Benchmarking Percentages in Chart Table with Set Analysis?

[110]

So I have two measures  and three dimensions 

Dimensions - 
Brand, ShopID , Week 

Measures -

Sales , Cost

My first metric  Cost Percent     
       Sum(Cost) / Sum (Sales) 

My second metric is Benchmark is the average of Cost Percent by Week / Brand -  
                Avg(Aggr(Sum(Cost) / Sum(Sales), Brand,Week))

My third metric is to compare the difference between Cost Percent and Benchmark - which have been turned into master measures 

            mmCostPercent  - mmBenchMark

This mostly works, however in a table the benchmark doesn't repeat over rows, ideally I'd like to ignore all filtering ( if user filtered on one brand / site etc ) in the benchmark field, other than date/brand.

Where I've used a set analysis with the master measure , it doesn't seem to apply .

   i.e.  mmCostPercent -   ({<ShopID=>} mmBenchmark )

If I was doing this in SQL, I'd just pre-calculate all the numbers in a CTE and recombine, but I'd like to allow users to filter / see it live. 

Any suggestions are most welcome, am hoping this is something fairly standard. 

Edit - I've added a picture to show what I mean -  I'd also like to be able to chart the 'vs Benchmark' metric over ShopID in a Time-Series line chart , to show the Shops relative performance to the benchmark over time.

[brunobertels]

hi 

maybe 

Avg(Aggr(nodistinct Sum(Cost) / Sum(Sales), Brand,Week))