Environmental volatility, development decisions, and software volatility: a longitudinal analysis

Although product development research often focuses on activities prior to product launch, for long-lived, adaptable products like software, development can continue over the entire product life cycle. For managers of these products the challenges are to predict when and how much the products will change and to understand how their development decisions influence the timing and magnitude of future change activities. We develop a two-stage model that relates environmental volatility to product development decisions and product development decisions to software volatility. The model is evaluated using a data archive that captures changes over 20 years to a firm's environment, its managers' development choices, and its software products. In Stage 1 we find that higher environmental volatility leads to greater use of process technology and standard component designs but less team member rotation. Earlier development decisions strongly influence current development choices, especially for product design and process technology. In Stage 2 we find that increased use of standard component designs dampens future software volatility by decreasing the average rate and magnitude of change. Adding new team members increases product enhancements at a faster pace than more intense use of process technology but adds repairs at almost the same rate as enhancements.