CLVTools vs WorthSignal

CLVTools is a free, GPL-3-licensed R package for probabilistic customer attrition modeling when purchases are not governed by contracts. It implements Pareto/NBD, BG/NBD, Gamma/Gompertz/NBD, and Gamma/Gamma models. Using transaction histories with customer IDs and purchase dates, plus prices in its example data, it predicts individual customer behavior such as expected future transactions and the probability that a customer remains active. Forecast horizons can be configured, supporting predictive customer lifetime value work, cohort analysis, and customer segmentation. Time-invariant covariates work across the models, while some models accept time-varying covariates. Extensions include correlated purchase and attrition processes, L2 regularization, and equality constraints for contextual-factor parameters. Stable installation is available from CRAN; a development version uses GitHub and devtools. Linux, macOS, and Windows are supported, though compiling from source requires the external GNU GSL C++ library. Documentation includes a walkthrough, reference manual, and vignettes.

WorthSignal is a free, open-source toolkit that turns spreadsheet and data files into customer value analyses for marketers. It reads XLSX, XLS, XLSM, CSV and JSON files, suggests column mappings and lets users correct them without changing the uploaded file. Its analyses cover nine areas, including RFM targeting, customer lifetime value, customer equity, acquisition and retention budgets, Markov switching ROI, contractual retention, BG/NBD, BG/BB and complaint recovery. Results can be downloaded as Excel or JSON. In local mode, analysis runs in memory on the user's computer; WorthSignal says uploaded customer data is not intentionally sent to its maintainer or a third-party API. Hosted deployments process files on the chosen server, and that operator is responsible for access controls, logs, retention and privacy obligations. The app requires Python 3.10 or newer and can be launched on Windows or Mac, through terminal instructions for any operating system, or with Docker. Its documented methods are described as classic, deliberately simple models. The project uses the AGPL-3.0-or-later license, with commercial use subject to its conditions.