retentionizer

Retentionizer is a Python package that applies the shifted-beta-geometric distribution model from Fader & Hardie's research to project customer retention rates and calculate lifetime value (LTV). It helps businesses forecast cohort survival and make data-driven decisions about customer value.

Key Features

• Retention Projection — Fits a shifted-beta-geometric distribution to historical cohort survival data to project future retention rates.
• Cohort Analysis — Displays projected retention rates and imputed beta distributions for each customer cohort.
• LTV Calculation — Computes the lifetime value of a given customer within their cohort based on the fitted model.
• Visual Output — Generates clear graphical representations of retention projections and distribution fits.

Philosophy

Retentionizer provides an accessible, open-source implementation of academic research on customer retention modeling, enabling practical application without proprietary tools.