Forum Discussion

Tazmastablasta's avatar
Tazmastablasta
Icon for Helper I rankHelper I
6 years ago
Solved

NPS last score per period

Hi All    I have a requirement to create a measure that calculates NPS based on the last score per period per customer (one vote per customer  either per month or per quarter), so if the line chart...
Need Help
amitchandak's avatar
amitchandak
Icon for Super User rankSuper User
6 years ago

Try Tried, but not able to match logic. Please see this screen shot. Can you explain by checking lines, maybe in excel,

What is considered as the promoter for Aug, Sep, and Q3. And final calculation

 

Tazmastablasta's avatar
Tazmastablasta
Icon for Helper I rankHelper I
6 years ago

Hi amitchandak 

 

I have simplified the example 

Let's say in October we have this:

 

The condition is to only consider the last score per ID , so in this case it will 3 promoters.

Now both my mine and v-juanli-msft  solutions are counting last Promoter,Detractor or Passive depending on a label,

which in the October example gives the total of 7 ( 3 Promoters ,2 passive , 2 Detractors) which is wrong 

what I am trying to achieve is count of labels which where last scores.

 

 

In terms of the Quarter  the result should be, 3 Detractors, 2 Passive,  3 Promoters.

 

Now in both examples ( mine and v-juanli-msft ) the distinct count of id with any label filters works well

NumOfRespon = 
SUMX(LASTDATE(VALUES('Calendar'[Date])),
CALCULATE (DISTINCTCOUNT(  'NPS'[ID] ),
    FILTER ( 'NPS', 'NPS'[Score] <> BLANK () || 'NPS'[Score] = 0 )
))

 

and return the desired value but  when we try to count the labels we have problems. 

 

Here is the link another version of PBIX 

https://www.dropbox.com/s/vbm6pndnengdgox/NPS%20last%20score.pbix?dl=0