Register | Log in

Subscribe Now>>
Home News Tech2Tech Features Viewpoints Facts & Fun
Download PDF|Send to Colleague

The value of one

Make better business decisions and save money with data mart consolidation.

by Imad Birouty

Would you feel confident making a decision if you knew that a missing piece of information could change its outcome? Probably not, because an uninformed conclusion might well result in the wrong decision.

The value of one

To take into account the effect of a decision on the entire enterprise, you need complete, reliable data that is integrated from across the company. Consolidating analytical environments into an enterprise data warehouse (EDW) through data mart consolidation not only improves the quality of information available for decision making but also reduces the cost of analytics.

Take this practical business example where decision making is improved with integrated data:

A grocery store wants to reduce its product offerings by eliminating one of two highly similar brands of clam chowder from two different suppliers. The grocer looks at the product line data to determine the per-unit profitability of each brand. Based on this single point of information, Brand A appears to be less profitable and the one that should be dropped. But when the grocer integrates transaction data to see the volume sold of each brand, he finds that Brand A generates more sales, which could ultimately mean greater total profit. So it seems he should not eliminate the brand after all.

When he integrates supplier data, however, he sees that dropping Brand B would decrease the total volume of business with that brand's supplier so much that the discount level for all purchases from that supplier would be reduced. This action would yield a dollar loss to the grocer that is much greater than the difference in profitability between the two brands. So keeping Brand B—not Brand A—would net the better outcome.

Clearly, leveraging data from across the enterprise enables the smartest decisions.

Figure: Superior business value of integrated detail data
The business value compounds exponentially while the enterprise data warehouse investment increases incrementally.

Business benefit of integration
The figure shows how multiple data sources can come together in an EDW to provide integrated, cross-functional information that can't be derived from disparate data marts. Why is this the case?

Data marts are, by definition, analytical environments focused on a single subject area. As such, they are usually owned, managed and operated at a departmental level to cater to the specific needs of that subject area and department. This limited scope by itself is not bad; data marts provide insight, analysis and value to their users. However, because these environments are set up at a departmental level, each has data definitions based on its departmental language rather than enterprise-wide definitions of the data. As additional departments find the need for analytics, they will build data marts with their own definitions and terminology.

The resulting problem is that the information that comes out of one departmental data mart is difficult—if not impossible—to combine with data that comes out of another. Monumental manual effort is required to reconcile and compare information across the enterprise. This is in contrast to an integrated EDW that, by definition, encompasses company-wide, agreed-upon definitions of data and stores a company's multiple subject areas in one relational data model. The result is the ability to join data (apples to apples) across multiple subject areas to answer questions that expand beyond a single subject area, as well as support decisions where the relationship between the subject areas is key to the outcome of the decision. Thus, information from an EDW is more valuable than that from data marts.

Added benefits of consolidation

Consolidating data marts into an enterprise data warehouse (EDW) helps drive better business decisions as well as save money, but that's just the start. Look deeper and you'll find additional benefits:
> Compliance. Having integrated data in one place gives you the ability to control it, helping you achieve many of the requirements for regulatory compliance including access and use of sensitive financial data, accuracy in reporting financial results, access and use of personal information, data security, encryption of sensitive data and disaster recovery planning.
> Going green. Maintaining separate data marts comes at a cost to the environment. Consolidating power-hungry servers, which are frequently under-utilized, will reduce electricity consumption as well as the amount of heat produced, which in turn reduces the amount of cooling required. It will also lessen the amount of floor space needed, lighting required to illuminate the area, paper utilized, etc.
> Master data management (MDM). Every company faces the challenge of defining and maintaining master records such as determining consistent identification of a data element like "customer name." This is difficult because records tend to be dispersed and inconsistent across operational systems. However, because the process of establishing an EDW includes the cleansing and standardizing of data to create common enterprise-wide identifications, the hard part of MDM is already done.


Value versus cost
If information from an EDW is indeed more valuable than information from data marts, does this higher value have a corresponding higher cost? The answer is no; it's quite the opposite. Contrary to common thought that higher value will cost more, it costs a company more to maintain disparate data marts than it does to maintain an EDW. Data marts provide less value at a higher cost.

Contributing to this reality is the fact that the efforts and cost drivers to establish an analytical environment for your first application are the same for both a data mart and an EDW. Some costs, such as hardware, software and networking, are easily identified and tangible. Labor costs are also required for planning, establishing, operating and maintaining the environments. This is the largest portion of the cost, because while hardware is a one-time expenditure, labor is a recurring expense. Labor tends to be more difficult to identify, especially because the data marts may be maintained by individuals within a functional department, not IT.

However, data mart environments are distinct. Each time a new mart is created, all of the costs are reborn. None of the initial planning, design, setup of the hardware and database, creation of the data model, data definitions, users and permissions or data loads can be leveraged from previous marts in a new data mart. Additionally, once established, all of the separate data marts must be maintained individually.

Also, because each data mart is distinct, common data that is loaded and stored in one mart must be reloaded and stored again in the next data mart, driving up costs for storage. The higher costs of operating data marts are driven by the redundancy in products, data storage and labor. Although these are the bulk of the costs, one more expense must be factored in: the amount of money required to reconcile information that comes from different data marts that have different data definitions. This is a significant and real task that must be recognized.

Complete information
Having complete information helps drive better business decisions. You can obtain complete information by integrating data from several sources across your organization into an EDW. For organizations with data marts, a data mart consolidation program can yield this benefit. The result is your ability to not only support better business decisions but save money too. T

Choosing wisely

Deciding to pursue data mart consolidation can be a powerful change for your business in terms of new business insights, cost reductions and reinvestment of the money you saved. Think of consolidation as a low-risk way to realize savings that can then be invested on projects that will yield higher return on investment (ROI). This holistic view reminds us that while budget dollars may be fixed, how you invest them can make a world of difference.

Once you decide that data mart consolidation is right for you, you need to start planning. Your success depends on your choice of vendor and its ability to help you execute. When selecting a vendor to partner with, ask tough questions, including:
> Do you follow a documented methodology?
> Do you have a business impact model to calculate and document the financial changes I can expect?
> Can you calculate a payback period?
> How do you determine which marts to consolidate?
> How many customers have you consolidated data marts for?
> How many total data marts have you consolidated?
> Can you provide a list of customers to call—in my industry and other industries?
> Have any independent analysts reviewed and rated your ability to execute?
> Do you have tools specifically tailored to automate the consolidation process?
> What hardware and software environments will you use? Are these proven capable for data mart consolidation?
> What consulting services do you provide?

Asking these questions is just the beginning. Data mart consolidation requires commitment from you and your vendor. If a vendor wants your business, make sure that company has proven ability to deliver on its claims and will be a true partner.


Imad Birouty is program marketing manager for Teradata's high-availability solutions and data mart consolidation program.

Teradata Magazine-March 2008

More Features

Related Links

Reference Library

Get complete access to Teradata articles and white papers specific to your area of interest by selecting a category below. Reference Library
Search our library:

Protegrity | About Us | Contact Us | Media Kit | Subscribe | Privacy/Legal | RSS
Copyright © 2008 Teradata Corporation. All rights reserved.