A stochastic programming approach for marketing campaign optimisation

Abstract

A marketing campaign is a series of activities that effectively promote an organisation's service or product. Optimising a marketing campaign by providing the right product to the right customer at the right time is challenging since there are multiple products and complex business constraints. One of these constraints is demand uncertainty, which deals with customer behaviour and is continuously impacted by numerous factors. Operations research focusing on optimisation under uncertainty is part of an ongoing effort to solve and optimise these types of complex problems. Optimisation under uncertainty consists of three primary approaches, and one of these approaches is known as stochastic programming. Stochastic programming models are developed to provide optimal solutions hedged against uncertainty. From a systematic literature review (SLR) conducted in the literature, it was concluded that there is no stochastic programming model that addresses the problem identified in this study. This study addresses the gap by proposing a two-stage stochastic programming model called a recourse model. First, two deterministic integer linear programming (ILP) models are identified from the literature for campaign optimisation. These two models are considered base models. Second, a uniquely defened deterministic model is formulated based on the marketing fundamentals of Model 1 and Model 2, respectively. The proposed deterministic model aims to maximise the profitability of a campaign while excluding uncertainty. Last, a deterministic model's counterpart, a recourse model, is developed. The proposed recourse model seeks to maximise the profitability of a campaign while providing solutions hedged against uncertainty. All four optimisation models are verified by critically evaluating each constraint's impact on the model's functionality and results. Subsequently, the model performed as expected in each instance, confirming that the models are correctly formulated and coded in CPLEX. The proposed recourse model was validated by showing that the model maximises the campaign's profitability while providing solutions hedged against uncertainty. Many factors continuously in uence customer behaviour, and retailers need to adopt approaches that accommodate uncertainty while maximising profitability. However, there is a limit to stochastic applications in the literature. Therefore, the main contribution of this study is the formulation of the recourse model and the value that this approach

adds when dealing with uncertainty in the decision-making process. The recourse model proposed in this study will provide retailers with an opportunity to make decisions hedged against uncertainty. Furthermore, the recourse model revealed a new horizon of future possibilities that should be investigated in the retail industry.