Hi @user033690 

You might be able to solve your problem using a windowing function however I'd recommend restructuring your data. I've done a previous write up regarding customized period over period data models. You can refer to https://dojo.domo.com/t5/Card-Building/show-percent-on-period-over-period-graph/m-p/50540/highlight/... for an in dept description of what needs to happen.

The basics being you'd have a period type column for Current and Last Year (which you'd filter on just these two) which you can then use a beast mode to calculate the total for that period and then subtract the two numbers. I'd highly recommend this method as it gives you a lot more flexibility in your analysis.

If you must use a beast mode you could try something like:

SUM(COUNT(DISTINCT
CASE WHEN `Year`= YEAR(CURRENT_DATE()) AND `Month #` <=`Month # for Today`
THEN `Customer Name`
END
))
OVER ()
-
SUM(COUNT(DISTINCT
CASE WHEN `Year`= YEAR(CURRENT_DATE())-1 AND `Month #` <=`Month # for Today`
THEN `Customer Name`
END
))
OVER ()

I've also replaced 2020 and 2019 with YEAR(CURRENT_DATE()) functions to automatically calculate the current year and last year (hence the -1) so you don't have to worry about changing the year value next year and in future years. (Unless you specifically want 2020 and 2019 then use those values)