Measure to identify new customers within date range

[losamfr17]

Hi,

I have this measure to identify all new users (i.e. "Uploaders") starting from Day >= 10/24/21.
I want to treat all user activity as totally new (regardless of history) starting 10/24/21.

New Uploaders (Starting wk39):=
(CALCULATE( DISTINCTCOUNT([customer_id]) , 
    FILTER( VALUES( Uploads[customer_id] ) , 
        CALCULATE( COUNTROWS( Uploads ) , 
            FILTER( ALLSELECTED( 'Retail Calendar' ) , [Day (454)] < MIN( 'Retail Calendar'[Day (454)] ) && 
                                                                               [Day (454)] >= DATE( 2021, 10, 24) ) ) = 0 ) ,
             FILTER( 'Retail Calendar', [Day (454)] >= DATE( 2021, 10, 24 ) ) )

The above measure seems to work well.
The issue lies when I use it in CALCULATE and apply another time FILTER which seems to cancel the filter above.
My goal is to track progress of new uploaders at a Week, Month, Season, and Year level starting from 10/24/21 (NRF Retail 454 calendar).
So for example, my week measure is:

Week New Uploaders:=
CALCULATE( [New Uploaders (Starting wk39)] , FILTER('Retail Calendar', [Previous Week (454)] = "LAST_WTD" ) )

Problem: see image. I should be getting 254 for week 40, but instead I get 263. I can't make sense of this.

Any thoughts please? Do you need to see the data model?