Forum Discussion
Help please - Cohort Analysis
How to do cohor analysis for a policies for data base which have the data as explained below:
Data file has records of policies taken by customers. Each record is populated with details of policy number.
Fields:
Crediting Date: date of transaction of amount if a policy is created or policy is updated or policy is cancelled cannot be considered as policy starting date. if a policy created early has an update in future, it will have another record with crediting date = policy update date and the policy start date changes to policy crediting date and policy end date changes to 6 months later date after the policy starting date which is equal to policy update date
Policy No: Unique policy no
Policy Start date: Policy start date- this date is tricky, if a policy is active an no update in policy this date will show the actual policy start date. if there is an update in policy, the policy no will have a new record with new crediting date = policy update date and the policy start date also changes to the policy updated date
Policy End Date: Policy start date- this date is tricky, if a policy is active an no update in policy this date will show the actual policy end date. if there is an update in policy, the policy no will have a new record with new crediting date = policy update date and the policy end date also changes to 6 months after the policy updated date
Starts_At : Date in this column is the Policy first started date, and this date does not change if there is any update in the policy, and this date remains same in every record of the policy no, i.e., in the event of policy creation and in the event of policy update.
My question is:
Customers when purchase the policy, the policy period is 6 months. We can calculate only the no of days (Tenure) the customer is active till date (lets say 87 Days) by (DATEDIFF of start_at and today()), even though his policy expiry date is in future.
Example: If a customer purchased policy in Apr-2023, he is retained until now, i.e., for that month (0 month - Apr2023), 1 month(next month May-2023), 2 month (June-2023).
His retention will be as follows
| 0 month (Apri) | 1 month(May) | 2 month(June) | |
| Apr-2023 | 1 | 1 | 1 |
How to divide those 87 days and as he is still active till now, how to populate the policy count (1) in every month from Apr-2023 as 1, 1, 1 in Apr, May, June.
Is there any possible way we can generate renetion count for a active policy, which will not records every month after the purchase of policy.
Thank you in advance
9 Replies
- Ashish_MathurSuper User
Hi,
Share some data, show the expected result and explain the expected result. I suggest that you create an MS Excel file and with the help of formulas/Pivot Tables/text boxes, show the expected result.
- sudheerbollaHelper II3 seconds ago
Thank you very much for the reply.
Sample Date:
Policy First start Date "column Name": STARTS_AT (Imported)
Outcome should be like this in Power BI:
- Ashish_MathurSuper User
Access Denied message.
- sudheerbollaHelper II
Sorry for that,