Predict customer spending at Time t?

[SSMK-wq]

I am using this package to predict the customer spending (revenue from the customer) in the next 3, 6 and 9 months

But I don't see any option in the package to declare the time horizon (t).

Is it not possible?

Let's say I have data from 2017-01-01 to 2022-06-30.

Now, I wish to train the model based on data from 2017-01-01 to 2021-12-31.

Test the model on 2022-01-01 to 2022-06-30

Once this is done, I would like to find out how much revenue can we expect from customers in next 6 months (which is till 2022-12-30)?

[bachmannpatrick]

Hi!

Thanks a lot for using CLVTools.

I guess your using the Gamma/Gamma model: gg()
The current spending model integrated in CLVTools (Gamma/Gamma) is stationary. It predicts the average spending for feature transactions. There is no time component in this model. If you are interested in future customer revenue you also need to estimate the number of future transactions. You can use a PNBD model to do that. Discounted expected future transactions (DECT) times predicted average spending results in the expected customer revenue.

I hope this helps!
Patrick