skills.rest
weisberg
weisbergCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

clv-modeling

Predict customer lifetime value from transaction data using RFM and BG/NBD models.

1|Updated Mar 19, 2026
Search Tags:#rfm#bayesian#customer-segmentation#predictive-modeling#clv-modeling#bg-nbd#gamma-gamma
One-click install
npx skills add https://github.com/weisberg/agile_agentic_analytics --skill clv-modeling
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: clv-modeling
Source: https://github.com/weisberg/agile_agentic_analytics/tree/main/plugins/marketing-analytics/skills/clv-modeling
Command: npx skills add https://github.com/weisberg/agile_agentic_analytics --skill clv-modeling

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires lifetimes, pymc-marketing, arviz, pandas, numpy, scipy, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Forecasts customer lifetime value by combining purchase frequency, monetary value, and churn risk to enable targeted retention and smarter marketing investment.

Core Features & Use Cases

  • End-to-end CLV modeling for non-contractual (transactional) and contractual (subscription) businesses
  • Per-customer CLV predictions with confidence intervals, CLV:CAC ratios, and segmentation
  • Supports both MLE (lifetimes) and Bayesian (PyMC-Marketing) approaches, plus holdout validation

Quick Start

Run the CLV pipeline on your transaction data by executing the run_prediction_pipeline script with appropriate paths.

Frequently Asked Questions about clv-modeling

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I forecast customer lifetime value using RFM and churn risk?▼

To forecast customer lifetime value, combine purchase frequency, monetary value, and churn risk to predict future revenue per customer. This skill supports both transactional and subscription data for retention optimization.

What is the difference between BG/NBD and Bayesian approaches for CLV modeling?▼

CLV modeling distinguishes BG/NBD and Bayesian approaches by offering both MLE via lifetimes and Bayesian inference via PyMC-Marketing. You can compare models using holdout validation to select the best fit.

Can I predict CLV for both non-contractual and subscription businesses?▼

Yes, you can predict CLV for non-contractual and subscription businesses. The pipeline handles both contexts by fitting BG/NBD and Gamma-Gamma models, or optional Beta-Geometric models for contractual churn.

How do I generate customer segmentation from CLV predictions?▼

Generate customer segmentation from CLV predictions by using per-customer revenue forecasts, confidence intervals, and CLV:CAC ratios. These outputs enable segment-based budgeting and targeted retention strategies.

Does this CLV pipeline support holdout validation for model fitting?▼

Yes, this CLV pipeline supports holdout validation for model fitting. You can evaluate the accuracy of MLE and Bayesian predictions by comparing forecasted revenue against actual holdout transaction data.

Related Skills

Explore More Marketing & Sales Skills→
liangdabiaoliangdabiao
264

rfm-customer-segmentation

Analyzes CSV transaction data to compute RFM metrics and segment customers via K-means clustering with Chinese-language support.

Community
Advanced
liangdabiaoliangdabiao
264

ltv-predictor

Predict customer lifetime value from order data using RFM features.

Community
Advanced
xjtulycxjtulyc
33

bayesian-stats

Build Bayesian models with PyMC 5.x and ArviZ using NUTS sampling.

Community
Advanced
tomtoto757tomtoto757
14

customer-segmentation

Segment customers by purchase recency, frequency, and monetary value using NTILE-based RFM scoring.

Community
Advanced