely the context behind them. The studies only appear to contradict each other, because each defines failure differently. More important than the rate is the underlying cause, and that turns out to be strikingly consistent across all surveys.
Depending on the definition, 17 to 19 percent of IT projects fail completely, and about half encounter significant problems with budget, schedule, or scope. The most common causes are unclear requirements and a growing project scope, cited by 82 percent of the companies surveyed. Agile methods make this issue apparent earlier, but they don't solve it.
How Often IT Projects Fail: The Numbers
The Standish CHAOS Report 2024 classifies 31 percent of the projects studied as successful, 50 percent as completed with problems, and 19 percent as failed (Standish Group). The PMI study for Germany from the same year reports a lower failure rate of 17 percent. However, only 68 percent of projects there actually achieve their business objectives in the end (PMI data via Haufe, 2024).
| Survey | Successful | With problems | Failed | What Is Considered a Failure |
|---|---|---|---|---|
| Standish CHAOS Report 2024 | 31 percent | 50 percent | 19 percent | Canceled before shipment |
| PMI Germany 2024 | 68 percent achieve their business goals | not reported | 17 percent | Business Goals Ultimately Not Met |
This difference is not a contradiction. It highlights a problem of definition. Standish counts projects as failures if they were terminated before anything was ever delivered. PMI, on the other hand, measures whether the original business objectives were ultimately achieved. That is a significantly higher standard. A project can be delivered on time and within budget and still be considered a failure if it missed the mark in terms of actual needs.
What overrun means in practice
Average budget overrun for large IT projects in 2024. The delay is 46 percent, and the delivered software provides, on average, 39 percent less value than planned.
McKinsey, 2024
This pattern is even more pronounced in SAP S/4HANA implementations. Sixty percent of projects exceed both their budget and their schedule, and 30 percent take longer than planned (Horváth study via WirtschaftsWoche, 2025).
S/4HANA projects are complex migrations with many dependencies on legacy systems. Therefore, this figure cannot be applied directly to every software project. However, it does illustrate how significantly project scope and system complexity can reduce the chances of success, even for providers with an established process model and a large partner network.
For companies in industry and retail, the S/4HANA figure is relevant because ERP migrations have been among their largest IT projects for years. A production plant that moves warehousing, manufacturing control and financial accounting onto a new system at the same time has a different risk profile than a pure software house building a single application. The dependencies between modules, interfaces to machine controls, and custom solutions that have grown up over time all add up. Anyone launching such a project should treat the 60 percent figure as a realistic baseline assumption for their own planning.
Why IT Projects Fail in Practice
Behind the percentages, there is usually one cause that occurs more often than any other.
of the companies surveyed name unclear requirements and growing project scope as the most common challenge over the course of a project.
SDZeCOM, 2024
This is consistent with an observation made in many ongoing projects. At the outset, there is no solid definition of what the system is supposed to do. Every subsequent requirement is then treated as an exception, until all those exceptions have turned into a second, unplanned project. To learn how to stop this from happening, see the article on Scope Creep, and the groundwork for that is laid by a clean requirements specification.
Despite ten years of agile methods, the overall success rate has changed little. It has stagnated at a similar level since 2017 (PM World Journal, 01/2026). This is uncomfortable for anyone who believed that sprints and retrospectives would solve the underlying problem. Agile processes make problems visible earlier. They are no substitute for a clear definition of the project scope and leadership capable of making decisions.
The Limitations of the CHAOS Methodology
The CHAOS Report has been the subject of criticism for years. Standish does not fully disclose the exact sample size or survey methodology, and the categories “successful,” “challenged,” and “failed” have been adjusted several times over the years. Anyone citing the 19 percent figure should be aware that it is only partially comparable to earlier editions of the CHAOS Report from the 1990s and 2000s, in which the failure rate was significantly higher. This does not invalidate the current figure. It merely limits what can be inferred from it in one’s own specific case.
What These Numbers Mean for Your Project
For management and IT leads in industry and retail, the practical question is less which study is right. What matters more is when to take your own warning signs seriously. One missed milestone is not yet an alarm. Two missed milestones in a row, plus a requirements list that grows every week, is a pattern the cited studies describe precisely. Anyone who waits until the budget has already been exceeded by 75 percent has missed the best moment for a course correction. Details on root cause analysis during an ongoing overrun are in the article on budget overruns in IT projects.
Did you hit two milestones?Then now is the best time to make a course correction. We'll assess your project during a consultation, with no pressure to buy.
Frequently Asked Questions
How many IT projects actually fail?
Depending on the definition, between 17 and 19 percent fail completely, and a further 50 percent or so run with significant problems in budget, schedule or scope (Standish Group 2024, PMI data via Haufe 2024). The exact figure depends on whether an abandoned project, or one that was delivered but missed its objectives, counts as a failure.
How much does a failed IT project cost?
The project costs alone are usually only part of the bill. McKinsey puts the average budget overrun in 2024 at 75 percent of the original budget, on top of which come the benefits lost through delayed or undelivered software. For larger undertakings such as SAP S/4HANA implementations, the overrun rate stands at 60 percent of projects (Horváth via WirtschaftsWoche, 2025).
When should you seek outside help?
As soon as two consecutive milestones are missed or the list of requirements expands significantly beyond the original scope, it’s worth seeking an external assessment. At this point, the causes are usually still clearly identifiable, and the options for action are broader than they will be three months later. How a Interim CTO We describe exactly how we proceed during such a phase on our Services page.
The Next Step
torck sees these patterns regularly in its own IT projects for industry and retail. At the first signs of trouble we take on technical leadership and implementation from a single source, backed by our own development teams in Maxhütte-Haidhof, Vienna and Rabat. Arrange a no-obligation initial consultation, to categorize your project.
German 
English 
French