Summary

In our recent benchmark research study “Operational Business Intelligence,” we found that for nearly all organizations it is important to expand business intelligence (BI) to all relevant functions of their operations. The business benefit that most organizations have as the goal of doing this is improved efficiency – if employees are able to spend less time trying to locate and access information, they will be more efficient. However, when organizations attempt to expand BI use or try to enable current users to analyze data from a wider array of sources, they often run into difficulty in providing them with consistent, high-quality data. In response, many organizations are deploying master data management (MDM) as part of information management efforts to improve the quality and consistency of data.

Most initial MDM implementations have focused on creating a reference master data source to improve the consistency of data that is exchanged between operational applications, such as customer relationship management (CRM) and enterprise resource planning (ERP). Recently, however, organizations have begun to pay greater attention to analytic MDM and its requirements, which include providing high-quality data for use with BI tools and improving the consistency of views, dimensions and hierarchies. We advise organizations to determine what analytic MDM, sometimes called master or enterprise dimension management, can do for them and to examine whether products currently available or soon forthcoming can help them improve operational efficiency through more effective use of BI.

View

As organizations grow, both organically and through mergers and acquisitions, they find it increasingly challenging to integrate data from diverse systems for both operational and analytic use. Different systems frequently have their own ways of defining and recording customer interactions, sales transactions and names and versions of products. Additionally, because data associated with these entities moves around and changes over time, the definitions and relationships between pieces of data associated with them also change. Thus, it is not surprising that in our MDM benchmark research, one-third of participants said that users in their organizations spend more time reconciling data than analyzing it; additionally, almost one-fifth – 17 percent – said that no one in their organization is held accountable for the quality of information.

Many organizations have deployed MDM systems for operational use, primarily to improve the distribution, synchronization and exchange of master data. But the standardization and data consistency gained through MDM implementations also can reduce the difficulties many organizations face in managing data warehouses for BI and analytics. A master data source can enable systems to locate data more rapidly and discover dependencies that should be reflected in the result sets. A master source of standard data names and definitions also enables organizations to run more precise extraction, transformation and loading (ETL) procedures, so that all relevant data can be brought into the data warehouse.

In our benchmark research on MDM we found that nearly two-thirds of organizations in the research sample have deployed an enterprise data warehouse. Yet despite this widespread investment in establishing a consistent data resource for BI, 70 percent of organizations in the overall sample reported they continue to experience inconsistent results from their information systems.

We label MDM implementations “operational” or “analytic” depending on the requirements they address. Some MDM tools, such as those from IBM, Initiate Systems and Siperian, have proven useful for operational purposes, particularly for managing customer master data, while Kalido has been used for analytic purposes, such as automating and streamlining ETL procedures and the delivery of dimensional data to users. However, as organizations extend the use of BI tools and analytics to more users, they find it challenging to keep up with the requests for management of views, dimensions and hierarchies, which are critical to reporting and analysis. Users often view data as part of hierarchies or tree structures, such as charts of accounts. Rather than rebuild these for every report or user, many organizations would prefer to maintain existing views, dimensions and hierarchies and manage changes to the information in them. They would also like to offer versions of these elements tailored to different user roles in business functions such as marketing, inventory or finance.

Many organizations therefore seek analytic MDM tools that will enable them to address these challenges. Oracle’s Master Data Management Suite and Data Relationship Management products, based on the former Hyperion/Razza tools, provide some capabilities for managing these elements. Microsoft, which acquired Stratature in 2007, offers EDM+ for enterprise dimension management and has announced plans to provide a more comprehensive MDM offering to accompany the next major releases of Office, SharePoint and SQL Server. Cognos, now part of IBM, has announced plans to introduce a master dimension management tool as part of the next release of the Cognos 8 BI and performance management platform. We expect that more tool providers will address specialized analytic MDM needs as demand grows.

Assessment

In our Operational BI benchmark research, we found that many organizations have ambitious plans to increase the size of their BI user communities in the next two years. Analytic MDM will be important for improving data quality and delivering to these newer users comprehensive BI views of key business data. And as users become more sophisticated in their data analysis, analytic MDM must support the creation, management and sharing of the dimensions and views they want to use. The technology must also support flexible hierarchy management for reporting and other BI activities. Without this kind of analytic MDM, organizations run the risk of recreating with BI and analytic applications the problems of data quality and isolated “silos” of data that have made spreadsheet use a major data management headache.

Ventana Research believes that organizations should include in their MDM plans a roadmap depicting the expansion of analytic MDM to cover dimensions, hierarchies and views. They should evaluate the progress of BI and information management vendors in providing technology over the next 12 months to support these capabilities and should consider participating in beta programs to test the software themselves. We believe an important goal should be to improve the user’s experience through flexible management and self-service rather than to limit functionality – and business benefits – because IT’s data warehouse management and MDM processes are unable to provide adequate levels of service.

About the Author

David Stodder is a Research Director for Ventana Research.