Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!

Reply
peekayza
Frequent Visitor

Checking conversions using patient names

Hi everyone

 

This seems so basic but it's getting the best of me. I'm trying to track conversions on all our marketing contact forms against our internal new patient records. The only field to compare unfortunately is name (which I realise will have a margin of error). I can't seem to link these values relationship wise so I was thinking a lookup could potentially work. Any assistance would be greatly appreciated. Thanks.

 

Screenshot 2022-01-11 122506.png

5 REPLIES 5
DataZoe
Employee
Employee

Hi @peekayza ,

 

In this situation I would create a table that has all the Name values from the other tables (which would be your lookup table or dimension table), de-duplicated. Then you can join that to the other tables (New Patients, SS Contact Form, GDC Contact Form).  

Modeling -> New Table:

Names =
DISTINCT (
UNION (
UNION ( DISTINCT ( Fact1[first_name] ), DISTINCT ( Fact2[first_name] ) ),
DISTINCT ( Fact3[first_name] )
)
)
 
(You can also create the Names table in Power Query too, which is a better option if you have millions of rows).
 
Relationships:
DataZoe_0-1641915670894.png

 

Make sure it's 1:Many from Names to the other tables.

 

Measures:

 

In Fact 1 = COUNTROWS(Fact1)
In Fact 2 = COUNTROWS(Fact2)
In Fact 3 = COUNTROWS(Fact3)
 
In Fact 1 and Fact 2 % of Fact 1 =
DIVIDE (
SUMX ( Names, IF ( [In Fact 1] = 1 && [In Fact 2] = 1, 1, BLANK () ) ),
[In Fact 1]
)
 
(you could even change this to >=1 if you have multiple matches by name)
 
Output:
DataZoe_1-1641915805742.png

 

Here I can see of the 25 people in Fact 1, only 1 is in Fact 2, or 4% of Fact 1.

 

I've attached a PBIX with this set up for you to look at.

 

Hope this helps!

Respectfully,
Zoe Douglas (DataZoe)



Follow me on LinkedIn at https://www.linkedin.com/in/zoedouglas-data
See my reports and blog at https://www.datazoepowerbi.com/

Thank you so much! I will look at this right away and report back soonest. Does it make a difference that my name field is both first and last name combined?

@peekayza No, it doesn't make a difference. You want to try and make it as identifying as possible so it's best to keep the first name and last name together.

Respectfully,
Zoe Douglas (DataZoe)



Follow me on LinkedIn at https://www.linkedin.com/in/zoedouglas-data
See my reports and blog at https://www.datazoepowerbi.com/

Hi @DataZoe 

 

Just checking in if you'd seen my latest update here. Many thanks.

Hi Zoe

 

Thanks again for your help and patience. So I think I took it a slightly different approach to you. I am wanting to check conversions per brand (TDIC, GDC and SS) against the new patients table names. 

 

So as first step I've managed to merge all the GDC names into a single conversions table with names. What I am struggling with is the next step of comparing it against the new patient names. I just can't see any easy way to bring the name in from that table. I tried this step below. Any assistance would be appreciated thanks. 

 

Screenshot 2022-01-20 174502.png

Screenshot 2022-01-20 173536.png

 

Helpful resources

Announcements
April AMA free

Microsoft Fabric AMA Livestream

Join us Tuesday, April 09, 9:00 – 10:00 AM PST for a live, expert-led Q&A session on all things Microsoft Fabric!

March Fabric Community Update

Fabric Community Update - March 2024

Find out what's new and trending in the Fabric Community.