Real Time BI in the Real World
By Tobin Gilman on Feb 16, 2011
One of my favorite BI offerings from Oracle
is a solution called Oracle Real Time Decisions. Whenever I mention this
product in customer meetings, eyes light up. There are some fascinating
examples of customers using it to up-sell, cross-sell, increase customer
retention, and reduce risk in real time, with off the charts return on
investment. I plan to share some of those stories in a future blog.
In this post however, I want to share some far more common real time analytics use case scenarios that are being addressed with widely deployed Oracle BI and data integration technologies
Not all real time BI applications require continuous learning, predictive modeling, and data mining. Many simply require the ability to integrate, aggregate, and access information that is current (typically within in few minutes or a few seconds). The use cases are infinite. A few I've seen:
· Purchasing agents need to match demand against available inventory
· Manufacturing planners need to monitor current parts and material against scheduled build plans
· Airline agents need to match ticket demand against flight schedules,
· Human resources managers need to track the status of global hiring requisitions against current headcount authorizations...you get the idea.
One way of doing this is to run reports or federated queries directly against transactional systems. That approach can be viable if you only need to access simple data sets on rare occasions. High volume and complex queries can quickly bog down performance of mission critical transactional systems. There is an architecturally simple way of solving the problem, and it's being applied by real companies around the world to solve real needs in real time.
Cbeyond is an Atlanta, GA based provider of voice, data and mobile business applications delivers. They deliver real time information to its call center agents as they are interacting with their customers. The data they need resides in production CRM and other transactional systems, but instead or reporting directly off the those systems, data is first moved to an operational data store (ODS). Rather than running data intensive, time consuming, and performance degrading batch ETL routines to populate the ODS, Cbeyond uses Oracle Golden Gate software to incrementally capture and move only the changed records from log files of the transactional systems every few minutes. There is no impact on transactional system performance, and the information needed by call center representatives is up to date. Oracle Business Intelligence software presents the information to services reps in a rich, visual, and highly interactive format.
Avea is similar to Cbeyond. They are a telecommunications company who integrates billing and customer information in an ODS that is accessed by their call center agents in real time using Oracle Golden Gate and Oracle Business Intelligence. They've taken it a step further by using the ODS to feed a data warehouse. The operational data store provides the current information needed by call center agents during "in flight" customer interactions. The data warehouse is used for more sophisticated analysis of historical data. For maximum performance, both the ODS and data warehouse run on the Oracle Exadata Database Machine.
These are practical illustrations of companies addressing real time reporting and analysis needs using established business intelligence/data warehousing methodologies and tools common to many IT departments. If real time BI could benefit your organization, you may be already be closer than you thought to having the pieces in place to solving the problem.Give us a shout if you are interested in learning more or if you have an interesting use or approach to real-time BI.