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

Get certified in Microsoft Fabric—for free! For a limited time, get a free DP-600 exam voucher to use by the end of 2024. Register now

Reply
Beckham
Advocate II
Advocate II

Weighted Average Revenue Per Unique Customer - Drill Down by Qtr and Year

Hey All,

 

I work for a subscription service and am looking to build a metric that shows weighted average revenue per unique customer over a time period (month, qtr and year). So basically if I only had one customer, and they paid $100 per month all year, then my chart would show $100 each month when drilled to month. When drilled up to quarter it'd say $300 each quarter, and year it'd say $1200. 

 

Here is how my data looks on the Transaction Table. Note that it's a monthly subscription so an ID only shows up once per month. 

 

1.PNG

 

I started with this Measure: 

Avg Rev Per Customer1 =

DIVIDE ( SUM ( Transactions[Revenue] ), COUNTROWS ( Transactions ), 0 )

 

Which works fine when showing the data by month, but breaks when showing by quarter because a unique customer would show up 3 times on the denominator. Essentially this only gives me a monthly average, not dynamic.  Which lead me to this measure:

 

Avg Rev Per Customer2 =

DIVIDE ( SUM ( Transactions[Revenue] ), DISTINCTCOUNT ( Transactions[ID] ), 0 )

 

This again works when drilled to month, but drilled to quarter and year it doesn't adequately weight for customers that weren't around for the full period. Example 35007 was around all of Q2, but 35002 was only around April and May. DISTINCTCOUNT puts them both as equals (1) on the denominator, where 35002 should be 66% of 35007. This is compounded even more at a year level.

 

So then I built this:

Avg Rev Per Customer3 - Qtr Only =
DIVIDE (
      SUM ( Transactions[Revenue] ),
      DIVIDE ( COUNTROWS ( Transactions ), 3, 0 ),
      0
)

 

Which works only for quarters. I can build another for years (replace 3 with 12), but now I have 3 measures which don't drill down, and makes my charts all clunky and gross. 

 

I need help thinking of a better way to measure this while still being able to use drill down.  

 

Thanks in advance. 

6 REPLIES 6
AlexChen
Microsoft Employee
Microsoft Employee

Hi,

 

I create a table like below. I replaced all your revenue to 100 for being easy to see result in chart.

 

1.png

 

In this scenario, your calculation need to be evaluated on row level. Please using SUMX() instead of SUM().

 

avg revenue = CALCULATE(DIVIDE(sumx('Transaction', 'Transaction'[Revenue]), DISTINCTCOUNT('Transaction'[ID]), 0))

 

Please refer to sample below:

 

2.png3.png

 

4.png5.png

 

It’s in Quarter level now.  When I focus on 35007, I can see that its “avg revenue” is 300.  When I focus on 35002, I can see that its “avg revenue” is 200.  This is what you said “where 35002 should be 66% of 35007”.   

 

6.png

 

7.png

 

If I drill down to Month, the result is correct too.  Revenue for 35007 is 700 while Revenue for 35002 is 400.

 

8.png

 

9.png

 

Best Regards

Alex

Hey Alex,

 

Thanks for looking into my situation and spending time on it.

 

Your solution calculates the same result as my 'Avg Rev Per Customer2' calculation, but it doesn't adequately discount the denominator. Also, I'm not looking to slice the data by customer, but to drill down by months, qtrs, and years. 

 

In the example with the revenue replaced with 100 for easy to see results (great suggestion by the way), when drilled by month you should always see an average of 100, by quarter you should always see an average of 300. Your example breaks when drilling to the quarter level.

 

 1.PNG

 

Does that make sense?

 

AlexChen
Microsoft Employee
Microsoft Employee

Hi @Beckham,

 

See data of quarter 2 in blue. The total revenue is 1400,  the distinct count of ID is 5.  Now you will find that the average revenue for quarter 2 should be 280.  The average revenue 300 I mentioned above is the for ID 35007 in quarter 2, not for all.

 

4.png

 

Best Regards

Alex

 

 

Hey Alex,

 

It's not fair to count 35002 as the same as 35003 on the denominator. Our business is subscription based which measures different than traditional point of sale. It's more of a weighted average by month, than a true average. 

 

Distinct count gives 35002 and 35003 both the same value (1), but 35002 wasn't around for the same amount of time as 35003. It was only around for 2/3 of the quarter. Where 35003 was around for 3/3 of the quarter. 

 

Thanks,

 

Clarke

AlexChen
Microsoft Employee
Microsoft Employee

Hi,

 

To get out the data you want, you need to add records for those missing month with "Revenue" blank.

Just like table below. I called it "Transaction2".

 

1.png

 

Now you can create a measure:

 

avgRevenue = CALCULATE(DIVIDE(sumx('Transaction2', 'Transaction2'[Revenue]), COUNT('Transaction2'[Revenue]), 0) * COUNTROWS(Transaction2) / DISTINCTCOUNT(Transaction2[ID]))

 

This is the result:

 

2.png

 

3.png

 

Best Regards

Alex

CahabaData
Memorable Member
Memorable Member

well as you note a little trial & error is involved, and it depends somewhat on the type of visual you prefer as to what its embedded capabilities are in terms of offering periodic values.....  but I would suggest considering adding 2 columns to your transaction table via the query editor:

quarter (values 1 thru 4)

year

 

rather than derive these in the calculations

 

if they exist in the transaction table then they are there for you to use as part of the visual sum/avg/etc features along with ID and revenue....   

 

this might help ......

www.CahabaData.com

Helpful resources

Announcements
November Carousel

Fabric Community Update - November 2024

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

Live Sessions with Fabric DB

Be one of the first to start using Fabric Databases

Starting December 3, join live sessions with database experts and the Fabric product team to learn just how easy it is to get started.

Las Vegas 2025

Join us at the Microsoft Fabric Community Conference

March 31 - April 2, 2025, in Las Vegas, Nevada. Use code MSCUST for a $150 discount! Early Bird pricing ends December 9th.

Nov PBI Update Carousel

Power BI Monthly Update - November 2024

Check out the November 2024 Power BI update to learn about new features.