Median of CPA

Hi there,

I have created two tables to show TOP and BOTTOM performance by audience type based on the CPP (cost per purchase)
The issue is that some audiences appeara in both tables (highlighted in red - see screenshot). What rule can I apply here so it doesnt happen?
My thought was to apply a threshold based on CPP Median BUT it is not possible to calculate the median of a calculated field.

Ayny thoughts?

1 Like

Hi @Patrizia_Buompastor1. You can achieve this using the Top/Bottom filter option.


Hi @Max,

this is exactly what I am using. The first table is using TOP filter and the second is using the bottom. However the rows are often the same

@Patrizia_Buompastor1 may I know how many unique “Audience Segment Name” you have in the table?

Hi @Patrizia_Buompastor1 Agree with where @royyung is going with this. Seems we need to limit how many Top and Bottom audiences we show in each.

Otherwise what do you want to behavior to be? If it is found in Top you dont want to show in Bottom? Theoretically if you have 25 audiences, then you would see 20 in Top and 5 in Bottom - is that what you want?

@Patrizia_Buompastor1 otherwise another way to get a Median of an aggregated value is to use:
percentileOver(CPP, 50)

Then you could compare values above/below that value (and filter on that field if you want):
ifelse(CPP>={percentile calc above}, ‘Top 50’, ‘Bottom 50’)

Note since I have an even number of products I should prob use > and not >=, but >= will work better if odd number. Choose what fits best for your data.