Power BI - Ranking on a Matrix Visual

The measure you've created seems to rank the parent and child categories separately, so when you expand using the + icon symbol, which drills down into the child values from a specific parent level, the rank measure might be taking into account only the child categories and ignoring the parent category. The use of ALLSELECTED could be part of the problem, since it may be only taking into account the level you've drilled down to.

A common workaround to this issue is to include an additional condition in the measure that checks for the level of hierarchy you're at and then calculates the rank accordingly.

Try modifying your measure like this:

Rank =
VAR CurrentLevel =
SWITCH (
TRUE,
ISINSCOPE ( 'Customer GP Data'[New MF Code & Parent Name] ), 1,
ISINSCOPE ( 'Customer GP Data'[Customer Name & Code] ), 2,
BLANK ()
)
VAR ismfFiltered = (CurrentLevel = 1)
VAR isCustFiltered = (CurrentLevel = 2)
RETURN
SWITCH(
TRUE(),
AND(ismfFiltered, NOT(isCustFiltered)),
IF(
COUNTROWS('Customer GP Data') > 0,
RANKX(
FILTER(
ALL('Customer GP Data'[New MF Code & Parent Name]),
[Total Est GP] > 0
),
[Total Est GP],
,
DESC,
Dense
),
BLANK()
),
isCustFiltered,
IF(
COUNTROWS('Customer GP Data') > 0,
RANKX(
FILTER(
ALL('Customer GP Data'[Customer Name & Code]),
[Total Est GP] > 0
),
[Total Est GP],
,
DESC,
Dense
),
BLANK()
),
BLANK()
)

Please note that the ALL function has been used instead of ALLSELECTED. The reason for this is that ALLSELECTED respects the context filters, while ALL ignores any filter that might be affecting the data model, which might cause an issue when ranking.

I hope this helps!