Accelerated risk management using statistical triggers

Abstract

In any large IT services provider's portfolio, it is not uncommon to find several deals that have a high customer satisfaction rating, but result in a financial loss. In many cases it is not clear that an IT engagement that is in serious trouble should be immediately terminated. The provider may have good reasons in continuing an engagement which may not be eventually profitable. For example, the provider wants to maintain a relationship with a client, or avoid an adverse impact on their reputation. It is important to not only identify that a project has become troubled, but also be able to predict whether the project can be salvaged from financial, quality or other perspectives. By drawing upon a historical database of services projects spanning several years, we are able to draw conclusions on the effectiveness of certain well established principles in risk assessment used by the project management community. We define derived statistical measures that can be used to predict the eventual outcome of an in-delivery project along several dimensions. In this paper, we will explore how quantifiable measures of project progress, gathered at several important stages of a project's life-cycle, can aid in early identification of troubled projects.