CASE STUDY - AmeriCredit Creates a Single Source of Reference for Key Information Assets
Commentary by Nikitas Gogos and Aiman Gurji, Data Architects, AmeriCredit
As a leading independent automobile finance company, AmeriCredit Corp. relies on information technology to support its decision-making capabilities for credit risk management. Information is core to AmeriCredit’s business, and it is critical that data is consistent, consolidated, and easily available to all employees for accurate and efficient decision making.
To create a single, consistent source of information for key information assets such as credit, payments, and collections, the data architecture team created a data warehouse that consolidated the dispersed data sources into a single source for reporting. This integration posed several challenges, according to data architects Nikitas Gogos and Aiman Gurji.
“Our biggest business benefit has been the documenting and publishing of metadata to our end users.” —Nikitas Gogos, Data Architect, AmeriCredit
The first challenge was technical integration across multiple database platforms. In addition to platform migrations, data types, and formatting, the data needed to be consistent in the target warehouse. With hundreds of database tables, this was no simple task.
The second challenge was organizational integration and information sharing. Metadata had to be documented and standardized across the organization. Once these definitions were created, a variety of teams needed access to the information—including business intelligence analysts, database administrators (DBAs), and data architects. Publicizing information across the organization and creating consistency was key to the success of the project.
Consolidating Information
AmeriCredit created a single view of information using graphical data models. Through reverse engineering, the company created an inventory of the source and target systems of the warehouse to have a comprehensive view of the information landscape. These models were then stored in a central model repository so that information could easily be shared, standardized, and reused.
Once an inventory of the data was created, additional business information such as definitions and code values were added. To show the right amount of information to the right audience, business information was stored in the logical data model, and technical details were managed in the physical layer. To further manage the volume of information, subject areas were used to organize information into smaller, more understandable groupings such as applications, account services, and funding.
Getting the Word Out
After creating this valuable store of information, it was important to share it with the extended team. With a data warehouse team of more than 300 members, and business intelligence analysts using the information to build reports, it was necessary to develop a publishing mechanism that was both intuitive and easy to implement. HTML reports were published on the Web so that business definitions, code values, and column details were easily accessible by end users.
Faster Access to More Reliable Information
Making information consistent and accessible saved AmeriCredit time and money as more and more users used the Web-based reports to find the information they needed. “Our biggest business benefit has been the documenting and publishing of metadata to our end users,” said Gogos.
With information published and easy to find, the data architects spent less time answering questions and were able to focus on their primary tasks. Questions that did come up were more detailed, and as a result, the data architects had more productive conversations with team members. Training of new employees was also streamlined, since project information was consolidated in a central place. The models and reports were able to provide new team members with the information they needed to get up and running quickly.
For a free white paper on this topic from CA ERwin Modeling, click here and choose the title “The Benefits of Data Modeling in Business Intelligence.” For more free white papers, click here.