en
Welcome Guest! To enable all features please Login or Register.
Options
Go to last post Go to first unread
Lotte Bjørn Jensen  
#1 Posted : Tuesday, February 19, 2013 11:36:12 AM(UTC)
Lotte Bjørn Jensen
Rank: Advanced Member

Groups: extranet\Forum
Joined: 4/8/2011(UTC)
Posts: 68

Hi

I would like to count how many days, our customers have bought one of our products the last 180 days.
Lets say we have 50.000 customers.
That makes a matrix of 9.000.000 cells, and Targit doesn't like that.

Here's and example with few customers and 7 days.

First I count number of days

UserPostedImage

Then I group them

UserPostedImage

And the final result

UserPostedImage

But isn't there an easier way to do it? Maybee a calculation in the cube? But how?

Regards
Lotte




Best regards

Lotte Bjørn
Páll Bjarkason  
#2 Posted : Tuesday, February 19, 2013 11:54:50 AM(UTC)
Páll Bjarkason
Rank: Advanced Member

Groups: extranet\Forum
Joined: 1/13/2012(UTC)
Posts: 164

Hej Lotte

You could get new a measure witch count how many days (on how many dates), your customers have bought (at least) one of your products.
In the criteria bar you could choose the scope your analysis to the last 180 days.
Then you could show customer on then y axis against the new measure. Thus you get 50.000 customers times 1 cell = 50.000 cells.
To show the final result you could use the advanced calculation like you have done already or you could get a new dimension which handles the day intervals for you.

Edited by user Tuesday, May 28, 2013 11:52:58 PM(UTC)  | Reason: Not specified

Best regards

Páll, CTS
Lotte Bjørn Jensen  
#3 Posted : Tuesday, February 19, 2013 1:16:23 PM(UTC)
Lotte Bjørn Jensen
Rank: Advanced Member

Groups: extranet\Forum
Joined: 4/8/2011(UTC)
Posts: 68

Hi Páll

Sometimes we would like to see the number of days per product, but sometimes per productcategory.
And I can't just summarize the product-days, because you can buy several products on the same day.

Productcategory could be Oddset and product could be Den Lange, Matchen, Live ....

Regards
Lotte
Best regards

Lotte Bjørn
Páll Bjarkason  
#4 Posted : Wednesday, February 20, 2013 11:03:22 PM(UTC)
Páll Bjarkason
Rank: Advanced Member

Groups: extranet\Forum
Joined: 1/13/2012(UTC)
Posts: 164

Hi Lotte

If there is a strict hierarchy relation between product and productcategory, then I think there is no problem in incorporating the MDX logic in the cube.

The hierarchical relation between product and product category should be:

productcategory A
product 1
product 2
product 3
productcategory B
product 4
product 5
product 6

Then you can choose which level of the hierarchy you present the analysis.

Edited by user Tuesday, May 28, 2013 11:53:09 PM(UTC)  | Reason: Not specified

Best regards

Páll, CTS
Lotte Bjørn Jensen  
#5 Posted : Thursday, February 21, 2013 4:06:29 PM(UTC)
Lotte Bjørn Jensen
Rank: Advanced Member

Groups: extranet\Forum
Joined: 4/8/2011(UTC)
Posts: 68

Hi Páll

Sounds good, except that I'm no expert in MDX. :-(

Do you know, how to do it?

Regards
Lotte
Best regards

Lotte Bjørn
Páll Bjarkason  
#6 Posted : Thursday, February 21, 2013 5:30:09 PM(UTC)
Páll Bjarkason
Rank: Advanced Member

Groups: extranet\Forum
Joined: 1/13/2012(UTC)
Posts: 164

Hi Lotte

In TARGIT BI Suite we are developing in the front end. You probably need some help from a backend developer for assistance you if you are no expert in MDX :)
If you don’t have an in-house back end developer, then it’s probably time to engage an external back-end consultant.

Edited by user Tuesday, May 28, 2013 11:53:54 PM(UTC)  | Reason: Not specified

Best regards

Páll, CTS
Lotte Bjørn Jensen  
#7 Posted : Friday, February 22, 2013 9:42:14 AM(UTC)
Lotte Bjørn Jensen
Rank: Advanced Member

Groups: extranet\Forum
Joined: 4/8/2011(UTC)
Posts: 68

Hi Páll

Unfortunately it's the same person developing front-end and back-end, and that's me. :0)

I'll try with an external consultant.

Thanks for your help.

Regards
Lotte
Best regards

Lotte Bjørn
Aristides Vassis  
#8 Posted : Thursday, May 9, 2013 1:18:06 AM(UTC)
Aristides Vassis
Rank: Member

Groups: extranet\Forum
Joined: 5/8/2013(UTC)
Posts: 6

Hello Lotte,

The most effective way to address this in the back end is to create a distinct count measure of your customer key from the transaction table (sales). When you group this measure by date it will show you how many different customers made a purchase on each day. When you also group by product it will show you how many customers bought the specific product on each date. The same goes with any other dimension attribute that is connected with your transaction table (sales)
Best regards,

Aristides Vassis
BinaryTree / TARGIT Greece

www.binarytree.gr

UserPostedImage
Users browsing this topic
Anonymous
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Notification

Icon
Error