Analysing research on Information Systems success and failure: A Machine Learning Technique

Abstract

The existing landscape of research on the outcomes of information systems (IS) implementation suggests that this domain is extensively researched. Results indicate that most IS implementations fail, yet the uptake of these systems is still on the rise. This inconsistency might be because of how IS success and failure are assessed, defined and framed in research; hence, there is a need for the systematic and comprehensive characterisation of research in this domain. To achieve this, we adopt a machine learning technique called latent semantic analysis (LSA) to identify the major themes and topics on IS success and failure in literature. The turn to a machine learning technique is valid given the ever-increasing volume of textual data in research, and the inability of traditional research approaches to keep up with this growth. For this study, abstracts from 379 journal articles from EBSCOhost were fed into the LSA process as input and 37 themes were identified as output. These themes are discussed under five broad categories labelled: model/theory, user, information technology (IT) artefact, factors and context.