For Zurich Kemper Life, non- traditional sales channels have sparked extensive change in our business. Instead of using insurance agents exclusively to sell products, we now use brokers, banks and even the Web to reach customers. As a result, we recently reorganized the company into four Strategic Business Units (SBUs) focused on specific customers and sales channels. We now need to track sales performance by SBU, not only to help us understand how well products move through the various channels, but also to help the SBUs meet their business plans. This need to use yesterday's business data for today's decisions led to the data warehouse. Zurich Kemper Life, a subsidiary of Zurich Insurance, develops and sells life insurance and annuity products in the United States. Zurich Insurance was founded in Zurich, Switzerland in 1872. Over the past 125 years, the Zurich Group has built an organization that holds high financial ratings and operates subsidiaries and branches in 49 countries, along with cooperative partnerships in some 40 additional markets. The Zurich Group offers customers global coverage in all lines of insurance (property/casualty, life, reinsurance) and asset management services.

We chose an enterprise approach mainly because we were skeptical about the value of data marts. Our concern was that marts would proliferate and we'd have a problem trying to get a single version of the truth. If you take a long-term view of your business, data marts may not really be appropriate. Building enterprise wide at an atomic level still allows us to build data marts later.

The data warehouse is fed with data from 20 source systems. We wanted to standardize the data transformation and extraction process because we were concerned about ongoing maintenance. The bulk of the transformation work is done by Prism Warehouse Executive. Prism programs are produced and tested much more quickly than standard COBOL coding with an investment of fewer staff hours. This has reduced our development costs by approximately 30 percent and has eased ongoing maintenance efforts.

In addition, Prism Warehouse Directory--with its capability to capture meta data--is going to provide a significant long-term benefit when our users start to do ad hoc queries. For this purpose, we are collecting technical meta data and all modeling information into the directory with links to operational meta data. We are establishing a cohesive meta data strategy for the North American operations.

Teradata is our database management system, and we use Business Objects for reporting and display. All the reports and displays that come out of the warehouse are delivered via a Web server, and all users connect to the data warehouse via a Web browser. We have two types of users: senior executives and what we call "information advocates" who perform sales analysis. There are 60 users now, but as we roll out data warehouse access company wide, we'll potentially serve 500 people via the Web.

Being able to use yesterday's business data for today's decisions is of tremendous advantage. We produce a series of monthly reports that present the state of the business. Under the legacy system, it took 16 days after the close of the month to produce them. Today, with the warehouse, we can produce the reports in one day. Another example is a report called "big ticket sales," which before the warehouse was produced weekly--now it's daily. Insurance applications are recorded electronically and get fed into the warehouse during the night; a report is issued the next morning. The CEO reviews this report each morning and can give immediate recognition to each individual sales office.

Not only has the data warehouse enabled us to do reporting by SBU, it has improved the integrity and timeliness of the data we're using to manage the company. We have a diverse number of policy management legacy systems, so data integration and routine periodic reporting has always been difficult. The data warehouse has helped us meet these challenges. It has also provided a foundation for other business applications that significantly benefit the company, such as target marketing, product profitability analysis and customer profiling. Now we're able to create more targeted marketing programs for different sales channels, better understand which products and customers are most profitable and see trends in investment patterns. All of this has enabled us to provide better service to our customers, agents and financial advisors.

There are two points I'd emphasize for anyone starting their first data warehouse project. First, it is imperative to take a long-term view of your data warehouse from an architecture and design standpoint, especially if you're thinking of starting with a small data mart. If you're not guided by an enterprise-wide architecture, you'll have difficulty bringing your marts together to get an enterprise-wide view of your business. This can be a significant trap for developers, and you need to design your environment in a way that allows for enterprise-wide expansion in the future.

The second point is that of all of the tasks involved--from selection of tools right up to production--data sourcing and transformation require the most time. You need to realize that business rules must be defined, data quality testing must occur, etc. Even if you've been extremely thorough and think you understand the requirements of this task, I'd recommend that you double the amount of time you expect to spend.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access