Uncertainty, information and decision making in product innovation

Abstract

Firms in market economies play an important role by shaping the economic experiments conducted in the marketplace. This thesis explores the microeconomics of the process through which firms sift and develop potential product innovations. Our departure point is a set of quite recent models that portray product development from a sequential decision making perspective. By assuming uncertainty reduction through development activities these models predict behaviour that can contravene standard investment criteria. However, they raise many, as yet unanswered, questions about both the magnitude of uncertainty facing projects and the extent of its stage-wise resolution. Obtaining empirical support for the necessary parameters is difficult given the need for probabilistic project-level information in a field noted for poor data sources and the problems of proprietary information. Our first study attempts to overcome some of these challenges by accessing a firm's internal project database produced using the methods of applied decision analysis. By using quantified measures for uncertainty, based on subjective probability distributions, we are able to derive parameters directly comparable to those for the sequential models and show that sufficient uncertainty is present for their theoretical effects to be potentially important. By analysing the composition of this uncertainty we also show, in a more rigorous manner than past studies, the relative importance of commercial versus technical uncertainty. We then introduce and explore theoretically a concept of intra-project alternatives. Unlike existing sequential models these recognise that projects may involve not only go/no go decisions but also decisions within the project. Importantly, we show that when these alternatives are uncertain they behave in a sufficiently different manner to require their separate identification and treatment in empirical work. Additionally, an evaluation of sequential development under risk-aversion highlights the need to also give attention to residual uncertainty in projects. We then develop a study design to explore sequential and intra-project alternative effects, again drawing upon the methods of applied decision analysis. Working closely with a New Zealand firm we then examine an actual early stage development project involving technical development uncertainty. We demonstrate significant value for sequential decision making, showing it to be comparable in magnitude to that demonstrated for the option value associated with irreversible investments under uncertainty. Two intra-project alternatives are identified and shown to yield a further similar increase in the value of the project. The two effects are thus complementary, jointly increasing the overall importance of information revelation effects. Comparing the observed uncertainty reduction profile for the project with that assumed by a particular sequential model we find the latter to be relatively optimistic about the extent of early information revelation. To examine some of the findings from the second study more rigorously we introduce a flexible two-stage model of development. In particular we explore the interaction between the incentive to wait and the incentive to develop created by information revelation. We demonstrate these effects to be counteractive, with sequential development and intra-project alternatives both reducing the option value of delaying investment. We find considerable information revelation within later, more commercial stages of the project and this raises the question as to whether the effects we demonstrate in development could be apparent in more general investment opportunities. We conclude that sequential and intra-project alternative effects are an important part of intra-firm innovation and should be included in any fuller picture of the overall innovation process. When examining uncertainty in innovation we demonstrate a need to clearly separate that portion which may be resolved through the firm’s activities since its effect is markedly different. The decision analysis based research methodology that we introduce appears to have merit for the study of intra-firm innovation processes and also perhaps for broader classes of decision.