City Research Online

Abstract

An experiment in strategy is a deliberate, systematic intervention designed to generate general knowledge about how a proposed action will perform, how customers or markets will respond, or whether a hypothesized mechanism of value creation holds. Unlike trial-and-error learning, which extracts information from ongoing activity, an experiment is planned in advance: what is varied, what is held constant, what outcomes will be observed, and what inferences may be drawn are specified before the test is run. Real experiments sit along a gradient, from textbook randomized controlled trials with pre-registered hypotheses and large samples to lightweight minimum viable products and landing-page tests, all of which satisfy in varying degrees the core criteria of deliberation, generality, and systematic variation.

The study of experiments in strategy draws on four intellectual traditions: the statistical design tradition, the decision-science tradition, the organizational-learning tradition, and a fourth tradition at the intersection of economics, strategy, and entrepreneurship that treats experiments as deliberate acts of learning under uncertainty. Across these traditions, experiments vary between calibration and augmentation experiments depending on the form of not-knowing the decision maker faces: refining probabilities within a recognized state space, building confidence in those probabilities when evidence is conflicting, discriminating among rival models, probing for evidence that the model is itself inadequate, or expanding the state space by discovering or creating states not previously conceived. The design, interpretation, and strategic value of an experiment depend on which of these purposes it serves. Strategic experiments are also embedded in organizations and competitive environments: the act of running one can commit the firm, disclose information to rivals, build capabilities, and shift stakeholder beliefs, so that experimentation and strategic action cannot be cleanly separated. The resulting perspective helps strategy and adjacent management scholars understand why experiments differ in purpose, design, and interpretation.

Publication Type:	Book Section
Publisher Keywords:	experiments, strategy, uncertainty, ambiguity, unawareness, Bayesian updating, thought experiments, synthetic experimental data
Subjects:	H Social Sciences > HB Economic Theory H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management
Departments:	Bayes Business School Bayes Business School > Faculty of Management
SWORD Depositor:	Symplectic Administrator

Export

Downloads