Skip to content

chrisclark/retentionizer

Repository files navigation

Try it live!


Using the techniques in How to Project Customer Retention by Fader & Hardie (2006), and an implementation of those techniques by JD Maturen, Retentionizer will fit a shifted-beta-geometric distribution to the data, show the projected retention rates for each cohort, show the imputed beta distribution for each cohort, and calculate the LTV of a given customer in that cohort.

Basically, it turns a sample of cohort survival rates:

t past 30
0 1.0
1 .81
2 .80
3 .76
4 .75
5 .72
6 .70
7 .67
8 .66
9 .65
10 .64

Into this:

sample-output

Retentionizer is built by David Chudzicki and Chris Clark.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published