Semi Additive Measure mixed up results after slicer

[-SAK-]

Hello, 

I have a table showing a category in each row, and I created a semi additive measure of a specific calculation for each category. 

All categories are present in 4 different stores, so each store has the same categories.

The semi additive measure calculates perfectly when the store slicer is selected all, but once I choose one store only, the semi addidtive measure returns the average of all three categories in each row. 

The first image is before selecting a specific store, the second is after selecting a specific store.

Thank you

[selimovd]

Hello @-SAK- ,

maybe you want to share the measure, more information or the file with us?

That would make it easier to analyze what could be the reason...

Best regards

Denis

[-SAK-]

Thanky ou @selimovd for your reply. 
The thing is, I'm not being able to create a similar file like the original one I have, the one I'm creating doesn't make the same problem I have in the original, which means I might have some sort of incorrect data modelling.
I also can't share my original file 😕 

Here is the measure: 

Semiadditive % by product =

AVERAGEX ( DISTINCT ( 'DIM Product'[Product Name]), [Percentage] )

Percentage = Divide( Returns, Orders)

In the DIM Product Table, these are the columns: Product ID, Product Name, Store, Product Key
In the Sales Table, these are the columns: Product Key, Returns, Orders, Date