Information technology-based innovation involves considerable risk requiring foresight; yet our understanding of the way in which managers develop the insight to support new breakthrough applications is limited and remains obscured by high levels of technical and market uncertainty. This paper applies discrete choice analysis to support improved empirical explanation of how and why decisions are made in information systems (IS). A new experimental method based on information acceleration (IA) is also applied to improve prediction of future IS service strategies. Both explanation and prediction are important to IS research and these two behaviourally sound methods complement each other. Specifically, the combination of IA and discrete choice analysis removes misspecification artefacts from response variability and generates more accurate parameter estimates that better explain IS decision making.