Dependent Data Mart
Data marts populated with data sourced from the enterprise data warehouse are called dependent data marts (Berson & Smith, 1997:124, and Han & Kamber, 2001:67). These are sometimes also called replicated data mart since the data mart is populated with a portion of the data warehouse’s data through replication (Gray & Watson, 1998:104). This is depicted schematically in Figure 2.
By integrating data from various sources into a data warehouse and then letting the integrated data flow to a number of dependent data marts each containing limited content tailored to the needs of the departmental user community gives users the benefit without central co-ordination being sacrificed (Dodge & Gorman, 1998:572-573). The information provided to each data mart is determined by the departmental or divisional needs for centralised information. It is also possible for locally generated and used information to be stored in the dependent data mart. Furthermore, because the use is localised, alternative analytical tools to those used to access the underlying data warehouse can be used (Simon, 1998:4-5). What is important to note is that data is “never transmitted by the owner of the data mart to other portions of the organization” (Gray and Watson, 1998:105).
Interdependent Data Mart
In order to address the problems experienced with both the enterprise data warehouse together with the dependent data mart, and the independent data mart, a co-ordinated set of data marts are used where the planning and design is at an enterprise level (Han & Kamber, 2001:67). According to Berson and Smith (1997:126) the “key to a successful data mart strategy is the development of an overall scalable data warehouse architecture; and the key step in that architecture is identifying and implementing the common dimensions.” Kimball, the proponent of the dimensional modelling approach, refers to the common dimensions as conformed dimensions: “the only … way to combine the data from these separate tables and achieve an integrated enterprise data warehouse is if the dimensions of the data mean the same thing across these tables” (Kimball et al, 1998:18-19). Kimball’s conformed dimensions are explained by Berson and Smith (1997:125-126) as follows: “For any two data marts in an enterprise, the common dimensions must conform to the equality and roll-up rule, which states that these dimensions are either the same or that one is a strict roll-up [or subset] of another.” By predefining the characteristics of dimensions to be used throughout multiple interdependent data marts a conformity or dependency between data marts is created which allows for “easy rollup into an enterprise-level data warehouse” (Microsoft Corporation, 1999:4).
According to Kimball et al (1998:18-19) the only feasible solution is to blend the top-down or ottom-up approaches, where separate pieces, i.e., interdependent data marts, are designed guided by a proper architecture. As Poe et al, (1998:18) state, organisations may choose to “begin their corporate data warehouse project with a small pilot project for a specific subject area (business function). In so doing, those organizations have taken a bottom-up approach to the implementation of a decision support environment—and they have essentially created both a data mart and their first data warehouse simultaneously.” This approach requires that a significant amount of work to be spent during the design of the first interdependent data mart in order to ensure future data marts can fit with the dimensions (Dodge & Gorman, 1998:572). If this is not done, the result will be multiple independent data marts. Note that the only difference between an independent and an interdependent data mart when building the first data mart is the intention and the design of the data model. The structure will appear the same when depicted diagrammatically.
CHAPTER 1 INTRODUCTION
1.2. PROBLEM STATEMENT, DEMARCATION AND METHODOLOGY
CHAPTER 2 THE ROLE OF INFORMATION IN ORGANISATIONS’ SALES AND MARKETING STRATEGIES
2.1. THE FOCUS IS ON THE CUSTOMER
2.2. SALES AND MARKETING IN ORGANISATIONS
2.3. MANAGEMENT INFORMATION AND DATA SOURCE REQUIREMENTS
2.4. DATA FLOW FROM CUSTOMER-FACING CHANNEL PARTNER FIRMS
2.5. CHAPTER CONCLUSION
CHAPTER 3 DATA MARTS AS MANAGEMENT INFORMATION DELIVERY MECHANISMS
3.1. DEFINITIONS OF DATA WAREHOUSES AND DATA MARTS
3.3. NATURE OF USAGE
3.4. USING DATA PROVIDED BY CHANNEL PARTNERS
3.5. CHAPTER CONCLUSION
CHAPTER 4 USE OF DATA MARTS IN MANUFACTURING ORGANISATIONS WITH THIRD PARTY DISTRIBUTION
4.1. CASE STUDY 1: BRANDED CONSUMER PRODUCTS INDUSTRY
4.2. CASE STUDY 2: PHARMACEUTICALS INDUSTRY
4.3. CHAPTER CONCLUSION
CHAPTER 5 COMPARATIVE ANALYSIS AND CONCLUSION
5.1. COMPARATIVE ANALYSIS OF CASE STUDIES
5.3. SUGGESTIONS FOR FUTURE RESEARCH