I am working for a bank that reminds its customers that they owe money back periodically starting from 3 days before their credit deadline, on the day of the deadline, 3 days after the deadline, a week after the deadline, etc. The reminders are text messages and voice calls. I have to check how effective the each text message and call is. I have all the possible data. There is no alternative group (customers that do not get reminders). What is the best methodology for this project? I was thinking on doing an interrupted time series, however there are would be too many interruptions, as there are many events taking place. Any input will be of help.
$\begingroup$
$\endgroup$
6
-
1$\begingroup$ If you have no control how do you hope to find out the effect of reminders? You do not know how the group with no reminders would behave. $\endgroup$– user2974951Commented Jul 25, 2023 at 7:01
-
$\begingroup$ You can draw an expected line and see how much it deviates from the expected. $\endgroup$– Nemo_the_scientistCommented Jul 25, 2023 at 7:28
-
$\begingroup$ And how do you obtain this expected line? $\endgroup$– user2974951Commented Jul 25, 2023 at 7:30
-
$\begingroup$ There are many methods - moving average, weighed average, arima, sarima... $\endgroup$– Nemo_the_scientistCommented Jul 25, 2023 at 8:26
-
$\begingroup$ But this is still based on the same data, customers who received reminders? $\endgroup$– user2974951Commented Jul 25, 2023 at 8:28
|
Show 1 more comment