Filter Two Fact Tables

[MightyMicrobe]

I have two fact tables that are related to a dimension table, like so:

'Campaign Clicks'  *<--1 'Customers' 1 --> * 'Inbound Calls'

I need to build a measure showing the average number of inbound calls per customer (total calls/total customers) for the customers who called in within 7 days of clicking on the campaign. 

I cannot change the direction of the relationships or create bridging tables, so everything will have to be coded in measure. I tried using TREATAS, but it does not give me the results I need. Propagating filters using TREATAS in DAX - SQLBI

Any tips?

[v-jingzhang]

Hi @MightyMicrobe 

Can you provide some dummy data about the three tables as well as the expected result? You want to count the customers who called in within 7 days of clicking on the campaign, do they only click on the campaign once? How many campaigns are included in 'Campaign Clicks' table? Sample data will help us understand more about what you want. 

Best Regards,
Community Support Team _ Jing

[MightyMicrobe]

Hi @v-jingzhang thanks for picking this up. 

Below is the screenshot of the model part in question, the three tables are linked on Customer ID, but there is no relationship between the two fact tables. 

One customer can click on an unlimited number of campaigns, but for simplicity sake, I'm working with just the latest episode of clicks/calls, so something like MAXX will do. 

[v-jingzhang]

Hi @MightyMicrobe 

It's still not very clear. Can I consider that in 'Campaign Clicks' table, we need to get the latest time of a click for each customer. If a customer made a call/calls after that time, we need to extract the earliest call time and calculate whether the duration between "latest click time" and "earliest call time" is less than 7 days? If it is less than 7 days, then we count this customer and count how many calls he made from 'Inbound Calls' table?

If my above understanding is correct, is it possible that a customer would click campaigns again after he had made a call? If so, only getting the latest click time is not appropriate as it would miss calls happened earlier than that. 

Can you provide some sample data to help me understand the problem better?

Best Regards,

Jing

[MightyMicrobe]

Hi @v-jingzhang below is the data samples from the campaign and call tables, I hope it makes sense. 

The business problem I'm trying to solve is a reduction of incoming calls. If a customer sees an online campaign they should ideally resolve their issue online instead of calling in. But some campaigns drive up call traffic instead. 

So the questions I'm trying to answer are these:

1. How many customers called within 7 days of clicking on a campaign. 
2. How many calls did they make in total (that's a straight row count on the calls table).
3. Which campaign codes are successful (i.e. clicks resulting in the least call traffic). 

I hope it makes sense, thank you for your help. 

CAMPAIGN CLICKS

Customer ID	Campaign ID	Date	Campaign Clicks
12345	ABCDE	2021-12-01	0
16582	ABCDE	2021-12-01	2
12345	AEEEE	2021-12-02	1
12345	ABCDE	2021-12-03	5
17777	EEFFF	2021-12-04	8

INCOMING CALLS

Customer ID	Date	Call Topic	Resolved
12345	2021-12-07	Support	TRUE
99598	2021-12-07	Complaint	TRUE
12345	2021-12-08	Sales	TRUE

[amitchandak]

@MightyMicrobe , Try a measure like

Measure =
var _min = calculate(Min('Campaign Clicks'[Start DAte]), allexcept(Customers[Customers]))
return
calculate(sumx(values(Customers[Customers]) , count('Inbound Calls'[Call ID]), Filter('Inbound Calls', 'Inbound Calls'[Date] >= _min && 'Inbound Calls'[Date] <=_max)))

[MightyMicrobe]

Sorry, this does not work. The last FILTER in the RESULT clause trips it up.