Insurance & Technology is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Data & Analytics

10:41 AM
Connect Directly
Facebook
Google+
Twitter
RSS
E-Mail
50%
50%

Rethinking Data Warehousing

Data warehousing has a checkered past in insurance, but more mature technologies, clear objectives and corporate support can bring future success.

Q: What recent technology advances/developments provide greater opportunities for success?

A: Schroeck, PwC Consulting: We have seen a few key technology advances that have, and will continue to allow insurance companies to deliver greater value from their DW, including:

Analytical Applications—We are confident that analytical applications represent the next generation of Business Intelligence and DW and have therefore created a dedicated practice area, called iAnalytics (short for Integrated Analytics). We believe that insurance companies will derive increased value from their previous DW investments, as well as accelerate their current DW efforts through the use of these predefined analytics. In addition, the adoption of these applications will change the manner in which insurance firms view their data. Instead of being primarily historical in nature, companies will view their data in a much more predictive fashion.

Scalable Architectures—In order to be successful in DW, insurance companies, by definition, must be able to aggregate and deliver granular information (e.g., policy information at a customer/household level). This requires the ability to capture, store and distribute a tremendous volume of data. Today, the hardware, software, database, network, and related technologies have all advanced to support these volumes.

Content Management—The insurance industry is at the forefront of requiring unstructured information (i.e., documents) to be an integrated part of DW architectures. Therefore, we are now seeing some of the leading firms combining content management with DW in order to deliver both structured and unstructured information throughout the enterprise.

Enterprise Information Portals—Leading insurance companies are expanding the use of portals to provide a ubiquitous vehicle that will allow users throughout the organization to access all of the information (much will be sourced from the DW) to make better decisions and become more productive.

Mobile Technology—While not widely used today, some insurance companies are exploring opportunities to deliver information from the DW to their mobile workforce through the use of wireless technology (PDA's, etc.).

Data Warehousing Technologies—We continue to see tremendous advancement in the core technology necessary to build and operate a effective warehouse, including data cleansing, data extraction/ transformation, data storage, meta-data, data access, data mining, etc. These improvements have been most pronounced in the areas of scalability and functionality.

A: Solomon, CIGNA: The bottom line is cheaper, faster, and bigger hardware. All DW are, by nature, intensive processing and storage consumers. The major relational database management software vendors are all offering increasing capability and scalability with regard to SMP (symmetric multiprocessing) and MPP (massively parallel processing technologies. Being able to "spread" the load across faster processors and across more servers is more imperative than ever, as the DW requirements bar gets set higher and higher. Multi-terabyte warehouses that were rare and a tuning nightmare 5 years ago, are now becoming more commonplace. Although certain software product suites can make DW development a little easier, I believe the principal enabling factor is faster and bigger hardware, accompanied by the technically skilled individuals required to design and construct the application.

A: Sibigtroth, AXA Client Services: The cost for DW has been declining since the early days of DW, mainly because the cost of storage devices has dropped significantly and the advances in storage technology has allowed for storing larger volumes of data. The advances in the ETL(extract, transform and load) software space have also enabled faster movement of larger quantities of data. In addition, newer technologies have also allowed for ""virtual warehousing or data marts"" that eliminate physicalmovement of the data itself, if this makes sense for your company. This approach, however, requires strong data administration skills and support for metadata repository and data quality resources, both people and tools. A ""virtual"" approach assumes that source systems have clean data with data formats that are consistent across systems. If they don't, then it will still require a significant amount of work in the ""virtual"" transformation process.

The rapid development of improved technology in the Business Intelligence and Data Analytics tool area has also made it much easier for an end user to have an improved customer experience when accessing data. Business people are more likely to be comfortable accessing and analyzing data with improved ""customer facing"" GUI's. This, however, further supports the need for strong data administration skills mentioned above. Understanding, standardizing and cleaning the data become even more critical to maintain business user confidence in the data in the warehouse.

It's still ""all about the data""!

A: Yorgensen, The Hartford: The maturation of active DW technology provides the opportunity for analytic applications to more effectively support operational decisions in customer touch-point applications. Integration of the analytic applications with the operational applications will increase the business value derived from the warehouse and lead to greater levels of success.

A: Teklitz, Sybase Business Intelligence: As stated earlier, insurance companies have very large and valuable data stores. Turning these large and valuable data stores into a DW pose formidable logistical challenges and huge data storage costs. Sun and Sybase are delivering scalable, high performance and flexible VLDW solutions with a much lower total cost of ownership that unlock the value of information. Sun and Sybase are doing this through industry leading technology and a best practices methodology called the IQ Multiplex Reference Architecture. Using this Reference Architecture, Sun and Sybase have delivered the world's largest DW that reduced storage requirements by up to 75 percent.

A: Hankison, Xbridge: Three major developments suggest greater success: First, the continued evolution of the extraction and cleansing tools. (The newer tools, for example, provide a simple manual mechanism to identify problem fields in the source data that even the automatic cleanup rules cannot correct.) Second, the advances in the math of finding and correlating data. (Modern relational and object-oriented databases support even the most complex queries much more easily than they used to. This has led to the revolution in "business intelligence.") Finally, the move to a standard relational storage architecture. (A common query language means the day might come where the tool is transparent, and use of the data is common and easy.)

Q: Should data warehousing be approached on an enterprise basis or in a more targeted way?

A: Schroeck, PwC: We believe strongly that you need to do both. This is the "top down/bottom up" approach. In other words, firms should begin with a targeted application of the DW, while establishing an enterprise design and standards that will ensure scalability and flexibility. In order to be successful, these companies then need to leverage "information assets" across their complete value chain.

A: Solomon, CIGNA: The track records show that the big bang enterprise approach is usually not successful. Each case must be considered on its own merits, but the best approach, and the one with the greatest value to the business, is a targeted implementation with an enterprise view. A DW specific to one line of business will have a lower total cost of ownership. Designing and incorporating the flexibility and extensibility to have that targeted "warehouse" implementation fit and scale into the future enterprise warehouse environment will require some additional cost. But almost invariably this cost will be more than recovered, as additional warehouses are built and when the inevitable cross line (or sub-line) of business questions need to be answered and enterprise-wide opportunities need to be pursued.

A: Sibigtroth, AXA: You can do a targeted approach, but it should be part of an enterprise strategy. Many companies adopt the "We'll-start-small-and-build-up" approach without an over-arching strategy, and then run into problems when they try to integrate the pieces. It is extremely important to have an enterprise data model that supports the particular business model. The correct model depends on the particular company, its organization (centralized vs. decentralized), support to the time and resources to develop enterprise architectures, tolerance for risk, and the profile of the company in its adoption of new approaches (bleeding-edge, leading-edge, follower, laggard).

A: Yorgensen, The Hartford: This remains an issue as companies strive to attain a competitive strategic advantage while simultaneously searching for rapid return on investment. The overall business strategy, especially as it relates to customer relationship management, the degree of senior executive commitment and the level of investment available are the primary factors to consider in selecting the most appropriate model.

A: Teklitz, Sybase: If a single-subject data mart can deliver business value that outweighs the cost, then this may be your best approach. The same may be true for a multi-subject DW. Does it generate value you can measure and justify? One way to have your cake and eat it too is to purchase a complete enterprise data model for the insurance industry. Buying an enterprise data model will save you years of design work. You will need to determine which parts of the model will generate the highest value and implement those parts first, adding other components incrementally as needed.

A: Hankison, Xbridge: However data warehousing is approached, the important thing to recognize is that it does not always make business sense to hold certain data in the system, which can easily become a breeding ground for corporate information pack rats. Storing data ""for data's sake"" can quickly explode the size and complexity of the DW, without proper regard for whether the incremental data has business worth. Fortunately, with the enhancements in data-storage systems and the ability to access the enterprise data directly in real time, the enterprise has more choices available at a variety of price points.

Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio

Previous
2 of 2
Next
Register for Insurance & Technology Newsletters
Slideshows
Video