Lean software development measures : A systematic mapping

Abstract: Context. Process improvement using lean software development uses measures to answer information needs. Measures are important in process improvement as they identify whether improvements have been made or further improvements are needed. Objectives. This study aims to identify the measures proposed in literature for lean software development and structure them according to ISO/IEC 15939. Methods. The research methodology consists of systematic mapping and uses thematic analysis. Results. Lean software development literature has become more frequently published from 1996 to 2013. The most common research types were evaluation research and experience reports. Themes were identified in measures resulting in the identification of 22 base measures, 13 derived measures and 14 indicators in total. Identified measures were then structured using a proposed meta-model adapted from ISO/IEC 15939. Using the proposed meta-model twelve models of measures were instantiated. Conclusions. Gaps exist in the lean principles for deferring commitment and respecting people. Structuring measures in models presents opportunities to identify shared dependencies in measures. Lean software development research guidelines were defined. Further research into the comparison of indicators, the industrial use of measures and the representation of models of measures is needed.