The Oracle Real Time Decisions (RTD) roadmap began with a quick review of Oracle RTD (see my earlier post on the role of RTD in e-commerce) – an analytical decision service that sits between the operational and analytic applications of an organization so that operational business processes can be analytically enhanced. Crucially it combines prescriptive decision logic (business rules) with adaptive decision logic – it is the system of record for decisions (not for customer data, not for content, not for CRM processes, just for decisions). It is a J2EE/web service deployment.
RTD is part of Oracle’s Business Intelligence stack within the Fusion Middleware offerings. BI, of course, is focused primarily on improving the decisions an enterprise makes. But decisions range from strategic to operational (as Neil and I described in Smart (Enough) Systems and that I discussed in this post on the need to focus on decisions). Some BI technologies are good at Strategic decision making, some at management decisions in the middle and some on front line operational decisions. RTD is aimed clearly at these high volume, repeatable, automated operational decisions. These decisions get embedded in operational processes for what IDC call Intelligent Process Automation.
Oracle feels that with OBIEE,and Hyperion, with its analytical applications, its database BI Technology as well as RTD it has a unique stack. In particular its ability to drive analytical decision making into execution. 11g brought more of these features together and made more of them easier to consume in applications to provide a single, unified set of BI applications.
Oracle RTD Product direction has two key elements:
- Tight integration with Fusion Middleware
RTD uses the Middleware stack whenever possible and exposes RTD deployments through the Middleware stack. This allows for easier integration to other elements such as Content Management and BPM. RTD uses common metadata foundation, security and access control, clustering, systems management etc.
- Integration of RTD into both operational and analytical application
The objective is that both analytical applications and operational applications/processes should be enhanced with RTD. Use case examples include the BPM tools are using RTD to do process analysis and optimization (RTD acting directly on process data) and to allow RTD decision services to be embedded into operational processes as well as content management and RTD together deliver personalized pages, emails, content etc
Besides these middleware-centric uses, new Fusion applications are being delivered that use RTD. Fusion CRM, for instance, uses OBIEE, RTD, Oracle Data Mining and Essbase to manage territory management in the new CRM applications. RTD is used to help sales teams by recommending sales approaches/offers based, in part, on ODM predictions as well as on both RTD rules and RTD adaptive models. OBIEE dashboards and reporting wrap this up for analysis and planing while Essbase allows for ongoing simulation. All Fusion objects have a web service API, an analytic API and a batch import/export interface. As a result one can use any Fusion or custom object as the basis for analytic models, feeding all instances of the object into an analytic environment, and have RTD or rules execute against them in a production setting. This means that Oracle, ISVs and customers can extend and customize Fusion applications using analytic and rules-based decisions instead of writing code. Nice to see this and to see some Fusion applications that take advantage of it.
Siebel Marketing is also integrated with RTD to deliver recommendations for inbound call marketing while many existing customer experience applications are being glued together with RTD – it is acting as “decision glue” for the customer experience to bring these together. Another (internal) use case is around using RTD to do lead scoring in Oracle’s own internal sales process – integrated with Siebel CRM, RTD scores leads as most likely to close based on rules and adaptive models.In each case RTD is adding value by providing specific, personalized recommendations for action at a transactional level for use in operational business processes.
Oracle’s interesting spin on enterprise-wide decision management is that they see decisioning technology (rules and analytics) being combined with more traditional BI as well as with CRM and content management. Clearly a lot of decision management is in customer experience/and some content/text analytics would seem in order too but the focus on content management integration is interesting and reflects their focus on e-commerce. At this time they see event processing (Oracle CEP) having a somewhat arms-length relationship with RTD but seeing good use cases for using them together. Interestingly they have worked on integration Essbase and the simulation with RTD so that one could simulate the roll-up of different offers while taking account of the impact of changing offers. This would allow what I like to call the Aggregation of Simulations rather than the Simulation of Aggregation (which is all most BI-centric simulations handle). This back end work is already in use but a front-end interface for managing it is required.