Embedded analytics, the end station for enterprise reporting?

Embedded analytics, the end station for enterprise reporting?

At Cubis, we see an increasing number of customers starting to upgrade their R/3 or other ERP environment to S/4HANA, in order to become a truly intelligent enterprise.

Since Cubis‘ focus is on analytics and not on enterprise resource planning, we should not be too bothered by this, except for the impact such an upgrade might have on the BI platform.

In fact, we are extremely interested in these upgrades, since an integral part of S/4HANA is actually to be found in our area of expertise.

S/4HANA enables its users to work data driven. They can get instant, real-time insights on their live transactional data, without the need for an additional Business Warehouse, nor its complementary ETL and batch loading processes.Embedded analytics, which is the collective name for these features, is clearly a great evolution in the area of SAP business analytics.

It’s perfectly suited for fact-driven process handling and offers great decision support, right inside S/4HANA’s transactional workflows.

Having your reporting data available in real-time and not requiring an extra system for analytics, and thus avoiding huge licensing and maintenance costs by enabling reporting on the operational system in an embedded manner is really awesome. This evolution in the area of analytics leads, not surprisingly, to following question:

Why would I need a data warehouse, when I can do my reporting on my transactional system straight away?

It is completely normal to start asking such questions, but unfortunately the answer is not always what you think or hope it would be.

Embedded analytics is great for getting quick insights on the very latest transactional data. These insights offer the business users immediate transactional decision support. But as always, there are limitations.

Numerous reasons exist for using a separate data warehouse, and these reasons cannot simply be disregarded when one gets the possibility to start with embedded analytics. Above all, S/4HANA is a transactional system. This means it is tuned, sized and maintained for optimal performance when it comes to day-to-day operational work. In addition to this, embedded analytics is available on data which is produced by those daily operations. It is available, but it is not the intended main use.

A good advice is to maintain the best practice rule of using only a maximum of 20% of your system’s capability for analytical purposes, to make sure the transactional operations are not negatively affected by the analytical ones. This also implies that reporting on huge datasets – reporting on the universal journal for various divisions over range of many years for example – should not be done on the S/4HANA system itself.

When companies have a reporting or analytics strategy, its probably not limited to delivering reporting on data coming from one single SAP ERP system.

It’s a perfectly possible scenario that there are lots of various data providers, like CRM systems, HR systems, or even multiple ERP systems from which the data might have to be combined in a harmonized manner.

With embedded analytics, reporting is only possible on the data which resides in the system itself (unless the combining/joining of the data is done in the front end by a BI tool, but performance-wise this is never the best possible solution, depending on the size of your data sets).

With embedded analytics, you only have access to the data in its current state. No historical trend reporting is available, unless you have replicated the data over time to another place, a data warehouse. Also, the use of snapshots, another showpiece of data warehousing, is not possible when reporting in real-time on live data.

So, when you are thinking about your analytics strategy when implementing an S/4HANA solution, please keep in mind what you just read. However reporting is available on S/4HANA in real-time, it might not suffice as the one, single analytics solution for you. If you don’t require any of the data warehouse usages mentioned above, you will be fine, otherwise a data warehouse will still be your way to go. The aspect of having your data available in real-time might get lost, but a product like BW/4HANA can deliver solutions which might get pretty close to a real-time scenario’s performance as well.

In any of the given scenarios, we at Cubis will be able to help and assist you with all of the above. Don’t hesitate to contact us in case you have any question regarding your SAP analytics.

Blog by – Joachim Dewaelheyns
Analytics Consultant at Cubis