Embedded Analytics Catching Fire

Drew Robb

Updated · Aug 01, 2013

For decades now, business intelligence has operated as a distinct application. But that is changing rapidly. There is a trend toward embedding BI into applications where users will benefit from analytics, including such enterprise standards as CRM and ERP.

Gartner predicts that within two years, 25 percent of analytics capabilities will be embedded in business applications. That is up from just 5 percent in 2010.

“BI for the masses is the mantra,” said Quaid Saifee, a business intelligence consultant with Wit Inc. “Vendors are embedding analytics in applications to provide more value to the customer.”

Gartner defines embedded analytics as the use of reporting and analytic capabilities in transactional business applications. These capabilities can actually reside outside the application, reusing the analytic infrastructure built by many enterprises. However, to be considered embedded they must be easily accessible from within the application itself.

Users consider such capabilities embedded as long as they do not have to log on to a separate application for analytics. But there are degrees of embedding, which show up in terms of ease of use and providing intelligence at the right place and at the right time.

“The more embedded the analytics, the more useful it is going to be and the more it is going to be used,” said Saifee.

‘Invisible’ Business Intelligence

Gartner analyst Wayne Eckerson noted that embedded BI has been a largely invisible part of the business intelligence market. But for vendors such as Logi Analytics, Jaspersoft, and Pentaho, it already represents a large portion of their annual revenues. He believes embedded BI represents the fulfillment of many of BI’s promises, because it can trigger decisions and actions that drive the business.

Open source business intelligence provider Jaspersoft has embedded its software in over 130,000 production applications, according to Karl Van den Bergh, vice president of Product and Alliances at Jaspersoft. This includes companies such as Red Hat, CA, Verizon, Tata, Groupon, British Telecom, Virgin and the U.S. Navy.

The Virgin Group, for example, founded a not-for-profit unit called Virgin Money Giving to provide a Web-based fundraising app for charities.

“We find open source technology to be quite comparable in terms of performance, capability, feature sets — and the costs are lower,” said Jeremy Walters, director, Systems Development, Virgin Money Giving. “We were able to present all components in a single-sign-on environment. The reporting is branded consistently with our core sites, and the charities don’t know they’re using a different server and different technology to do reporting and analysis.”

Van den Bergh said the main drivers behind the rise of embedded BI include end-user demand for broadly available analytics, application providers wanting to differentiate their products from the competition, and a technology evolution which is making it easier to do so.

The simple fact is that users no longer want to wait for a BI specialist to provide them with insight on organizational matters. Nor do they want to log into a separate analytics app to get the data they need.

Similarly, commercial ISVs and IT groups who are building internal applications or customer-facing websites/portals see embedded BI as a means of market differentiation. For example, Jagex Studios, a multi-player online game provider, created dashboards that monitor in-game behavior so studio leaders can develop future game features with more appeal.

On the technology side, the addition of rich Web APIs based on standards such as REST and JavaScript make it simpler and more affordable to add interactive reports and dashboards to Web applications. Van den Bergh stressed that there are specific levels of embedded intelligence:

  • Level 1: Static reports
  • Level 2: Managed interactive reports (i.e. parameter-driven)
  • Level 3: Interactive dashboards
  • Level 4: Self-service reporting and analysis
  • Level 5: Advanced analytics

As embedded BI becomes more pervasive, vendors are moving upward on the scale to add richer functionality. Some are further along than others. For level 5, Van den Bergh believes you need to have the full spectrum of self-service BI capability within one tool, built upon a Web architecture where the analytics are delivered in a Web browser rather than using desktop software.

“Embedded intelligence will be the next hot trend in the BI market,” said Van den Bergh. “It is only when BI is invisible to the end user (i.e., they don’t realize they are using BI) that BI will become pervasive.”

Embedded Analytics’ Mobile Future

Sascha Schubert, director of Technology Product Marketing at SAS, provided examples of actual integration of BI and analytics into business processes. Take the case of online tariff calculations for car insurance prospects on an insurer’s website or pre-approval of prospects for credit cards or loan offers based on risk score calculations. Everything is arranged automatically, with predictive models being applied on-demand by a push of a button from the application or activated by triggers.

“Organizations are now starting to embed predictive analytics into a variety of operational business processes such as financial transaction systems,” said Schubert. “With real-time fraud prevention systems, analytics are applied to every single credit card transaction to automatically detect anomalies and flag transactions that should be investigated. This saves financial organizations millions of dollars.”

In a manufacturing company, too, analytics might be embedded into process monitoring for preventive maintenance in order to predict risk of process failure or even events that might have a negative impact on product quality. This approach can be extended to intelligent monitoring of all kinds of processes in extreme environments.

For example, off-shore oil platforms use predictive analytics to prevent breakdowns of their production processes by analyzing impacts of combination of events and optimal times for intervention. Based on historical data on actions taken to fix problems of the operational oil production process, the embedded analytics predict the risk of breakdowns, pinpoint the most likely cause and suggest the best action to fix the problem.

The point to appreciate, Schubert added, is that the analytical application does not need to take data out of the operational process to pipe it to the analysis engine. With embedded BI and analytics, intelligence is brought to the process data and not the reverse.

In terms of where embedded analysis is going in the future, Schubert believes it will most likely become much more mobile. As we are beginning to see in high-end vehicles, cars will carry more devices to support the driver and make decisions, such as initiating an emergency brake in case of an impending collision.

Recommendation engines are another example. “While recommendation engines are quite extensively used today in online shops, in the future embedded analytics will go mobile,” said Schubert. “Shoppers will get recommendations on their mobile devices, based on their historical shopping behavior, their preferences and their location.”

Drew Robb is a freelance writer specializing in technology and engineering. Currently living in California, he is originally from Scotland, where he received a degree in geology and geography from the University of Strathclyde. He is the author of Server Disk Management in a Windows Environment (CRC Press).

  • Business Intelligence
  • Financial Services
  • Manufacturing
  • Research
  • Drew Robb
    Drew Robb

    Drew Robb is a writer who has been writing about IT, engineering, and other topics. Originating from Scotland, he currently resides in Florida. Highly skilled in rapid prototyping innovative and reliable systems. He has been an editor and professional writer full-time for more than 20 years. He works as a freelancer at Enterprise Apps Today, CIO Insight and other IT publications. He is also an editor-in chief of an international engineering journal. He enjoys solving data problems and learning abstractions that will allow for better infrastructure.

    More Posts By Drew Robb