Critical role of measures in decision processes: Managerial and technical measures in the context of large software development organizations

Abstract

ContextToday, many software development organizations struggle to establish measurement programs to monitor their projects, products and units. After overcoming the initial threshold of establishing the measurement program organizations stand before the questions of which measures should be collected in order to lead to actions or at least effectively trigger decision processes. ObjectiveThe objective of this paper is to investigate how to use measures in an effective way in decision processes. This dependency is examined through a case study – Ericsson in Sweden. Two models of these dependencies are recognized a priori – metrics-push and metric-pull – and in the study the models are used to describe how measures affect decisions and vice versa. MethodThe research method is a case study of the measurement program of one of the product development units of Ericsson in Sweden. The participants are carefully selected from the management teams at different levels of organizations. The objects are measures and decisions at these management levels. The instruments are interviews and observations. The results obtained at Ericsson are validated through interviews at another company – RUAG Space. ResultsThe results show that effective use of measures as evidence for decision processes does not require a large number of measures (ca. 20 at the top management level). It was found that there are four types of measures which are used in different ways in the context of decision formulation and implementation (which we call decision-measures dependency model). The critical aspects of effective measures in decision-making context are completeness, reliability and providing early warnings. It was also found that the time between the decision and when its results can be observed via measures (length of the feedback loop) is a crucial aspect determining at which organizational level a measure should be placed. ConclusionsAfter overcoming the initial threshold of establishing measurement programs the organizations demand non-functional properties from the measures. These non-functional properties like completeness, providing early-warning or trust determine whether decision processes are triggered by measures or not.