AI and Analytics for Business
(Closed for Submission)
Collaborative Research Opportunity: Customer Lifetime Value for Business-to-Business Customers
AI and Analytics for Business (AIAB) is announcing a brand new collaborative research opportunity with A Leading Supplier of Industrial Products & Services, who is seeking research proposals on customer lifetime value (CLV) research for business to business customers who purchase non-contractually through several channels. Customers in this B2B setting are notably different than typical consumers. They vary widely in their size and tend to make recurring orders with high seasonality, making this a challenging environment for predicting CLV.
The project sponsor is open to any research that will improve their ability to predict CLV and interpurchase times and manage their relationship with these customers. Specific issues the sponsor is interested in include:
- Measuring salesforce effectiveness
- Predicting the effect of pricing on CLV
- Exploiting multi-channel behavior in the prediction of CLV
- Incorporating other customer characteristics (e.g., firm characteristics, attitudes) and marketing activities into CLV predictions
- Accounting for nonstationarity in customer behavior, e.g., latent customer characteristics that can change over time.
- Relaxing traditional assumptions about the distribution of purchases and drop-out rates
- Applying new machine learning methods to CLV prediction including multitask learning and deep learning
- Developing Bayesian non-parametric estimation methods
- Understanding the trade-off between model accuracy and computational overhead
The available data includes at least two years of sales data for 10,000 non-contractual customers, along with website clickstream data and e-mail/marketing data.
The project sponsor intends to select two research teams and collaborate closely with them While both teams will be working on CLV, they will be methodologically distinct (e.g. one applying new machine learning methods and another extending traditional timing models). Teams will meet regularly with the sponsor, providing opportunities to share knowledge, clarify the business context and refine the data.