DM Review welcomes Joyce Norris-Montanari as a new columnist.

Where did the operational data store (ODS) go? Nowhere! For those of you who have invested a tremendous amount of time and energy into building an ODS, I applaud you! Not only does the ODS have business functionality that benefits the organization, it is used for a lot more than just reporting, such as data integration, data quality checks and feeding other systems good clean data.

An assessment of business needs based on the available data and interviews with business and technical personnel is a great place to start with any data endeavor. I like to interview management separately from the technical people. This ensures that all responses are open and spontaneous. The interviews include questions such as:

  • What is your business process?
  • What information do you need to make your job easier?
  • Where is the data?
  • What stores of data are currently being used?
  • How do you use the data?
  • How do you like your current applications?

The answers to these questions will come in the form of tactical or strategic business answers. Tactical answers involve lack of access to data, no data integration (they use spreadsheets) or software issues (functionality does not meet the business requirements). Strategic answers may be in the form of a business question, such as:

  • What is sales information by customer and product on a monthly basis, and how does it compare with the same time period last year?
  • What customers were the most profitable?
  • Who are my top 10 customers?
  • What products sell best in Colorado?
  • What products sell best together?

Some tactical business questions include:

  • Is there an easier way to report or integrate financial data on a spreadsheet?
  • What is the current customer list?
  • How do I get the most current tactical data?
  • How many times have we interacted with a particular customer this week?

Creating an ODS is not a small project. Integration of corporate data includes data quality. Technologically speaking, now we have master data management (MDM) and customer data integration (CDI) specialization software that is proving very useful in data integration. These products do the work of many extract, transform and load (ETL) programs, using specialized integration routines. This software can integrate customer, product, facilities and human resources from many sources in real time, in some instances. Much like the ODS, MDM and CDI software can create a master store (hub) of enterprise-ready integrated customer and/or product data. Let's all get on the same page with definitions.
MDM and CDI. MDM is the integration of multiple sources of master data (employee, customer, product, cost centers, etc.) from the entire enterprise. CDI specializes in customer data. MDM does not include transactions, just more dimensional data.

Operational data store. The ODS is a store of data that is updated and does not contain history, unlike the data warehouse. It is used to integrate data for tactical purposes, such as reporting and provides functionality that cannot be achieved in source systems. This data includes customer, product and other integrated dimensions required by business as well as transactions.

Data warehouse (DW). The DW is historical, subject-oriented (customer, product, facilities, human resources, etc.), nonvolatile (not updated) and integrated, thus requiring data to be added at various daily or weekly intervals. The DW is the store of data used to propagate data into data marts. In most cases, analytical change affects the data mart, not the DW.

Let me share an example of a client who had some financial issues. This global corporation with multiple subsidiaries needed to create a store of data that integrated financials from multiple places around the world. The reporting requirements were both tactical (especially during the budgeting and accounting periods) and strategic (comparing financial information from one accounting period to another. An architecture that includes the ODS and DW (with data marts) was chosen as the way to go. By bringing in the global financial data multiple times a day, ETL programs were written that handled any integration or quality issue. The ODS was used to reconcile global financial data, which was one of the original requirements. The ODS included cost center information, transactional data, budgeting data, customer data, vendor data, etc. In essence, this company wrote their own MDM. Could they have used MDM software? Yes! However, what they created is flexible and continues to grow and enhance this corporation's assets.

Is MDM considered a source for BI, or is it part of BI? I don't think there is one true answer. Your next question should be, "Well, can I get the new toy even if I have an existing ODS and DW?" The answer is yes, but understand that data flows and ETL programs require change.

The ODS is still a viable means of integration of transactional data for any corporation. MDM and CDI software and methodologies can be part of the enterprise and the BI processes. Just redraw your architecture and the data flows, and move on. 

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access