Keywords: Time series analysis, gender, industry, labor market
Developing a loyal customer base happens in many ways, and one option is to provide customers a well-directed service based on their needs. The ensuing questions are 1) which customers would benefit the most from each of the offered services, and 2) how we measure success. We use a state-space model to infer the impact of the services and identify the conditions where the service provision is optimal. The end goal is to increase customer loyalty, and we measure the growth rate of selected customers against the next set of customers to whom we would provide directed services had the resources been doubled. The recommendation algorithm takes into account the historical performance of services and revenue from all customers to assess the potential impact of service provision and identify those who would develop loyalty to the product offerings.