FROM CHAOS TO ORDER: A study on how data-driven development can help improve decision-making

Abstract The increasing amount of data available from software systems has given a unique opportunity for software development organizations to make use of it in decision-making. There are several types of data such as bug reports, website interaction information, product usage extent or test results coming into software-intensive companies and there is a perceived lack of structure associated with the data. The data is mostly scattered and not in an organized form to be utilized further. The data, if analyzed in an effective way, can be useful for many purposes, especially in decision-making. The decisions can be on the level of business or on the level of product execution. In this paper, through a literature review, an interview study and a qualitative analysis we categorize different types data that organizations nowadays collect. Based on the categorization we order the different types of decisions that are generally taken in a software development process cycle. Combining the two we create a model to explain a recommended process of handling the surge of data and making effective use of it. The model is a tool to help both practitioners and academicians who want to have a clearer understanding of which type of data can best be used for which type of decisions. An outline of how further research can be conducted in the area is also highlighted