by Doug Walker
Beginning immediately after the spring semester, I had the pleasure of teaching an intersession course focused on data-driven marketing with twenty students. The 3-week course revolved around an actual customer database (scrubbed of identifying information), consisting of almost 200,000 customers and 60 variables for each customer – about 12 million data points. The course was designed to simulate the experience of an internship. It was very hands-on – the students worked with the data for 2 ½ hours every day.
The course consisted of three parts. The first week, we worked on cleaning and appending the data. The second week we analyzed the data, calibrating and validating models estimating the probabilities of response to marketing activities and the expected size of the response, incorporating both logistic and linear regression. We spent the third week using a decision analysis framework to make resource allocation decisions based on our analyses. Over the final few days, each student specified, calibrated, and validated a probability model and a size model that could be used to decide which customers should receive a sales call in the future. On the final day of class, the students competed by applying their models to a holdout sample of customers to see who could produce the highest profits.
The objective of the course was to begin to prepare these students to bridge the gap between the marketing and sales managers they will one day become and the analysts that are gaining insights from the data being collected. Understanding how data can ultimately enhance decision-making is critical for contemporary managers. The students learned about managing data, analyzing data, and making data-driven decisions; as well as greatly enhancing their Excel skills along the way.
The course will be regularly offered beginning in the Spring of 2018, under the title of Customer Relationship Management, providing students with insights into the role of applied data analytics in marketing and sales.