Developing a Theory of Digitally-Enabled Trial-Based Problem Solving Through Simulation Methods: The Case of Direct-Response Marketing.

摘要

In turbulent business environments, change is rapid, continuous, and unpredictable. Turbulence undermines those adaptive problem solving methods that generate solutions by extrapolating from what worked (or did not work) in the past. To cope with this challenge, organizations utilize trial-based problem solving (TBPS) approaches in which they continually probe their environments or problem spaces and rapidly update their solutions in response to detected changes. Trial-based problem solving, in practice, takes advantage of new digital methods to rapidly and iteratively develop and conduct trials that take advantage of existing business intelligence (BI) infrastructures for data capture and analysis. Due to the level of practitioner interest in trial-based problem solving, and the important role of information systems in TBPS, there is a need for a new theory that identifies and explains the mechanisms linking digitally-enabled TBPS to problem solving performance in turbulent environments.;The emerging use of digitally-enabled trial-based problem solving is brought to the forefront in the direct-response marketing sector, traditionally characterized by the use of low-cost market trials to probe customer responses to different permutations of advertisements. Leading firms in this sector are now using business intelligence technologies to transform direct-response marketing from an entrepreneurial trial-and-error method into a professional, digital model of TBPS that may offer lessons for other situations in which organizations face turbulent environments. Furthermore, the case of direct-response marketing surfaces important theoretical issues that have not been adequately resolved in the existing literature on problemistic search in changing environments: the manner in which trials as combinations of solution elements explain the costs and benefits of TBPS; the hierarchy of inferences a problem solver must make in TBPS; the perishability of trial-based data in turbulence; and the role of business intelligence in the TBPS feedback loop. It is a fitting context, therefore, in which to develop a preliminary model of trial-based problem solving as a stepping stone to a more comprehensive theory.;In this dissertation, I employ simulation methods to develop a preliminary theory of digitally-enabled trial-based problem solving in turbulent environments, using the case of direct-response marketing, and following the methodological roadmap of Davis, Eisenhardt, and Bingham (2007). I integrate a number of simple theory building blocks into a logically precise simulation model of a digital TBPS process, including: exploration and exploitation, search scope and depth, interdependency between inferences about advertising elements and product potential, and new concepts regarding digital differences in TBPS. After verifying the simulation model's consistency with the theory, I elicit through analysis and virtual experimentation a number of new theoretical propositions about TBPS that follow from the theoretical model.;The model and new propositions together constitute a preliminary theory which contributes to our understanding in a number of ways. First, they add new nuances to the simple theories that they incorporate, in particular, distinguishing two often conflated interpretations of the exploration/exploitation trade-off, and showing that some propositions hold for the compromise between dual search objectives while others hold for the issue of conservatism or radicalism in the search for novel solutions. Second, they offer new concepts for modeling problemistic search and adaptive problem solving, including trials as combinations of solution elements, knowledge perishability, and a hierarchy of inferences, and suggest new research questions for business intelligence researchers. Third, they serve as the kernels of a design theory for a digital capability for TBPS, directing researchers and practitioners alike to its critical challenges—intelligence of the feedback loop and rapidity of iteration between trials and analysis. The specific knowledge contributions of my research are a preliminary theoretical model of digitally-enabled trial based problem solving and a design theory for a digital capability for TBPS.