• TechnologyInnovator
• NextInnovator
• EnterpriseInnovator
• SecurityInnovator
• DefenseInnovator
• WirelessInnovator 
• HPinnovator
• EnergyInnovator
• TransportationInnovator
• SMBinnovator (beta)

• NextInnovator(at)Live.com

• Ghost City
• Frontline Sentinel
• Innovation Insights
• WebInno
• Over the River
• Enderle Group
• Security Insights Blog 
• McAfee Audio Parasitics
• Rethinking Security
• Ovum
• iSuppli
• Canalys
• eMarketer 
• CRM Help Desk SW 
• Rethink Research
• The Gadgeteer
• Master the Moment

 

At its recent analyst conference in Boston, IBM announced the launch of IBM Decision Management, a solution that combines several complementary products from its acquisitions to make analytics and predictive capabilities an integral part of decision-making. IBM is promoting analytics as a critical capability in strategic and, more importantly, tactical decision-making. Ovum believes IBM’s step is a reflection of overall business-user sentiment.

IBM’s solution meets a growing business need

Although operational decision-making systems have existed for more than a decade, many systems focus purely on instant gratification and yes/no decisions. However, businesses and customers rarely decide and purchase in a binary fashion, and the failure to identify “gray areas” in decision-making, or take a learning-and-discovery-based approach to selling, can cost businesses dearly in terms of opportunities missed. Better management and the ability to frequently change business rules, or employing a complex event processing (CEP) engine, can help companies to make sound, current, and well-researched decisions. However, most rules/event-based systems do not incorporate a predictive or heuristic element. Traditionally, rules systems evolved from operations management and were constraint-based. Rules were rarely tied in to strategic planning, which was opportunity-based, and did not interact with strategic tools such as predictive analysis. In addition, rules needed to operate at runtime while predictive analysis was never instantaneous. There was therefore a marked disconnect between the practices.

The last few years have seen a substantial change in consumer behavior and a shift in buying patterns, channels, and methods. In this purchasing paradigm, the average attention span and loyalty of online/mobile customers is smaller and much more difficult to sustain. To succeed in volatile markets, capitalize on emerging demand trends, and engage non-sticky customers, organizations must make decisions faster, while still applying the same degree of rigor and research behind every decision. IBM’s Decision Management solution is targeted at these scenarios.

Because the concept of using predictive analytics in rules systems is relatively new, use-cases are slowly emerging. Current relevant use-cases include dynamic discounting at online retail chains, identifying claims fraud at insurers, and identifying the next best action.

Umbrella product reduces complexity for end users

IBM Decision management integrates a host of products under one enterprise platform. These include SPSS Predictive Modeling, Business Rules and Optimization from IBM WebSphere-Ilog, IBM’s homegrown Social Network Analysis and IBM Cognos Consumer Insight, and an upgraded Entity Analytics module that helps spot fraud and reduce duplicates by finding non-obvious relationships. This is a welcome step. Ovum has often found customers baffled by the sheer number of complementary/competing products that IBM offers. While it’s easy for IBM Global Services to integrate different IBM products to suit customer needs, this approach does not make it the simplest portfolio to understand and work with.

Pricing for IBM Decision Management could, however, do with some simplification. Pricing is highly variable, depending on the scope, number of users, and size of the organization, meaning that there isn’t really a base price.

IBM offers loose or tight integration according to specific needs

IBM Decision Management offers two different types of integration between components, adapted to suit customer needs and adoption patterns. For example, some organizations may prefer to use a loose integration using BPM, where BPM creates a workflow between the operational and analytical elements of Decision Management. An example would be an insurer that uses loosely coupled integration to first check policy and enforce compliance, search for quotes and pricing, and examine the deterministic risk. Before the final offer is made to the client, the system could also check the clustering, segmentation, and propensity of the client to buy according to previous behavior and interaction history. These factors could then all be weighted to arrive at a final decision.

Organizations may also prefer a tighter integration between the analytical and operational components. For example, retailers or insurers can select the best product offers by blending analytical propensity with eligibility, while retail banks can use tighter integration to understand and react to complex fraud by blending analytical and deterministic risk. The workflow runs in true SOA fashion, bringing together two services and combining their inputs at runtime to arrive at a decision.

Scoring latency does not affect the operational aspect

Business rules and predictive models frequently arrive at the same decision, but may take a different amount of time to get to the result. Predictive analytics involves data modeling on larger, historical data sets, and usually takes longer than business rules, which react to a specific set of conditions at runtime. For blended decision-making, however, these parts need to act together to ensure that one does not delay the other. IBM specifies that while it may take longer to create and build data models for predictive analytics, the model execution/scoring at runtime is instantaneous and therefore not prone to delay. In cases where millions or even thousands of decisions need to be taken per minute, the vendor recommends using InfoStreams as the analytical infrastructure. For specific workloads, IBM Global Services can suggest alternatives based on target latency, the nature of the decision, the complexity of the model, the need for model training, and so on.

Watson promises to make things even more exciting

Watson provides the final touch to the IBM Decision Management platform by providing the cognitive analytics element. While the exact nature of integration with Watson remains sketchy, IBM has designed a Ready for Watson program to prepare and accelerate client development toward a Watson solution deployment via key factors such as industry context, use-case alignment, data management prerequisites, and infrastructure issues, including interoperability and integration with existing analytics infrastructures. Watson can now be delivered as a service accessible through the cloud and integrated into IBM platforms.

IBM has developed a few use-cases for Watson, most notably in healthcare and financial services, which have the potential to revolutionize the way these practices operate. While widespread enterprise adoption may take a long time, Watson has enough proof-points to show that it is no longer a purely academic exercise. However, it is premature to talk about definitive pricing for Watson because it is still looked at on a project-by-project basis. Over the longer term, IBM’s goal is to ensure that pricing reflects the value being delivered, and provides for global accessibility.