Ooh-Lala: Embedded Analytics with In-Memory applications

Embedded analytics take a leap these days and that is for a reason. As build-in analytical capabilities into in-memory platforms, the embedded analytics gain in popularity. So what is so interesting about embedded analytics and how do they differ from “regular” business analytics. Is the one better than the other? Time for some deep-dive.

Embedded Analytics: a definition

Defining embedded analytics is like a set of capabilities tied into an application that bring additional awareness, context or analytics that support decision taking for very specific tasks. These capabilities are typical represented in charts, graphs or lists of overviews and tend to represent the core data in real-time. The embedded analytics are often associated with a single type of data. For example, in a purchasing application, the embedded analytics only bring awareness on the purchasing data processed through that application. They don’t provide insights in wider-surrounded applications like logistics, to mention one. This, opposite to business analytics that are designed and used to overlook various sources of data, applications and consolidated views.

Which is better and when to use the one or the other

Naming one the better over the other doesn’t make sense: both serve different purposes. Your organization probably cannot go with solely embedded analytics nor would only business analytics be the ultimate route. The embedded analytics are tailored to their accompanying core application providing insights on very specific tasks or objectives. These could be customer-opportunity follow-up analyses when it regards embedded analytics on a CRM application. Some of the advantages of the embedded analytics are:

  • the user experience is aligned with the experience on its application

  • no extensive set-up required: the embedded analytics applications quite often are fully prepared and come with the accompanying applications

  • the most important analytics are pre-built and ready for instance use by end users; no ETL or batch processing needed. By all means one (1) copy of the data.

  • No latency: occurrences are available for insights as they happen

Referring to the famous Dutch football player Johan Cruijff, we can state that “every disadvantage brings its advantages”; embedded analytics on the contrary do have their limitations:

  • downsize: embedded analytics do often not allow to be embedded in corporate analytics; they are tailored to their core applications and are difficult to access for embedding in corporate business intelligence platforms

  • downsize: given their own dedicated look-and-feel, they do not “fit” in corporate business analytics initiatives which might confuse users

  • downsize: given embedded are deeply embedded in their accompanying core applications, they also “inherit” these applications behaviours and architectures. It could be the case for example that end users want their embedded applications to be available on mobile devices; it the core application does not support mobile, the request cannot be fulfilled easily

  • downsize: extending the reach of embedded analytics is “almost per definition” tailored work

Embedded Analytics and in-memory

Embedded analytics are often applied in in-memory platforms. They excel over there for the fact that they support the closed loop scenario so well. Using the power of in-memory platforms, they allow for real-time operational reporting on specific data sets or applications. Think of order management or retail cash transactions, for which embedded analytics are game changers:

  • real-time insights based on high volume data processing (orders, transactions, traffic, social media etc ..) with immediate results and overviews

  • real-time simulation and if-then-else scenarios available: “if I change supplier and order this list of new items, what will be the effect on shipment lead-times

  • adding embedded predictive analytics adds to the above “…….. and which is the best price I set and what will the effect on customer demand …

SAP is leading in both disciplines: embedded analytics in in-memory platforms and state-of-art business analytics with their SAP BI Suite. The embedded analytics have grown up to maturity especially with SAP S/4HANA or SAP’s Smart Business Cockpits.

Have a look at this link. It gives a very good overview of what can be done with embedded analytics as part of in-memory platforms.

So they embedded analytics are there, and there to stay. Powerful, tailored and instantly ready to use, they have a promising future ahead. Maybe I need to come back to the conclusion of my recent article on what’s up for 2016 and add another trend :-). Am I a big evangelist for the embedded analytics? Hey, you know me by now: I am a business analytics guy. I like corporate aligned analytics running fully in-memory that covers the closed loop to its full extend: Cloud for Analytics is my cup of tea !

tweet this article

Share on Facebook

Highlighted articles