2 Replies Latest reply on Sep 14, 2018 8:29 AM by Maggi Varsho

# Uptake Rate calculation for "Utilization" report

Hi all,

I have a report that I am working on that wants to analyze the utilization of the discounts my company offers. I currently have the data structured in a long- dataset where the discount_name and item_cnt are two columns.

Groupdiscount_namecount
GROUP1auto_pay10
GROUP1new_customer3
GROUP1paper_free5
GROUP2new_customer1
GROUP2paper_free2
GROUP3auto_pay13
GROUP3paper_free11

The problem that I am running into is computing my uptake rate. Calculating the numerator is no problem. But the denominator has been an entirely different story. I have the total item count for each region (lets say GROUP1=10, GROUP2=4, GROUP3=15) calculated separately. So ultimately...

Groupdiscount_nameuptake
GROUP1auto_pay10/10
GROUP1new_customer3/10
GROUP1paper_free5/10
GROUP2new_customer1/4
GROUP2paper_free2/4
GROUP3auto_pay13/15
GROUP3paper_free11/15

What I am curious about is if anyone has built something like this and if so, how they structured their dataset and the calculations they used. I have tried a couple different options but haven't found an efficient solution so I am hoping that there might be some lessons learned from others that I can benefit from myself.

• ###### 1. Re: Uptake Rate calculation for "Utilization" report

Hi, Maggi

Please find my solution attached based on my experience.

Below is the step by step

1st, create a table of total count for each group as provided

2nd, left join this provided count to your data by using the group as key

3rd, create calculation field for utilisation

Hope this helps

ZZ

• ###### 2. Re: Uptake Rate calculation for "Utilization" report

ZZ,

Thanks for the quick response. I had considered that approach initially but was advised against because of performance issues. My "group" is actually a collection of 18 different attributes that I would need to perform the join/merge on. I never actually tested this out (although I am trying to this morning) on our servers but I was lead to believe that this would be particularly troublesome for Tableau. Does anyone have experience performing this kind of join?

Assuming that the above approach does in fact have a significant decrease to performance, I would be curious if anyone else has any other approaches to this problem that wouldn't involve joining two datasets.

Much appreciated,

Maggi