Customer Lifetime Value - Pareto/NBD (BTYD) Model

CLV and Buy Till You Drop Models

Customer Lifetime Value (CLV) models and Buy Till You Drop (BTYD) models are used to estimate the total value a customer brings to a business over their entire relationship. These models help businesses make informed decisions about customer acquisition, retention, and marketing strategies.
PPC
data science
digital marketing
quantitative marketing
intelligence
Author

Oren Bochman

Published

Tuesday, September 14, 2021

Keywords

CLV, BTYD, Pareto/NBD, BG/NBD, customer lifetime value, churn prediction, retention strategies, customer segmentation, predictive analytics, Bayesian modeling, Bayesian data analysis, hierarchical models

Figure 1: customer churn time series
NoteTL;DR - Too Long; Didn’t Read on Buying till you are dying

Buy till you die in a nutshell

Buy till you die in a nutshell

The Buy Till You Die (BTYD) model is an early customer lifetime value (CLV) model that estimates the total value a customer brings to a business over their entire relationship. The pain point it addresses is the invisible churn event – the fact that your when you customers leave they don’t tell you they are leaving.

The model is based on the idea that customers will continue to make purchases until they “die” or simply churn. This model helps businesses understand customer behavior, predict future sales, and make informed decisions about marketing and retention strategies.

In this post I look at the first such model, the Pareto/NBD model in which the authors mapped out the behavioral story and developed a hierarchical Bayesian framework for “counting your customers”. As such the model has some challenging mathematics as well as some baked in assumptions.

In the next two posts we will cover two models that made the mathematics much simpler and those are the ones usually implemented in practice. So in this post I’ll go over the assumptions, the behavioral story, then try to present a simple hierarchical model with a plate diagram that can help understand what is going on at a glance.

Besides this I will also try to explain the terminology, unpack the business questions. Finally I would also like to outline why the model got the technical name Pareto / NBD, and the authors approach to estimating its parameters. (There are two parameters per customer)

Citation

BibTeX citation:
@online{bochman2021,
  author = {Bochman, Oren},
  title = {Customer {Lifetime} {Value} - {Pareto/NBD} {(BTYD)} {Model}},
  date = {2021-09-14},
  url = {https://orenbochman.github.io/posts/2021/2021-09-10-clv-models-1/},
  langid = {en}
}
For attribution, please cite this work as:
Bochman, Oren. 2021. “Customer Lifetime Value - Pareto/NBD (BTYD) Model.” September 14, 2021. https://orenbochman.github.io/posts/2021/2021-09-10-clv-models-1/.