# Determine percentage of repeat callers (beast Mode or ETL?)

Options
Contributor

Good morning. We've been trying to create a beast mode to help us determine the percentage of repeat callers and while I can picture in my mind what needs to be done, we've been unable to determine a working calcuation for it.

In the image below, there are 14 distinct companies/phone#'s and a total of 33 calls. Of the 14 distinct companies/phone#'s, a total of 8 have called in more than once, so that would equate to 57% of our callers being  repeat callers.

Can this be accomplished with a beast mode, or is it something we'd need to do in ETL, i.e. collpase columns to aggregate and then create a beast mode at the card level? I'm not a math whiz, but thought it would be easier than this!

If anything is needed to help, please don't hesitate to ask, and any assistance will be greatly appreciated; thanks in advance...

Just a few of the many attempts made are:

• CASE WHEN
(COUNT(DISTINCT`PhoneNumber`) - COUNT(`PhoneNumber`)) = COUNT(DISTINCT `PhoneNumber`) THEN 1 ELSE 0 END / `Callsid`
• CASE WHEN
(COUNT(DISTINCT`PhoneNumber`) - COUNT(`PhoneNumber`)) THEN 1 ELSE 0 END / `Callsid` - 100
• (COUNT(`PhoneNumber`) - COUNT(DISTINCT `PhoneNumber`)) / COUNT(DISTINCT `PhoneNumber`)

1) https://dojo.domo.com/t5/Domo-Dojo-8482-Documentation/Calculate-percentage-of-a-subset-to-the-total-in-Beast-Mode/ta-p/1922 (this would likely work fine for a single bar chart, but unable to get it to work since we want the % of total in a bar, and the total # of calls as a line)

• Contributor
Options

We're working on doing this in an ETL. Thanks

• Contributor
Options

This would be the top level of the card, and we'll use a table to drill down to view the repeat offenders in the hopes of finding ways to minimize the issue.

• Contributor
Options

Any thoughts or suggestions on this? Apologies if the submission was long winded or possibly confusing, trying to simplify the request here:

It's understood COUNT(DISTINCT `PhoneNumber`) will return a count of unique phones #'s, which obviously we'll require, but the problem we're having is determining the count of the unique phone #'s returned that have called in to our call center more than 1x.

From there, it's likely a simple calculation of dividing those distinct phone#'s that called in more than 1x by either total calls/phone #'s to provide us a percentage of repeat callers, but we've been unable to figure out how to get that piece of the equation, i.e. a count of unique phone #'s that have more than one (1) call in.

Thank you...

• Contributor