Paper Type

Complete Research Paper

Description

Information systems (IS) projects are famous for experiencing severe cost overruns, which amongst others are often caused by inaccurate ex-ante cost estimations. Against this background, this article presents a descriptive case study located in an IS transformation program at a major German financial services provider. In this case study, a multi-stage cost estimation process, which was applied to 79 IS projects, is described and the estimation accuracy of the cost estimations of all IS projects is determined using different estimation accuracy measures: Estimating Quality Factor, Forecast Error, and Mean Absolute Percentage Error. Depending on the concrete estimation accuracy measure used for the evaluation, the overall estimation quality of the program turns out to be evaluated as good or at least average "“ which seems to be contrary to most studies in scientific literature. However, the results further reveal that the estimation accuracy also depends on the estimation accuracy measure chosen for the evaluation. These differing judgements are discussed from a management perspective.

Share

COinS
 

ESTIMATION ACCURACY IN LARGE IS PROGRAMS - INSIGHTS FROM A DESCRIPTIVE CASE STUDY

Information systems (IS) projects are famous for experiencing severe cost overruns, which amongst others are often caused by inaccurate ex-ante cost estimations. Against this background, this article presents a descriptive case study located in an IS transformation program at a major German financial services provider. In this case study, a multi-stage cost estimation process, which was applied to 79 IS projects, is described and the estimation accuracy of the cost estimations of all IS projects is determined using different estimation accuracy measures: Estimating Quality Factor, Forecast Error, and Mean Absolute Percentage Error. Depending on the concrete estimation accuracy measure used for the evaluation, the overall estimation quality of the program turns out to be evaluated as good or at least average "“ which seems to be contrary to most studies in scientific literature. However, the results further reveal that the estimation accuracy also depends on the estimation accuracy measure chosen for the evaluation. These differing judgements are discussed from a management perspective.