CoChat

Why did your project fail?

Open access
Cerpa, Narciso, Verner, June M. · 2009 · Communications of the ACM
Cited by 220

The DOI matches its Crossref record.

Abstract

Introduction We have been developing software since the 1960s but still have not learned enough to ensure that our software development projects are successful. Boehm 2 suggested that realistic schedule and budgets together with a continuing steam of requirements changes are high risk factors. The Standish Group in 1994 noted that approximately 31% of corporate software development projects were cancelled before completion and 53% were challenged and cost 180% above their original estimate. 13 Glass discussed 16 project disasters. 5 He found that the failed projects he reviewed were mostly huge and that the failure factors were not just management factors but also included technical factors. Linberg in 1999 found that 20% of software projects failed, and that 46% experienced cost and schedule overruns or significantly reduced functionality. 8 Later, Glass revisited failed projects and found that poor estimation was high on his list of failure factors. 6 In 2007 the Standish Group reported that 35% of software projects started in 2006 were successful compared with only 16% in the corresponding 1994 report; however, the 2007 CHAOS report still identifies 46% (53% in 1994) of software projects as challenged (having cost or time overruns or not fully meeting user's requirements) and 19% (31% in 1994) as outright failures. 12 The validity of the Standish Group findings has been questioned as not consistent with cost overrun results of other surveys. 7 Jørgensen and Moløkken-Østvold suggested that there are serious problems with the way the Standish Group conducted their research and that the findings were biased toward reports of failure because a random sample of top IT executives was asked to share failure stories when mailed confidential surveys. Recently Charette 4 commented that "billions of dollars are wasted each year on failed software projects" and that "we have a dismal history of projects that have gone awry." 4 Charette suggests that from 5%-15% of projects will be abandoned before or shortly after delivery as hopelessly inadequate. 4 Other recent studies, suggest various failure rates for software development projects up to 85%. 7 Stories of software failure capture public attention and in general there is a perception that software quality is not improving but getting worse. Developing software systems is an expensive, and often a difficult process. Although there are many guidelines for successful software development, 9,11 few project post-mortems are conducted, and little understanding is gained from the results of past projects. The project manager (PM) and the development team must deal with many pressures from project stakeholders (for example, upper level management, marketing, accounting, customers, and users) during the software development process. These pressures impact both the cost and the quality of the software produced. There are generally more than one or two reasons for a software project to fail, and it usually is a combination of technical, project management and business decisions. Many software project failure factors have been described in the literature. 1-13 However, most organizations do not see preventing failure as an urgent matter. It is not understood why this attitude persists. 4 Because most of the literature is based on a handful of failed project case studies, in our research we analyze 70 failed software projects to determine those practices that affected project outcome. We are interested in providing, from the perspective of software practitioners, quantitative evidence targeting those aspects of the development process that contribute to project failure. Essentially, we are interested in updating the results of prior studies and testing the validity of previously reported anecdotal evidence. To date, no one has taken a set of project data and teased it out to identify, for a whole group of such projects, the most common failure factors. We are interested in everyday projects that are not high profile enough to be reported in the literature. Our work builds on that previously reported by Boehm, 2 Charette, 4 Glass, 5,6 Jørgensen and Moløkken, 7 Linberg, 8 and the Standish Group, 12,13 among others.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1145/1610252.1610286

Cites

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cited by

See all 220

Metadata

Authors
Cerpa, N., & Verner, J. M.
Venue
Communications of the ACM
Published
2009-12
Volume
52
Issue
12
Pages
130-134
ISSN
  • 0001-0782 (ISSN-L)
  • 1557-7317
OpenAlex
W2063065565
Cited by
220
References
0
FWCI
35.31
Type
journal article
Language
en
Open access
green

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Why did your project fail? https://app.cochat.ai/papers/10.1145/1610252.1610286
The paper
Cerpa, N., & Verner, J. M. (2009). Why did your project fail? Communications of the ACM. https://doi.org/10.1145/1610252.1610286

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.