The Payoff of Data Warehousing

As companies merge, acquire, and divest from other companies, the faces of these firms continually change shape. This can make any attempt to consolidate data an even more daunting task. But, according to a new study by the Data Warehousing Institute (TDWI), the small number of organizations that have completely consolidated are reaping huge returns, to the tune of $3.34 million average ROI on consolidation projects. TDWI's second report in its 2004 Report Series, "In Search of a Single Version of Truth: Strategies for Consolidating Analytic Silos," surveyed 615 qualified respondents, including corporate IT professionals, independent consultants, and business-sponsor users. Industry experts, including end-user organizations, business intelligence consultants, industry analysts, and report sponsors, were also interviewed. Based on the report's findings organizations have yet to consolidate an average of two data warehouses, six independent data marts, and 28.5 spreadmarts, with only one third of all structures actually consolidated. "I don't think the numbers clearly portray the extent of the problem," says Wayne Eckerson, director of research at TDWI, who wrote the report. "We had 28.5 average number of spreadmarts and in reality, most companies have no idea how many they have. The fact that there are almost five times as many spreadmarts as there are data marts that need to be consolidated I think is [disproportionate]. It could be five to ten times or more." Eckerson also believes that most of the trouble with data consolidation sits with the spreadmarts. "I think a lot of companies understand how to consolidate data marts and data warehouses. "A merger and acquisition forces a company to actually consolidate their data warehouses or data marts," Eckerson says. "But spreadmarts--there are just so many and...they're so hard to get rid of. It's like a...weed or a virus. You just can't get rid of these things. They keep coming back." The report found that only 11 percent of organizations have completed an analytic silos-consolidation project, while 56 percent are in the midst of designing or implementing one, 26 percent are looking into the idea, and 7 percent are not planning one. Eckerson attributes the 11 percent to the ongoing challenge organizations face, because data repeatedly gets fractured by mergers, acquisitions, and separations. "It's a never-ending problem and that's probably why only 11 percent said they've actually finished it." Eckerson says. "In reality it's very difficult to come to a point where you've consolidated everything." For Todd Higginson, manager of Teradata professional services marketing, these numbers are an accurate portrayal of consolidation. "A data mart-consolidation project is very rarely a big bang project where all the markets are consolidated into one specific time frame," Higginson says. "Typically what we see is a phased approach, where an organization will prioritize a subset of data marts for consolidation. You perform that consolidation, you achieve some cost-savings, and then you use that cost-savings to fund the consolidation of additional marts. So it's a cyclical approach that can actually...fund itself by spreading the investment in a consolidation effort over a lengthy period of time." Related articles: Integrating Silos of Data Don't Integrate, InTRAgrate How Can We Build a Foundation for Successful Customer Data Integration?