How can carriers capitalize on the convergence of service-oriented architecture (SOA) and business intelligence (BI)? Insurance Networking News asked Mark Gorman, strategic research advisor, insurance, and David West, research area director, insurance, for TowerGroup Inc., Needham, Mass.INN: More and more carriers are seeing the value of SOA for BI (i.e. business intelligence services). Why? What are the drivers?
MG: Business intelligence services represent the convergence of two very important trends in the insurance marketplace. Insurers today are looking for the ability to make better decisions, and to apply those decisions when, where and how they want, in order to gain competitive advantage.
Insurance carriers are increasingly looking to the promise of BI to help them make more precise, more accurate, more consistent and more timely decisions across a wide array of business processes. To do so they are enhancing their capabilities at capturing, analyzing and utilizing data.
Carriers are also looking to apply those decisions at the appropriate time, in the appropriate place and in the appropriate manner to enhance their opportunity for competitive advantage.
The promise of an SOA is that the service is potentially available to any person, process or transaction regardless of the timing, location or type of transaction being executed.
DW: In addition to improvements in consistency, accuracy and timeliness of decisions, a key value proposition of SOA for BI is its financial return, which is realized in two ways. First, service-oriented BI systems carry a lower total cost of ownership. This architecture enables organizations to optimize their use of hardware so that components of BI systems run on the platform that makes the most sense from a cost/performance perspective.
Furthermore, the cost of maintenance is reduced as changes can be made in a central repository rather than on multiple desktops. Changes can be made more rapidly and with less disruption than with older generation systems.
Second, the benefit of greater access to information, timely and accurate decisions, and consistency comes from competitive advantage. Competitive advantage yields greater revenue. The bottom line is that SOA for BI produces benefits on both the revenue and the cost side of the profit equation.
INN: How does SOA for BI help insurers move from an application- to a function-centric mentality?
MG: If you are talking about moving from technology selection and purchase decisions focused on buying applications (i.e. claims systems, billing systems, policy administration systems, etc.) versus buying reusable functionality (i.e. BI solutions), I'm not sure it helps to move from an application-centric to a function-centric mindset as much as it clearly differentiates the two.
Many of the legacy applications don't adapt easily to BI functionality due to data formats, data access difficulties, lack of a robust "what if" environment, etc. Insurers have historically viewed BI functionality as something separate from legacy applications, and as functionality to be applied or accessed separately.
The addition of the SOA paradigm has enhanced the functional view of the technology purchase decision. A service-oriented architecture allows the BI functionality to be developed and sourced internally, or it can be developed and sourced in a vendor environment. This not only increases the flexibility of the insurer in utilization of BI without necessarily needing a particular application to do so, it also increases the ability to develop BI functionality without necessarily having the development resources in-house.
This enables an insurer to easily leverage utilization of BI into more functional processes and real-time transactions, and apply BI capability into more strategic decisions.
DW: If you are talking about business functions (such as claims, marketing, etc.), I don't think insurers are, or should be, moving toward a function-centric mentality. I think many of them are too focused on functional silos currently, and should be migrating toward a customer-focused mentality.
During the past few decades, most companies talked a great deal about being customer-focused, yet their IT departments continued to be organized around the broad functions of the business, such as paying claims, collecting premiums or maintaining policyholder information and all of this was done within product lines, such as personal auto. The change that successful SOA implementers must bring about is a greater alignment between IT and the business. This doesn't mean the business will start producing programmers and systems analysts, rather, IT must improve its understanding of exactly what goes on in each part of the business. They'll need to understand the processes by which new customers are acquired and claims are adjudicated. Only when IT achieves this alignment will they be able to design and implement systems that leverage the full potential of SOA.
INN: We know SOA reduces operational costs and increases efficiencies. As a foundational underpinning, where would the SOA/BI marriage be most effective? Underwriting performance improvements (reducing leakage), customer data that is used first in claims processing but funnels back to enterprise systems, etc.
MG: The answer to this question is truly based on an insurer's situation and unique needs. I'm seeing some insurers who are utilizing BI services for enhancing operational or transactional capabilities. These include marketing applications, such as lead generation or lead assignment, product selection for cross-sell or up-sell opportunities, or channel selection for new product offers. These also include automated risk assessment and pricing precision capabilities during new business or renewal application processing.
For claims transactions, BI services are being used for a variety of functions, such as fraud identification, for adjuster assignment and for reserving requirements. In all of these cases, providing BI on demand in real time is critical to the effectiveness of the organization in transaction processing.
Increasing in importance is the utilization of BI services for performance management. Insurers are utilizing real-time management reporting capabilities for demand management, for resource allocation decisions, for portfolio risk evaluation and for regulatory reporting.
Having access to the information in real time, rather than monthly, quarterly or even weekly, increases organizational responsiveness and decreases organizational risk.
For some insurers, BI services are also being used for more strategic applications like channel optimization, territory evaluation, product pricing precision, market strategy assessment, etc. In these cases, the ability to evaluate alternatives, do "what if" analysis and model various outcomes, is critical. Being able to do so regardless of the location of the resources or the timing of the analysis provides significant benefit.
DW: So far, the benefits of SOA for BI seem to come from reductions in total cost of ownership and more rapid access to data. These benefits can be realized by any company and, ultimately, any competitive advantage gained by first movers will diminish. For longer-term benefits, carriers that incorporate predictive capability into their applications will achieve the greatest success.
There's no doubt that the ability to produce and distribute accurate reports about the company's key performance metrics has become mission critical, however, the instantaneous result of any metric is much less important than the trend that can be identified and predicted using more advanced analytics.
It is good to know where you've been, but it is more important to know where you are going. Integration of predictive analytics into BI applications is the only way to do this.
INN: Explain how SOA creates "right-time" BI on demand, and where is this element most beneficial within an insurance enterprise?
DW: A basic benefit of BI is the access to information that it provides. By giving people the ability to explore information when they need it, rather than only producing standardized reports each week, month or quarter, companies gain the ability to respond more rapidly to a dynamic marketplace. There is a downside to this, though.
Most companies rely on standard reports for the vast majority of their information needs. A company whose people have enough time on their hands to play with information and who spend a significant amount of their time exploring, might not achieve the full benefit of BI systems.
INN: What are some of the things insurers should be looking for when they develop an SOA for BI approach?
MG: Our research suggests that insurers should look first internally at the organizational governance structure to ensure it is sufficient to leverage and reuse the BI services functionality.
Critical is the ability to drive organizational consensus around the utilization of BI services, and a methodology for capturing the organizational knowledge being developed. Also of importance is the determination of the finite outcomes the BI services are to provide.
Feedback we are receiving from the market indicates that data issues are next in importance. This includes the preparation of data for analysis, the validation of the data used in transaction processing, the enhancement of the data used for BI applications and the retrieval of data from external sources.
A number of vendors have emerged with the capability to assist insurance carriers with the preparation of data for analysis.
From a transaction-processing standpoint, insurers should be looking for assistance in integration of the BI services into their transaction-processing environment. This includes determination of the sourcing strategy for BI services (either internal or outsourced) and evaluation of the impact on performance (scale issues, contention issues and recovery issues are examples).
Finally, insurers that indicate support for rapid adoption of the BI services within current or new business processes speeds the realization of benefits from a BI services solution.
Examples include pre-integrated vertical content, analytic resources with industry experience, insurance process redesign experience, organizational change management expertise, services infrastructure support (including security, monitoring and management expertise) and services deployment expertise ("round trip" testing, performance management, backup and recovery services, etc.).
DW: I'll have to echo Mark. The keys to successful SOA for BI are governance, platform and cross-functional communication. Furthermore, the incorporation of predictive analytics into the BI implementation will magnify the ROI.
