The data warehouse development effort includes source system analysis as one of its steps. During the development process, the contents of the source system are analyzed and appropriate transformations are designed and programmed to migrate the data into the data warehouse. But what happens after the data warehouse is built?
After the data warehouse is in operation, changes in the operational environment can have a significant impact on the quality of the data which reaches the decision makers. There can be changes to the business processes which use the transactional systems that can impact the data warehouse even if the system itself does not change. The data stewards play a very important role in ensuring that the business changes are transmitted to the data warehouse team.
One area which may seem harmless is the use of classifications within the transactional system. Customers may be classified, for example, to indicate if they are either large businesses (companies with over $1,000,000,000 in revenues) or small businesses (others), with the classifications based on the organizational hierarchy within the sales department. To illustrate how the data warehouse quality may be jeopardized by a classification change, consider the two reports which could be produced from the data warehouse on consecutive months. (See Figures 1 and 2.)
|Data warehousing Report: 1-31-96||Annualized Sales||Number of Customers|
|Data warehousing Report: 2-28-96||Annualized Sales||Number of Customers|
In the marketing department, meanwhile, the vice president is jumping for glee! In late January, the marketing department implemented a new promotional campaign aimed at large businesses. When the vice president saw the report from the data warehouse, she was amazed at the success of the campaign. (The slight dip in sales to the small businesses was not considered significant.) The vice president, therefore, decided to extend the campaign another month, and the following month was surprised by the report (see Figure 3) showing virtually no change.
|Data warehousing Report: 3-31-96||Annualized Sales||Number of Customers|
In my December 1997 column, data stewardship was described within the context of the Zachman Framework. In that article, "Data Stewardship Using the Zachman Framework," a stewardship council to provide an enterprise view and a business data steward to be responsible for each subject area was recommended. With a data steward responsible for the customer data, the sales department would be aware that any change to the data requires the approval of the data steward. Even prior to the actual reorganization, a sales department representative would have met with the data steward, and a decision would be reached concerning the classification of the data. Options include reclassifying the customers or providing another means to accomplish the sales department's objective of getting reports aligned with its new organizational structure.
Once the decision for the classification scheme is reached, the data steward would be responsible to ensure that processes are implemented to reflect the changes in all systems--not just the data warehouse. Further, the data steward would assume responsibility to ensure that the change was effectively communicated.
One way of documenting the change so that all data warehouse users can be aware of it is through the meta data. The meta data can contain the domains for each of the classification schemes. By updating the domain values and reflecting them in the meta data, users of the warehouse can become aware of the change. Many users do not access the meta data unless they are exploring new territory.1 These users, therefore,would not be aware of the definition changes and could misinterpret the data warehouse reports.
To ensure that all users are aware of the changes, some proactive steps are needed. Examples of proactive steps include e-mail notifications, messages provided to data warehouse users when they log on or when they execute certain requests, and articles within company newsletters.
Business process changes can impact the quality of the data warehouse. Even though the source for a quality problem may eventually be traced to a business process, user confidence in the data warehouse may suffer. An effective data stewardship program can help companies understand the impact of business process changes on the content of the data warehouse. Through this understanding, conscious decisions can be made prior to implementing the change, and information about the change can be disseminated to those impacted.
1 Inmon, W. H. "Knowing your DSS End-Users: Tourists, Explorers and Farmers." DM Review, October 1996, p. 29.
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
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access