c***z
发帖数: 6348 | 1 How to calculate LTV of existing and potential customers? This is a question
I got frequently asked.
My idea is to calculate the following first
1. time to stay with us
2. time to first spending
3. frequency of spending
4. mean amount of spending
Since 3 and 4 are independent according to e-commerce literature, we
calculate
LTV = (1-2)*3*4
Any thoughts? Thanks a lot1 | c****t
发帖数: 19049 | 2 差不多吧。能想到这很不错了，这年头是个人resume上就有lifetime value。能说出这
个框架的大概一，二百个candidates里有一个。
可惜的是实际上1,2,3,4都互不独立嘿嘿。2并且不是个survival distribution。不知
你用什么model "frequency"。如果有其它可用指标这里可以做个hidden markov | c***z
发帖数: 6348 | 3 大牛能展开说说么
特别是hidden markov这块
我没有什么好主意model "frequency", what I have in mind is Poisson... |
|
