CLVTools

CLVTools is an R package for analysts modeling customer purchasing and attrition in non-contractual businesses. Its appeal is a focused set of probabilistic models and customer-level predictions; it is less suited to anyone seeking a ready-made customer platform.

Overview

In a non-contractual setting, customers can stop buying without formally cancelling. CLVTools estimates expected future transactions and the probability that each customer remains active, making it useful for predictive customer lifetime value work as well as cohort analysis and segmentation. The forecast horizon is configurable.

The models are generally used on cohorts whose members joined at the same time, so the package is a natural fit for aligned customer groups. It takes transaction histories with customer IDs and purchase dates; the example also uses prices. Analysts should expect to shape and analyze data in R rather than work through a turnkey interface.

Key features

Choice of probabilistic models

CLVTools implements Pareto/NBD, BG/NBD, Gamma/Gompertz/NBD, and Gamma/Gamma. That range gives analysts several approaches to purchase and attrition behavior in one package, with Gamma/Gamma also supporting monetary-value modeling. Individual-level estimates can distinguish customers that a single cohort average would treat alike.

Covariates and model controls

Time-invariant covariates are supported across the models, while time-varying covariates are available for some. Analysts can also use extensions for correlation between purchase and attrition processes, L2 regularization for contextual-factor parameters, and equality constraints between those parameters. These options provide room to tailor model specification, but they are most valuable to users comfortable making statistical choices in R.

Installation and documentation

The stable release is installed from CRAN; a development version is available from GitHub using devtools. Walkthroughs, a reference manual, and vignettes on modeling, advanced techniques, and package classes provide routes into the package. The project welcomes contributions, with bug reports directed to GitHub Issues and contributions handled through pull requests.

Pricing

CLVTools R package: free. The package is open source under GPL-3, with no paid tier described. R is required. Installing the stable release from CRAN avoids the source-compilation step; compiling from source requires the external GNU Scientific Library (GSL). That dependency matters most to users setting up a development environment or building the package themselves.

Platforms

CLVTools supports Linux, macOS, and Windows. Its source installation guidance covers Linux, Intel and Apple Silicon Mac, and Windows. The package requires R, so operating-system support does not make it a standalone desktop application.

Who it's for

CLVTools suits analysts and researchers with customer transaction histories who need probabilistic forecasts for non-contractual purchasing, especially when working with same-start cohorts or individual-level estimates. It is a poor match for teams that want a visual campaign tool, an all-purpose customer platform, or analysis without R.

Pros and cons

• Pros: Four model families cover purchase, attrition, and monetary-value analysis within one R package.
• Pros: Customer-level transaction and alive-probability predictions support more differentiated CLV work than cohort averages alone.
• Pros: Covariates and model extensions offer additional control for users who need to tailor their specification.
• Cons: R is required, which rules it out for users seeking a no-code workflow.
• Cons: Source compilation depends on GSL, adding a setup requirement for users building from source.
• Cons: The emphasis on non-contractual settings and same-start cohorts narrows its fit for other customer-analysis needs.

Alternatives

For a broader selection, browse Customer Lifetime Value Software.

• PyMC-Marketing is another free, open-source modeling library; choose it if an Apache 2.0 Python library is a better fit than an R package.
• Pushwoosh offers a free plan with push notifications, email, and AI requests; choose it for messaging capabilities rather than CLV modeling.
• RevenueCat has a free Pro plan up to $2,500 in monthly tracked revenue, then charges 1% of tracked revenue; it is the alternative to consider for that revenue-based plan structure.
• SaaSFlow has a free plan with all features, unlimited users, and unlimited data sources up to €120k ARR (€10k MRR); its Pro plan is €100 a month with annual billing. Consider it if those SaaS metrics and terms fit your needs.
• Siftmo offers a free Starter plan capped at 50 monthly orders, with Growth at $29 a month for up to 500; choose it if those order-based tiers match your requirements.
• Klaviyo has a free plan capped at 250 active profiles, 500 email sends per month, and $5 of mobile messages per month; it may fit a small messaging workload within those limits.
• WorthSignal is another free option.
• VOZIQ AI Platform is available free of charge on a BYOL basis, with AWS infrastructure costs potentially applying.

Verdict

Choose CLVTools if you work in R and need established probabilistic models with individual-level predictions for non-contractual customer histories. Its modeling range and specification controls are the case for choosing it; the R workflow, source-build dependency, and cohort-oriented scope are reasons to look elsewhere.