Key influencers analyzing measure

[Sasha]

Hi,

I'm trying to implement the new feature of analyzing measure in the key influencers visualization.

Suppose I have a table of sales with columns Region, Month, Units Sold, Total Revenue.

I would like to analyze Revenue/Unit.

I defined a measure of SUM(Total Revenue)/SUM(Units Sold)

Put it in the visualization.

The thing is that every segment it finds is compared to the overall Revenue/TEU calculated on a row level.

Example: Sales of 100 units with total Revenue of $100 and another row of 200 units with total Revenue of $100

For the visualization the benchmark is $0.75 per unit (100/100 + 200/100)/2

Obviously I want it to be compared to $0.67 ((100+100)/(200+100))

Is there a way to make the visualization consider the measure as a measure?

Thanks

[v-xuding-msft]

Hi @Sasha ,

Can you please share a dummy file? If you can't, please share some sample data, your expected results and screenshots that will make us help you quickly and accurately.

How to Get Your Question Answered Quickly

Best Regards,

Xue Ding

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Sasha]

Hi @v-xuding-msft 

Here's a sample of my data:

I add a measure: 

Revenue/Unit = DIVIDE(SUM(Sales[Total Revenue]),SUM(Sales[Units Sold]))

I then want to analyze this measure all columns:

Without filtering total revenue is $594100 and total units sold 144661

So Revenue/Unit = 594100/144661 = $4.11

However, the visualization compares the segments to other average.

What it basically does - it calculates the measure on a row level and then averages all results resulting in $4.39.

I would like it to compare to $4.11 - is it possible (otherwise - why am I given the option to analyze by measure?)

[v-xuding-msft]

Hi @Sasha ,

How about adding CALCULATE per sum values? Did you try this? Did it work?

Revenue/Unit = DIVIDE(CALCULATE(SUM(Sales[Total Revenue])),CALCULATE(SUM(Sales[Units Sold])))

Due to the sample data is a part of your data, I'm not sure the result I tested is what you want. You could try this firstly. If there is anything wrong, please point it.

Best Regards,

Xue Ding

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Sasha]

@v-xuding-msft  Hi,

it didn't change anything 😞

[v-xuding-msft]

Hi @Sasha ,
Please try the following formula.

Revenue/Unit = 
var a = CALCULATE(SUM(Sales[Total Revenue]),ALL(Sales))
var b = CALCULATE(SUM(Sales[Units Sold]),ALL(Sales))
return
DIVIDE(a,b)

Best Regards,

Xue Ding

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[Sasha]

@v-xuding-msft 

Thank you!

However, this will result in a constant number of Revenue/Unit - meaning for every filter it will return the grand average - so there won't be any difference between a particulr group average and the total average.

By the way - in the file you uploaded - if I remove Months (or other columns) from "explain by" and then return it back the results change - it seems like the visualization is not stable.