In-Depth
Drill Down: Business Empowers Data Warehouse
In 1997, Nova Scotia Power executives felt that they did not have an accurate picture of individual customers and the decisions they make, despite regular customer surveys. According to Paul Oliver, who was then a market analyst for Nova Scotia Power, "We didn’t have enough information about individual customers so we could not tell why the market was moving in the way that it was."
The solution, the company believed, lay in a data warehousing project. "We wanted to have one version of the truth about each customer," explains Oliver. What he meant was that the company wanted to be able to assemble all the appropriate information about individual customers from several different data sources into a single view that could be shared by authorized personnel, regardless of their specific area.
Over the past two years, the company has been on a systematic journey that has lead to the first iteration of its data warehouse. Additional functionality will be developed over the next several months. When development is complete, a data warehouse administrator, data warehouse analyst and a business analyst will form a core group to maintain, operate and enhance the data warehouse.
The first step, said Oliver, who was tapped as the business intelligence project manager, was a request for information, which was issued in mid-1997 to a wide range of technology vendors. While power officials believed that a data warehouse could be the solution they needed, Oliver and others first wanted to educate themselves about what a data warehouse could do, and how the technology was being applied in the utility industry.
Following the request for information, the company issued a request for proposal. It selected IBM because of the company’s experience in the utility arena. IBM, in turn, used data warehouse software technology from ShowCase Corp. (Rochester, Minn.). ShowCase provides end-to-end data warehousing solutions on the IBM AS/400. Its signature product is ShowCase STRATEGY.
At that point, a project team was put together with Oliver representing the business perspective and heading the project. While the information technology staff from both the corporate and business unit levels were represented, business users made up most of the team.
The next step was business discovery. The team went to different departments to determine what questions needed to be answered and what questions could not currently be answered. The team then asked where the data that could answer important questions was located and what were the barriers involved in getting to that data. "I wanted to know what I could and could not answer," Oliver said.
Following the discovery process, a business case was made for each analysis. Each case was reviewed by a senior management team that included the project sponsor. Not only were the benefits for each analysis laid out, an effort was made to determine how each analysis fit into the overall strategic direction of the company.
After the business questions were prioritized, phase one of the implementation began. ShowCase promotes a hub-and-spoke approach to data warehouse projects. In this method, a centralized data warehouse is created by a single process of collection, cleansing and organizing all needed data from operational applications. After the centralized data warehouse is developed, it can populate independent data mart databases with subject-specific data. "You have to focus on the business user," maintains Roger Bottum, Vice President of Marketing at ShowCase.
Too many times, people implementing data warehouse projects try to "boil the ocean," Bottum says, by trying to include all the information that users could possibly need. Instead, the initial data warehouse should only include information needed to answer specific questions. But the architecture should be extensible to include the data that will be needed to answer other questions in the future.
That is the methodology Nova Scotia Power used. Currently, the eight to 10 users of the system are focused on very specific questions. The company has been able to better analyze the way meters are read. And it can now more accurately determine the appropriate billing rates for commercial customers. In the past, some commercial customers were being billed at one rate when their usage patterns actually dictated that they should be billed at a different rate. But usage statistics and billing were not easily compared before the data warehouse was created.
Oliver expects that the number of users of the data warehouse will grow to 25 or 30 by the end of the year, and to 100 by next year.
As Oliver sees it, the project’s success relied on three primary factors. First, the project had high-level sponsorship. That meant the data warehouse was designed to meet the needs of the entire organization and not a specific department. Secondly, it was driven by business needs. Although IT was represented on the project team, 75 percent of the direction came from the business users. Finally, the company knew it needed better information about its customers. With deregulation in the future, Nova Scotia Power has to be ready to compete.
About the Author: Elliot King is an Associate Professor of Communications at Loyola College in Maryland. He can be reached at (410) 356-3943, or by e-mail at eking@loyolanet.campuscwix.net.