1
$\begingroup$

I'm tasked with performing a lifetime value analysis for the clients of a small to medium sized business, however I have an issue that client turnover is actually quite slow and there are not many 'deaths' so far to beging the process of estimating mean survival time.

Are there any techniques I can apply or should be aware of for estimating the lifetime value of these clients, givent he relatively young age of the company and the low level of client drop off.

(The company has approximately 450 clients and probably fewer than 50 drop offs so far, for reference).

$\endgroup$
  • 3
    $\begingroup$ 50 out of 450 seems reasonable for 'regular` survival models. Some of your estimates may have large confidence intervals, but I don't see a way around that (although others may). $\endgroup$ – Peter Flom - Reinstate Monica Oct 1 '13 at 10:25
0
$\begingroup$

This seems to be a perfect case for Survival Analysis. I think Cox regression without any censor would be a perfect match.

The crux of survival analysis is hazard function (or survival function) which is exponential function indicating average probability of survival/deaths for a group.

Refer this paper for implementing Cox regression in R.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.