One of the greatest challenges facing industry in the new century will be t
he process of selecting which new products to develop. Capital resources an
d manpower are limited. Stakeholders are demanding ever-increasing rates of
return. These problems are especially difficult in highly regulated indust
ries such as the chemicals and life sciences businesses, where development
times are long and costly. In formative industries such as biotechnology, r
egulatory requirements continue to tighten, as public perception is often m
ore influential than science in the approval process. Engineers are comfort
able building process models. However, they infrequently think about the de
velopment of new products or the selection of new products as processes. Th
is study is an attempt to get the engineers involved in the new product dec
ision making process. Using the pharmaceutical industry as an example, prob
abilistic network models are used to capture all the activities and resourc
es required in the 'process' of developing a new drug. The data representin
g each new product candidate are then combined into a simulation model of t
he new product development pipeline. This simulation model can be used by m
anagement to obtain insights into a new product portfolio, which will provi
de high rates of return at an acceptable level of exposure to risk for the
corporation. (C) 2000 Elsevier Science Ltd. All rights reserved.