Next Step to BI - Embedded Analytics
SAP Business Intelligence has already made its presence as the analytical tool in almost all places where R3 has been implemented as its transactional system and in other OLTP systems as well, to serve the users with analytical reports. These analytical reports have been used to make decisions after analysing various details. Recent release in BI being NW2004s (BI 7.x) has lot more features that overcame the disadvantages in its previous releases, hence, serving more to the users. So, typical usage involved analysing various aspects of the business using the BI reports and that's it. Dead end? Yeah, you can call that way. Since they are just reports that can only give information and in order to react, BI is not used, rather, the user decision needs to be implemented through a separate system, typically an OLTP system. Let's take an example, where the personnel responsible for approving cancellation of an order need to approve/reject lot of cancellation requests. In order to achieve this, the personnel would ideally have a system to view all requests, but, to make decision, he should have more details, say for example, Is a particular customer cancel the order often? What is the customer's order value etc? As of now, these personnel would open his application to list the requests pending to be approved/rejected, analyse the required details through the relevant BI report and take action accordingly. Remember, BI and the application for approval/rejections are separate systems. What if they both are married together as one application and living as a happy couple? Wouldn't that serve the user more efficiently?
Here comes ‘Embedded Analytics', the next step to Business Intelligence. This sweeps across the entire BI space and is sure to revolutionize and change the way people use BI today. It's not that Embedded Analytics replaces the current BI, but, uses it as a component. This is based on Enterprise Service Oriented Architecture. Tools like Visual Composer, WebDynpro etc. are available for implementing Embedded Analytics. It ideally makes use of Enterprise Services that are available in Composite Application Framework. These Services connect to disparate systems to provide relevant details. In our example, the Services should ideally be connected to the OLAP as well OLTP systems. This way, the Service will be able to provide data from both the systems.
What we finally get is one integrated application where the personnel views all pending requests waiting for approval/rejection and has the capability to view other related details like the details about the customer cancellation history, order value etc. And make decision as to whether approve or reject, all in the same screen, instead of separate systems.
Designing such applications would ideally start from the user perspective and drive backward. This way, it makes the user feel easier to use such applications.
Some of the available connectors to connect to SAP and non-SAP sources are XMLA Connector, ODBO Connector, JDBC Connector, and SAP BI Connector. The challenge in this implementation is to identify the right services and knowing the sources from where it fetches data.
In fact, when I wrote my blog BI in Bank Analyzer, titled ‘BI in Bank Analyzer', I was trying to emphasize on the BI component as not being as a standard BI like SAP BI 3.x or SAP BI 7, but, it's just embedded itself as an analytical component to analyze the risk factors in Banking sector. If you understand this right, you will see that the BI component is embedded very much into the business process of Banking and it doesn't stand alone.
In years to come, we are sure to witness tremendous growth of this technology hence, offering sophistication in terms of implementation and usage. No doubt, SAP Business Intelligence space is getting more excitement. Availability of BI Accelerator makes such applications a great value to the business process.
The core idea of Embedded Analytics being embedding the analytical capability deep into the business process, both go hand in hand complimenting each other. It would be apt to call it as Process Driven Business Intelligence.
