Analysis of agile software development skills by analysing product backlogs using INVEST and DEEP

Abstract

The transparency of processes and rapid reaction to changing customer requirements can be enhanced by adapting agile software development methods. The agile methods provide many different tools for improving management of processes and projects.

One of these tools is a product backlog. A backlog is a prioritized list used for managing of user stories. Within a project there can be as many backlogs as there are their administrators and from time to time qualitative and structural analysis of these backlogs is required.

One way of analyzing backlogs is using criteria of INVEST and DEEP acronyms. The aim of this thesis is to investigate what kind of measurement models can be created based on these criteria to assess the quality of backlogs and its items and how viable these measurement models are for backlog analysis.

The study conducted in this thesis is based on the backlogs of two development teams from the case company. Created measurement models are applied to these backlogs and the results are assessed to determine whether or not the created measurement models are viable to evaluate the quality of a backlog.

This thesis concludes that creation of measurement models according to INVEST and DEEP criteria is not straight forward because of the abstract nature of some of the criteria. Evidently, few of the measurement models created from the more abstract criteria do not serve their purpose satisfyingly whereas the models created for others can be used to some extent.