Organizing product releases in time-driven development processes: a probabilistic analysis

In this paper, the problem is addressed of how to organize a new product development process that is fast on the one hand and that provides good-quality results on the other hand. Several planning techniques, like PERT and CPM, are available to analyse the completion time or product release time of such complex processes. Although these techniques are all well known and are widely used in project management, they do not address the iterative mode of operation that is characteristic for such uncertain processes. Neither do they offer a tool or guidelines to ‘design’ a composition of iterative processes because they are analytical. In this paper, a quantitative concept is presented for modelling the release time of a single uncertain iterative activity, as a random variable to deal with the probabilistic aspect in a simple way. From this simple model, the complexity is extended systematically to model fundamentally different configurations, that are on the one hand simple enough to be studied analytically and on the other hand exhibit their fundamentally different release characteristics, as experienced in real life. From the analysis, guidelines are formulated for organizing or (re)configuring a complex process configuration. It is demonstrated that organizing uncertain processes for a fast product release requires a balance between the exploitation of the principles of concurrent engineering and the risk of overrunning time targets. An important factor in this balance is the decision structure for the release of intermediate results. The ‘empowerment’ structure, a structure where decisions are not clustered as milestones but are made without any delay, offers great opportunity in terms of small mean release times and small variances thereof.