Wayne Eckerson published a nice report for TDWI recently called "Best Practices in Operational BI: Converging Analytical and Operational Processes", which you can download from the TDWI site here (free registration required). It's a very interesting report and the summary goes like this (emphasis mine)
"Operational business intelligence (BI) represents a turning point in the evolution of BI. Traditionally, BI has been the province of technically savvy business analysts who spend many hours with sophisticated tools analyzing trends and patterns in large volumes of historical data to improve the effectiveness of strategic and tactical decisions. But operational BI changes this equation: it moves BI out of the back room and embeds it into the fabric of the business, intertwining it with operational processes and applications that drive thousands of daily decisions. This report describes the promise of operational BI and provides suggestions about how to surmount the challenges involved in converging operational and analytical processes."
This is a great summary of the difference between what you might call "traditional" BI and enterprise decision management or EDM. Where BI is somewhat separate, reporting on the business, EDM is woven into the fabric of your business. As Wayne goes on to say
"Using predictive analytics and rules engines embedded within a real-time data delivery environment, companies can streamline processes and automate responses to various types of business events."
Wayne has a nice model of maturity with multiple levels and, in the section on level 4 (Execute Processes), he says that this is "perhaps the most exciting stage of operational BI". This is the top level of his framework (which is a nice way to view growing sophistication in BI) and he talks about such a platform needing business rules and scoring code (I would say deployed predictive analytics, perhaps using PMML) and being action oriented. While he has a slight focus on event-centric decisions at the expense of process-centric ones - me I think there is a need for both - he uses decision services very much the way I would. Perhaps because of this event-focus he does not include the business rules management system vendors (Fair Isaac et al) as platforms for operational BI but obviously I would - I wrote an article on rules as a platform for operational BI here.
Wayne ends by giving some great advice on the kinds of processes/decisions that lend themselves to decision management (I call these micro decisions) and makes a strong recommendation not to maintain a rigid boundary between analytic and operational processes. If you enjoy the report, check out these posts on a webinar he gave on Decision Automation from TDWI and this post about a previous report From BI to Predictive Reporting to Predictive Analytics. You might also enjoy the book Neil Raden and I recently published, Smart (Enough) Systems.
One last thing. Wayne makes the point that operational BI implies "more users, more decisions, and lower latency data" and I think the latency issue was described very succinctly by Richard Hackathorn in an article I referenced in a post on my other blog - Decision Technologies and Active Data Warehousing.
Technorati Tags: analytic application, analytics, automated decision making, BI, BRMS, business intelligence, business rules, data mining, decision management, decision service, decision-centric, event processing, event-driven architecture, micro decisions, predictive analytics, Smart (Enough) Systems, smartenoughsystems, TDWI, Wayne Eckerson
