Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
cancel
Showing results for 
Search instead for 
Did you mean: 

The previous blog New Business Process Analytics features with SolMan 7.1 SP5 gave some overview about all the new features of Business Process Analytics. Within this blog also the new filtering capabilities were mentioned. Meanwhile we got a lot of first customer feedback on the new Business Process Analytics. Besides all the very positive feedback we received, we also heard several times that people are not sure whether the new filtering works correctly or not. At the end everytime it worked correctly, but as mentioned in the previous blog it is important to have an eye on the filter sequence in order to get the expected results! The filter sequence has to be read from left to right.

This blog will give you a practical example to show how different the analysis results can be depending on the filter sequence you have chosen.

Benchmarking without filter

We take the key figure Overdue Outbound Delivery Items as an example (which is also used during the TechEd season 2012), although it is not really of importance which key figure is chosen as the filtering logic is independent of the key figure. In our example we can see that if no filter is set, the number of identified document items is 788 in total. The data can be split per Sales Organization and Delivery Type.

If you visualize the data, then you can easily see that 5 sales organizations are sticking out (1000, 3000, 3020, R100 and R300). The first 3 are mainly related to delivery type LF and the later 2 are only related to delivery type NL (so STO replenishment process and not sales process).

Benchmarking: Filter top5 first

Now we want to set some filters. We want to focus on our top5 sales organizations, i.e. we set the Top N filter for Sales Organization to 5. This reducec the result set from 788 down to 694. In a second step we want to focus only on the sales process and hence filter the Delivery Type on LF and LO. This reduces the final result set to 454.

When we update the graphical chart we only see 3 sales organizations 1000, 3000 and 3200 but not 5, is this correct? Where are the other 2 top sales organizations R100 and R300? Well, everything is correct here. If you remember the un-filtered result set then the sales organizations R100 and R300 were only related to delivery type NL which was filtered-out. These 2 sales organizations are not related to LF and LO, so R100 and R300 are no longer showing up in the final result set.

Benchmarking: Filter top5 last

As was mentioned in the introduction - the filter sequence is of importance and filters have to be read from left to right. So we change the sequence and shift the Delivery Type to the left and the Sales Organization to the right. Now we "formally" set the same filters as before, i.e. we filter on delivery types LF and LO (which reduces the result set from 788 to 481) and then set the TOP N filter for Sales Organization to 5. This top5 filter is set only on the remaining 481 document items and further reduces the result set down to 470. So obviously the filtering that looked "formally" the same, is practically not the same because of the different characteristic sequence. The first filtering example brought 454 records back and the second example shows 470 records.

If we update the graphical chart, then we also see that the result set returns a different picture and this time we can really see 5 sales organizations. We now see the top5 sales organizations only for those delivery items with delivery type LF as well as LO and which are overdue. In contrast the first filtering example showed us how many delivery items of type LF and LO belong to the top5 sales organizations with overdue deliveries.

Conclusion

I hope that this simple example helps to understand that the sequence of filters is essential for what you see as a final result set. Always read filters from left to right. Delete filters that you don't need and add more characteristics for filtering if required. In most use cases you probably want to set a Top N filter for the last characteristic to the right. Avoid setting Top N filters for characteristics that stand to the left - otherwise you will encounter "surprising" situations like in the first filtering example where the chart doesn't show all elements from the Top N filter that you initially set.

More Business Process Analytics details on all (new) features and the new authorization objects can be found on the complete (SMP login required) Setup Guide - Business Process Analytics 7.1 SP5 at http://service.sap.com/bpm > Media Library > Technical Information. 

 

More on SAP TechEd 2012

If you want to learn more on the new Business Process Analytics functions and how they can be used for improving business processes you can visit the following sessions during this year's SAP TechEd sessions:

  • ALM109 - Improving Your Business Processes with Business Process Analytics (1h lecture)
  • ALM267 - A New Generation of Business Process Analytics (2h hands-on)
  • ALM802 - Improving Your Business Processes with Business Process Analytics (1h Q&A)

Further information

Frequently Asked Questions about Business Process Monitoring and Business Process Analytics are answered under http://wiki.sdn.sap.com/wiki/display/SM/FAQ+Business+Process+Monitoring and

http://wiki.sdn.sap.com/wiki/display/SM/FAQ+Business+Process+Analytics respectively.

The following blogs (in chronological order) provide further details about Business Process Monitoring / Analytics functionalities within the SAP Solution Manager.

2 Comments