Bart Baesens

← Back to courses

Customer Lifetime Value Modeling

📅 September 24-25th, 2020 (9am -5pm)
🌍 English

About This Course

In this course, participants learn the essentials of Customer Lifetime Value (CLV) Modeling. We start by setting the stage and defining CLV and its key parameters. We then elaborate on RFM analysis, a key building block to many CLV models. We extensively zoom in on churn prediction: how to define it, predict it and evaluate the resulting analytical models. Next, we cover response modeling for both customer acquisition and deepening customer relationships. Markov chains are discussed as a very handy and intuitive approach for CLV modeling. We then review some probability models for CLV such as the Pareto/NBD model. Survival analysis is covered in depth since estimating accurate survival probabilities is key to calculating well-calibrated CLV values. We zoom in on profit-driven machine learning for CLV Modeling and discuss ProfLogit and ProfTree, both developed in our research group. We then briefly review some of our research on CLV modeling and churn prediction. The course concludes with some closing thoughts on the topic.

The course provides a sound mix of both theoretical and technical insights , as well as practical implementation details. These are illustrated by several real-life case studies and examples. Throughout the course, the instructors also extensively report upon their research and industry experience .

The course also features code examples in both R and Python and R/Python tutorials are also provided.

Course Outline

  • Introduction
    • Instructor Team
    • Our Publications
    • Course Outline
    • R/Python Software
    • R/Python Tutorials
    • Disclaimer
  • Setting the stage
    • Customer Lifetime Value: Drivers
    • Customer Lifetime Value: Definition
    • Customer Lifetime Value: Key Parameters
    • Customer Equity
    • CLV Modeling: Example
    • CLV: Strategic Actions
  • RFM Analysis
    • RFM Framework
    • Recency
    • Frequency
    • Monetary
    • RFM Interactions and Correlations
    • Operationalising RFM
    • RFM Usage
    • RFM Measurement Level
    • RFMPD
    • RFM Out of the Box
    • Closing Thoughts
  • Churn Prediction
    • Churn Prediction: Basic Idea
    • Defining Churn
    • Types of Churn
    • Churn Prediction Analytics Model
    • Churn Prediction: Data
    • Churn Prediction: Feature Engineering
    • Churn Prediction: Target Definition
    • Churn Prediction: Analytical Techniques (with example in Python)
    • Churn Prediction: Social Network Effects
    • Analytical Churn Prediction: KPIs
    • Churn Prediction versus Churn Prevention
    • Uplift Modeling for Churn Prediction (with example in R)
    • Maximum Profit Measure: General Definition
    • Maximum Profit Measure for Churn
    • Expected Maximum Profit (EMP) (with example in R)
    • Expected Maximum Profit for Churn (EMPC)
    • Churn Prediction: Scientific Impact
  • Response Modeling
    • Response Modeling: Basic Idea
    • Marketing Campaigns
    • Response Modeling: Definition of Target
    • Response Modeling: Data
    • Response Modeling: Analytical Techniques
    • Response Modeling: Evaluation
    • Response Modeling: Uplift Modeling
  • Markov Chains
    • Markov Chains: Basic Idea
    • Markov Chains: Example
    • Markov Chains: Simulations (with example in Python)
    • Markov Reward Process (with example in Python)
    • Markov Decision Process Approach (with example in Python)
    • Markov Chains and Customer Heterogeneity
    • Customer Migration Mobility (with example in Python)
    • Modeling Customer Migrations
    • Markov Chains: Evaluation
  • Survival Analysis
    • Survival Analysis
    • Censoring
    • Time Varying Covariates
    • Survival Distributions
    • Kaplan-Meier Analysis (with example in Python)
    • Accelerated Failure Time (AFT) Models
    • Proportional Hazards Model (Partial Likelihood) (with example in Python)
    • Discrete Survival Analysis
    • Competing Risks
    • Mixture Cure Modeling
    • Evaluating Survival Analysis Models
  • Profit-Driven Machine Learning
    • ProfLogit: Profit Driven Logistic Regression (with example in Python)
    • ProfTree: Profit Driven Decision Trees (with example in R)
  • Closing Thoughts
    • Customer Accounting
    • Sample Bias
    • Model Risk
    • Deep Everything
    • Leader versus Follower
    • Privacy
👩‍🏫 Lecturers

Prof. dr. Matthias Bogaert
Professor at Ghent University

Prof. dr. Bart Baesens
Professor at KU Leuven

🏢 Location

Van der Valk Hotel Brussels Airport (Belgium)

Culliganlaan 4b
1831 Diegem

🏫 Organizer

Bart Baesens

💼 Register

The price is 1,000 Euro (VAT Exclusive) for both days. This includes:

  • a copy of the course material (including R/Python code)
  • lunches and coffees

Online registration

Price and Registration

The price is 1,000 Euro (VAT Exclusive) for both days. This includes:

Please register through the link below. After processing your payment, you'll be sent a confirmation e-mail to confirm your registration.

Online registration