Return to search

Theorizing outliers : explaining variation in IT project performance

IT projects are temporary organizations of strategic importance. Companies invest large amounts of money, time, and resources into business-embedded IT projects in order to change and gain a competitive advantage. Extreme cases of failures were previously only analyzed as case studies, e.g., Denver Airport, London Stock Exchange Taurus, London Ambulance Service. The research poses an important question: What is the risk of these outliers, that is markedly deviant observations of IT project performance? What causes outliers in IT project performance? Only very few studies problematized the frequency of outliers directly. Reported numbers range from 33% to as low as 0.2%. The variation has been explained through biases in planning processes of organizations and as artefact of data collection. An alternative explanation is that the true nature of IT projects contains more variation than commonly assumed. A rich body of organizational, project management, and IT project management literature offers antecedents of outliers. The extant literature falls broadly into three schools of thought: (1) system-centric, (2) event-centric, and (3) process-centric theories of why outliers occurred. System-centric explanations focus on the question of system design, based on theories of normal accidents and high reliability organizations. Event-centric explanations focus on how organizations respond to rare events that impact the organization, based on theories of crisis management, management of organizational turbulence, and strategic surprises. Process-centric explanations focus on the role of managing uncertainty and risk over time, based on theories of man-made disasters, escalation of commitment to a failing course of action, and the normalization of deviance. The study is based on the archival research of 4,307 IT projects from 190 organizations. The findings show that the tail of the cost, schedule, and effort performance distributions is best fitted by a power law, with overwhelming goodness of fit. Moreover, the findings show that system-centric explanations and process-centric theories offer explanations for the thickness of the tail and the odds of an outlier occurring. In particular five variables were associated with outliers: estimated cost and duration, perceived uniqueness of the project, the qualification and motivation of the project team, and the effectiveness of monitoring and controlling. The results show that outliers are not chance events; they follow patterns that are describable. The study showed how design factors, that are often conceptualized as system complexities, and execution factors, that are often conceptualized as the effectiveness of project processes, explain project outliers. Lastly, the thesis draws implications for research and practice.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:627834
Date January 2014
CreatorsBudzier, Alexander
ContributorsFlyvbjerg, Bent
PublisherUniversity of Oxford
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://ora.ox.ac.uk/objects/uuid:9fd44230-32a0-41f0-861e-4ef999aea22f

Page generated in 0.0017 seconds