سنجش موفقیت و عدم موفقیت در برون سپاری IT
Abstract
1- Introduction
2- What is ITO success anyway
3- Outcome prediction
4- Intervention
5- Rigid factors
6- Timing the interviews
7- Conclusions
References
Abstract
We implemented five easy-to-complete questionnaires in Excel, which could serve as early warning signals for practitioners interested in the odds of their IT-outsourcing deals and could serve to redirect their course when still possible. The questionnaires are based on our earlier published longitudinal, observational study on 30 representative ITO-deals in the Netherlands, of which we know whether they failed or not. Our questionnaires predicted their outcome correctly. To help redirect the course of a dubious deal, we developed a questionnaire estimating the odds in relation to boosting strongly significant critical success determinants. Another questionnaire guides practitioners how to further improve on less critical factors. There are no specific reasons that limit our results to the Dutch situation, which makes it promising, therefore, to apply the Excel as an aid in improving ITO deals in other contexts.
Introduction
In 2016 we published an article in Science of Computer Programming which dealt with the research findings from a longitudinal study on IT-outsourcing (ITO) deals in the Netherlands (Delen et al., 2016). About 60 organisations participated: clients (also called outsourcers), suppliers (also called vendors) and intermediaries (also called sourcing consultants). The research sample is a representative cross-section for 700 IT-outsourcing deals in the Netherlands. Representativeness was statistically proved through validations that the sample reflected the Dutch economic sectors, the duration of the deals and the type of outsourced work of the total Dutch ITO-deal population reasonably well. For more details, see Delen et al. (2016). This is a very important result, for it implies that findings of the sample generalize to the entire population. So our Excel-tool can be used for the entire population. Not everybody will immediately realize what this actually means, so we shall elaborate on this fundamental statistical rule below. When statistical tests are used we usually accept a 5% chance that although the null hypothesis is true we still reject it. The probability of making that (type I) error is often called α. Vice versa when the null is false and we fail to reject it, this is called a type II error, and the probability for such errors is often called β and we usually accept a 20% chance of making a type II error.