The Interplay of the Levers of Control in Product Development

This paper contributes to the research on management control systems (MCS) in product development (PD) by exploring the performance effects and the interplay of the levers of control (LOC), i.e., interactive, diagnostic, beliefs, and boundary control systems. We use data from a survey of 468 senior managers from the manufacturing industry and identify the direct and indirect effects of the LOC on PD performance and organizational performance. Therefore, we first scrutinize the possible types of interplay of the levers by comparing by structural equation modeling a base model of unrelated LOC (additive model) with (a) a model of mutual association of LOC (interdependence model), (b) a model of joint use of LOC (interaction model), and (c) a combined model of mutual association and joint use of LOC (combined interdependence/interaction model). The results show that the interdependence model is the most suitable in terms of fit to the data and parsimony, providing evidence of the fact that the LOC move together and influence PD performance and organizational performance both directly and mediated through each other. Interestingly, PD performance is positively affected by both the beliefs system and the diagnostic use of performance measures, as these LOC have a significant direct effect and generate the largest total effects on PD performance. However, the boundary system and the interactive use of performance measures, the last being praised in previous literature for its suitability for PD, have no significant direct effects, show lower total effects than beliefs and diagnostic control systems, but have rather meaningful indirect effects on PD performance. These results hold across different industries and firm sizes. The paper contributes to research and practice by uncovering the interdependence of the LOC in PD, thus clarifying the role of each of the levers for PD performance and organizational and therefore providing suggestions for a more successful design of LOC for PD. Furthermore, it directs the attention of researchers to the consequences of potential misspecifications of LOC models