Forum Discussion
Helper IICalculation theory - **bleep** cost per unit
Hi, this is more of a question on the calculation/analytics than anything. . Hoping someone can provide insight or feedback.
I am trying to calculate cumulative product warranty cost per unit (CPU) with the X axis as Months In Service (MIS).
MIS = DateDiff [In service date (ISD), repair date]
The ISD can be different for the products but it is not a dynamic variable. Once a product is in-service the ISD does not change.
It is easy to calculate CPU shown by MIS for products that have a warranty repair.
The challenge and question I have is below.
1) How to take into account the products which never have a warranty repair? They should be included in the denominator as they are part of the population to compute CPU, but how do I define their MIS at any point in time?
2) Should I use a cohort analysis for this type of calculation?
An example with simple fake data is shown below. Notice the ISD’s can vary and the repair date is dynamic in some products. Some products do not have repairs so therefor $0 warranty cost.
I am testing this in excel but will use Power BI when I have a solution.
1 Reply
- lbendlin
Super User
1) Excellent question. On the one hand since the product didn't yet have a warranty event it should not be considered in the pool of devices. On the other hand if you are planning for future potential warranty events then you may want to preallocate funds.
Let's say you have 100 products, and one had a warranty event for $100. That would peg the cost at $1 across the population. Does that make sense for you?
The MIS question is easier to answer. In the absence of a repair date you take the current date ( TODAY() ). There's even a function for that in DAX - COALESCE().
