topic Measure to display returning customers in DAX Commands and Tips

Measure to display returning customers

avanderschilden — Tue, 02 Jul 2019 20:35:31 GMT

Hello,

I have a very simple model;

I want to calculate the distinct number of customers that did a sale in the past, but not in the last 180 days.

As you can see in the following table, customer with id 1 does a second sale on 2019-07-02 and therefore it is the only "returning" customer in my data set;

The result I want to achieve is;

2019-07-02 | 1

In the result, 1 is not the id of the customer, but it means there is only 1 returning customer on this date, which did a sale on this date, does not have a sale in the last 180 days from 2019-07-02, but does have a sale older than 180 days from 2019-07-02.

The following measure is not producing any result;

What is the mistake in my measure?

Many thanks in advance!

Re: Measure to display returning customers

Greg_Deckler — Tue, 02 Jul 2019 20:53:22 GMT

marcorusso wrote a quick measure for something like this: https://community.powerbi.com/t5/Quick-Measures-Gallery/New-and-Returning-Customers/m-p/168297

Re: Measure to display returning customers

avanderschilden — Tue, 02 Jul 2019 20:59:07 GMT

Thanks! I've seen this before but unfortunately it does not solve my problem. In Marco's example, a returning customer has an easier definition than in my example. My measure contains a mistake in one of the filter arguments which is the reason it does not produce the result I want.

Re: Measure to display returning customers

marcorusso — Tue, 02 Jul 2019 22:35:21 GMT

Can you copy the code as a text instead of as a bitmap? It would be simpler to edit showing you possible improvements. There are two issues:
- calculation: you filter Sales[Date_ID] instead of Date[Date]
- performance: you have a very slow implementation of the calculation

Re: Measure to display returning customers

avanderschilden — Thu, 04 Jul 2019 07:18:26 GMT

marcorusso

I ended up with the following measure that produces the result I was looking for. As you mentioned on the sqlbi forum, this measure can be improved performance wise, so I will read the article you send me to see if I can implement the suggested performance improvements.

Reactivated Customers =

VAR CurrentUsers =

VALUES(Sales[customer_id])

VAR UsersLast180Days =

CALCULATETABLE(

VALUES(Sales[customer_id]),

FILTER (

ALL ( 'Dates' ),

Dates[Date] > MIN( Dates[Date] ) -180 && Dates[Date] < MIN( Dates[Date] ) -1))

VAR UserBefore180Days =

CALCULATETABLE(

VALUES(Sales[customer_id]),

FILTER(ALL(Dates),Dates[Date]<MIN(Dates[Date])-180))

RETURN

CALCULATE(

DISTINCTCOUNT(Sales[customer_id]),

FILTER (

ALL ( Sales[customer_id] ),

Sales[customer_id] IN CurrentUsers

&& Sales[customer_id] in UserBefore180Days

&& NOT Sales[customer_id] IN UsersLast180Days))