Leveraging attribution models for enhanced scenario planning in strategic decision-making

Abstract

If scenario planning is the capability of perceiving what is going on in the business environment, thinking of the consequences of what this means and taking action in order to have a competitive edge when times get difficult from understanding trends to developing policy, forecast and foresight are essential activities and state of mind for any company and profoundly influences its strategic planning and budgeting. Defining scenario planning for the tourism industry represents a big challenge because of the implications that it involves for different interdependent economic activities. Nevertheless, as extant research indicates, the adoption of scenario planning techniques is strictly related to the need to depict future trends above all in period with great uncertainty, instability and unpredictability. In our research (still in progress) we intend to better efficiency of demand predictive models and tourism scenario foresight using sales determination measures (attribution models) and we propose this starting at the level of a single hotel in a determined destination. Attribution models play a crucial role in marketing analytics by attributing credit to various touchpoints along the customer journey. However, their utility extends beyond marketing analysis, as they can also be leveraged for scenario planning in strategic decision-making. By integrating consumer behaviour analysis into scenario planning, organizations can ensure that their scenarios are grounded in a deep understanding of consumer dynamics. In particular, we propose to capture the customers’ decision process along the emblematic stages of the conversion funnel through a typical attribution model, the self-exciting point process. This type of approach is largely used to describe phenomena for which the occurrence of an event increases the probability and manifestation of other events, close in time and/or space. Given this distinctive trait, it is use can allow us to model not only the effect of a particular click on a future purchase, but also the dynamic interactions among clicks themselves, and among them and the final purchase activity with the final aim of depicting future scenarios.