Another top priority is the replacement of policy administration systems to enable more agile product development. One of the unintended consequences of the growth of data warehousing and business intelligence solutions over the last 25 years is that reporting has become less and less a focus of transaction system vendors. Separate data repositories have been built to provide support for the reporting and analytical needs of organizations. Initially, these solutions provided summarized data to support a high level view of information to executives. Over time, increasingly granular data became available in these repositories and more and more reporting needs are being met in this way. Policy administration systems typically contain much of the detailed policy, premium, and customer data that insurance companies need for many different types of analysis.

(For more ideas on business intelligence, click on this Tech Decisions article.)

Many policy administration vendors also offer claims and billing; thus, providing detailed data across multiple transaction systems which may already be integrated. However, it is not unusual for insurance companies to choose policy administration, claims and billings applications from multiple vendors or even to have multiple policy administration systems in place. Even with integrated data, companies that develop policy administration systems often lack the desire or skills to deploy BI because BI requires different skill sets and design methodologies to implement. The lack of a global data model is a common reason for failure. Multiple unintegrated systems or integrated systems that lack a global data model typically mean that business intelligence solutions have the responsibility to incorporate data from multiple systems as well as third party data providers.

With insurance companies focusing on both BI and policy administration, it makes sense that vendors of policy administration systems (PAS) are looking to expand their capabilities with Business Intelligence offerings. Business Intelligence can add significant additional revenue and make carriers more loyal to the PAS vendor. Making this wealth of data available and providing easy to use analytical capabilities, the options for these vendors are:

? Build their own business intelligence solutions

? Partner with a business intelligence vendor to provide an easy interface between the transaction data and the BI tools.

? Acquire a BI Vendor

? Partner with a BI vendor to build analytics and reporting as an additional insurance application vendor offering

These are not new options. Earlier this decade the major vendors of accounting and finance software looked to buy or build Business Intelligence solutions to increase their business applications capabilities. Hyperion acquired Essbase; SAP built their Business Warehouse then acquired BusinessObjects; PeopleSoft built their Enterprise Performance Management (EPM) modules and were then acquired by Oracle. This seemed to make sense because financial reporting and general ledgers are often seen as the one place in an organization where the broadest set of information is available. However, the usefulness of that data is often limited by two things - the breadth and depth of the data in the ledger, and the impacts of adding more data to the ledger, especially as it impacts the time and effort required to "close the books" . By acquiring BI vendors, these larger vendors hoped to strengthen their portfolio for their installed base. One of the downsides of this trend, however, is that many of the largest BI companies are shifting their focus from their traditional base to support their new owners' proprietary software stack.

The realities of the insurance application market shows that similar trends have already begun with PAS vendors and that movement in this direction is increasing. Policy administration and other insurance transaction systems have the benefit over financial systems of requiring very detailed data for their own processing. This same data can be leveraged for reporting and analytics - the core capabilities of business intelligence. As insurance companies replace policy administration systems, they must either replace or re-source much of their reporting; consequently, many are looking at business intelligence offerings at the same time.

One option for PAS vendors is to build their own BI applications. This may be a suitable option when a carrier has all, or most, of their application systems with the same vendor. The benefit to insurance companies is easier integration of transaction systems with BI applications. From the PAS vendor perspective, the more products that a carrier buys from them, the greater the loyalty to that vendor. However, the skills required to design, build and maintain BI solutions are different than those needed to build application systems. The software to distribute and visualize reports, scorecards, dashboards, etc exists in the marketplace and would be expensive to build and maintain. Consequently, even where a vendor builds the application, they are typically left to determine which BI vendor's tools to support and interface with. Supporting multiple BI vendors can become expensive - each time a BI tool vendor has a new release, the PAS vendor must support the newest version of the product while still maintaining support for the older version. . If they choose to support only one BI tool vendor, they may end up limiting the marketability of their own BI application. An alternative is to build their own tools to distribute and visualize reports, scorecards, etc. This will drive up the cost of developing their BI offering and potentially limit the marketability of their solution. This solution is most advantageous when the PAS vendor has the development staff with the time and skills to build and maintain BI, or when the PAS vendor is building a BI application that is truly unique and cannot be easily met by tools in the market place.

The second option is to partner with a business intelligence vendor to provide an easy interface between the transaction data and the BI tools. There are two typical means of accomplishing this. One is to build an adaptor or connector that transforms the policy administration (or other system) data into a format that the BI vendor data can use. The second is to develop a fast path for connecting the BI tool directly to the transaction data base, assuming that the vendor has a strong global insurance data model that supports all of their insurance applications. Both methods ensure that the insurance company can get access to their data from the BI tool fairly quickly. One downside is that, in either case, any non-standard data or customized fields in the PAS data requires customizations to the interface. If the BI tool is connected directly to the transaction data base, this raises the question of how to handle other source system or third party data. Each vendor must then react to new releases of the other vendor - whether major or minor. An area of concern for the insurance carriers may be the impact of using BI tools for direct, mass query against the transactional data bases. This solution may be advantageous for carriers who have most, if not all, of their insurance applications on software from a single vendor and the implementations have been mostly straightforward involving little to no customizations within the data.

The third option is for the PAS vendor to acquire a BI vendor to provide a quick path to delivering reporting, analytics, and other business intelligence capabilities. This gives the vendor a fast path to providing strong BI to their clients. However, the vendor will need to focus on continuing to support existing BI customers, at least until the expiration of their current contracts. The vendor will also need to focus on providing seamless integration of its own pre-existing transaction systems with the BI tool. These efforts can slow down delivery of new and competitive business capabilities - both BI and PAS - until the integration of the BI tool is complete. One of the risks of this option, is that the insurance application vendors may lose focus on the need to include data other than their own. This may, therefore, be a better solution for providing standard reporting and ad hoc query into the application vendor's own data than it is an enterprise business intelligence solution.

The fourth option is for the PAS vendor to partner with a BI applications vendor. A partnership of this type means that the applications vendor continues to focus on providing strong policy administration and other transaction systems. The BI applications vendor builds a solution targeted specifically for the customers of the insurance applications vendor. Insurance companies should benefit from the best features of each vendor, while each of the vendors improves their offering based on the interaction of the two. Often business intelligence applications fall short when transaction systems do not capture the data needed for reporting or analysis. The BI vendor should provide feedback to the applications vendor when this is the case. PAS vendors can give advance warning to their BI partners when new data (think mold, cyber risk, etc.) is going to start becoming available and driving new analytical needs. The interface or API between the two vendors should be part of the partnership with each vendor responsible for maintaining the integrity of that interface over time and with new releases. The insurance company is likely to have more choice of reporting and query tools as the BI application vendor continues to support the tools for all of its customers. As long as the relationship is not an exclusive relationship, the BI vendor will also need to consider supporting reporting and analytics sourced from a broad range of applications and vendors. This may be the most advantageous solution for insurance companies with multiple policy admin or other transaction system vendors and for those who want to combine internal and third party data.

In recent studies, Novarica has found that two of the major initiatives of insurance companies are policy administration system projects and business intelligence projects. Additionally, they have found that insurance companies are being driven by the desire for increased business capabilities rather than cost reductions. The increasing efforts of insurance application vendors to include Business Intelligence in their suite of offerings, points to an overall maturing of technology in the delivery of a broad set of business capabilities for the industry. Business intelligence, which was once considered a competitive advantage that only the largest carriers could afford, has quickly become a must have capability for both carriers and insurance application vendors.

Aviva Philips, ([email protected]) is the insurance lead and business architect for insurance at Advanced System Designs, developers of BI Express and BI Express Insurance Solution. She has over 20 years experience delivering business intelligence and business performance management solutions to the insurance industry.