Measures

Hi Khaoula_Farissi - I am not sure I have the formula right, but you can change it. I modeled this properly, so bear with me as I explain.

First, I created a dimension table with just the brands A, B, C, and D. So the model looks like this:

Brand and BrandMonth1 are active, Brand and BrandMonth2 are inactive. That is why one is dotted vs solid.

Then I created 3 measures:

Vague Brand 1 = 
DIVIDE(
    COUNTROWS('Table'),
    COUNTROWS(ALL('Table'))
) 
--------------------------------------------------------------


Vague Brand 2 = 
CALCULATE(
    DIVIDE(
        COUNTROWS('Table'),
        COUNTROWS(ALL('Table'))
    ) , 
    USERELATIONSHIP(Brand[Brand], 'Table'[BrandMonth2])
)
-------------------------------------------------------------


Vague Difference = [Vague Brand 1] - [Vague Brand 2]

Vague Brand 1 counts the rows for each row in the table (so all A rows, then all B rows, etc.) and divides that by all of the rows in the table. That is the %.

Vague Brand 2 does the same thing, but this time it uses the Brand/Brand Month 2 relationship. USERELATIONSHIP turns on that inactive relationship and turns off the Brand/Brand Month 1 for this measure only.

The Vague Difference is simply the differnece between the 2 measures. The results:

Here is my file so you can play with it as desired. If the math is wrong you can fix it or explain it better to me so I can assist.