Business Innovation Homepage > Information Management

At many organizations, various departments and divisions have been busy building data marts — databases that are smaller than data warehouses and generally focused on a particular area or process.

While these data resources are useful for helping managers make decisions, they don’t necessarily benefit organizations as a whole because the information often is not made available outside the domain of the data mart. Also, with a large number of data marts in different parts of the organization, there’s a possibility that some of the information will be inaccurate or out of date. For example, a company might have multiple addresses for the same customer.

Data mart consolidation can enable organizations to optimize their information assets, and get more value out of applications such as business intelligence. According to research conducted this summer by Aberdeen Group, the top challenges companies face in expanding the use of BI applications throughout the enterprise are the ability to integrate data and to address data quality issues. Data mart consolidation can address both those concerns.

Some, but not all, data mart consolidation efforts involve the deployment of an enterprise data warehouse, a large information resource that can be shared throughout an organization. “Depending on the technological capabilities available and the nature of the existing data marts/sources for a given project, consolidation can be achieved with or without a traditional data warehouse approach,” says David Hatch, research director, business intelligence, at Aberdeen.

Regardless of the approach, several benefits can be had from consolidation, Hatch says. One is that it gives organizations new and useful views into data resources. A consolidated view across information that hadn’t been combined in the past often leads to new insights into the business, he says.

“For example, in many hospitals, the ability to combine clinical outcomes data with existing financial and operational business rules has led to [the] discovery of cause/effect relationships,” Hatch says. “The business rules surrounding patient flow and the admitting process behind it directly affect clinical outcomes. Everything from patient satisfaction to the time it takes to administer life-saving treatment is affected.”

By consolidating the data marts involved, management can gain a clearer understanding of the cause-and-effect relationships among data that haven’t been combined and analyzed in the past, Hatch says.

Another benefit is that users can get much speedier access to information. “Consolidated data marts and data warehouse environments should, if they are built properly, provide faster access to critical information for improved decision-making ,” Hatch says. Timeliness of decision-making is a key competitive differentiator in many industries — particularly where response time to customers affects key business metrics, he says.

Finally, consolidation can give organizations improved data quality. “A consolidation process can and should include an application of data quality and integrity standards,” Hatch says. “This should address issues surrounding ‘many versions of the truth,’ where decision-makers are pulling data from disparate systems at different times.” Having a single version of information also can help organizations address regulatory and internal compliance requirements.

Companies face several challenges when consolidating data marts, Hatch says. One is the lack of data format standards. Data marts are often built using different data formats that aren’t easily merged. The development process involved in tackling this problem can often cause budget increases and delays for consolidation projects.

Another challenge is managing data volumes. Consolidation often results in massive data stores that might overtax an existing computing, storage and network infrastructure, Hatch says. “The solution may involve an overhaul that was not budgeted for as part of the project, and may delay or eliminate the ROI that had been predicted earlier in the planning stages,” he says. “A careful consideration of how much data is enough should be made prior to defining the data parameters related to consolidation.”

Click here for more Information Management articles