Business Intelligence: Data Warehouse

Follow FullPress on Google

Add us to your preferred sources.

This last definition is the most widespread and recognized, as it shifts the focus to a series of peculiar aspects of the DWH: 1. oriented to the object of investigation (the customer, sales, …) in the sense that it includes all the data that will be used in the control and decision-making process, grouping them by areas or themes of interest (Datamart) and targeting them to those who use them and not to those who generate them; in other words, while traditional information systems support specific operational functions or activities (inventory, billing, order management, etc.), in the DW the data are organized by analysis objects deemed relevant: products, customers, agents, sales points, and so on, in order to offer all the information relating to a specific marketing phenomenon or relevant fact; 2. integrated, that is consistent with a global conceptual data schema, the corporate glossary, and with respect to measurement units and decoding structures shared at the corporate level; in other words, while data stored in operational information systems are often heterogeneous in terms of coding and format, in a DW data are homogeneous and consistent; 3. time-variant, that is the data in the DW have a historical time horizon of 3 or 5 years, include current data and often forecast data relating to the immediate future; non-volatile, meaning that data is loaded periodically offline, that is, once correctly stored it can be accessed but not modified by the user; in other words, operational data of transactional systems are continuously updated and are valid only at the moment they are extracted (i.e., for example, invoice data extracted one hour earlier or one hour later can be very different); in the DW, data relating to each object or phenomenon to be analyzed generally refer to a specific period, are loaded periodically in bulk and then analyzed: the original loaded data are never modified and maintain their integrity over time, because they refer to events that occurred (e.g., sales in the week, number of receipts in the month, etc.) that must therefore not undergo any modifications, to be preserved accurately and be reusable at different times.Methodology The life cycle of a DW and Marketing Intelligence system is structured, like all information systems, into three macro-phases of design, realization, and management: in this context, we will focus only on the design phase, which has a more managerial and less technical approach than the others. The first critical phase of DW design is to identify the Business Requirements or the Business Model of the marketing intelligence system, defining the set of information, indicators, quantitative measures, and dimensions of analysis most relevant and significant for the entire identified user recipient group. Following the typical “top-down” design logic, the starting point must necessarily be the definition of marketing informational and functional needs, then subsequently seeking the elementary input data necessary to meet them, although this requirement gathering activity is structurally difficult and uncertain: there are specific techniques in this regard (guided or free surveys, brainstorming, analysis of the user information system, prototyping, etc.) that help to conduct the process but do not definitively resolve the problem of defining marketing informational needs which in any case evolve, gradually appear passing to successive levels of analysis, understanding of marketing phenomena and, therefore, experience, also in the use of these more advanced marketing intelligence systems. The next two phases, data modeling and the identification of all data sources necessary to produce the defined information and indicators, present another set of critical issues mainly concerning the design of the DW database architecture, founded on the concepts of “breadth” and “depth” of data. Breadth is defined by the number of phenomena, information, indicators, and measures the system must be able to produce; depth depends on the desired level of data detail (in jargon, maximum granularity), e.g., day or hour in the time dimension, individual customer, individual product item, minimum territorial area (municipality or census zones), etc. Once the architecture of the DW and marketing intelligence system is designed, it is necessary to focus on the main phases and processes of the realization macro-phase:The loading or population process of the DW The goal of a DW is to bring together data from operational systems with those from external databases and funnel them into the management environment to make information available through business intelligence systems. Data acquisition for the DW is performed through complex operations of data capture from source systems, their cleansing, and subsequent transformation based on business rules defined in the project’s development phases (for example, uniformity of measurement units). Once transformed, the data are mapped and transported into the various DMs or the DW.The search and production process of marketing information Constitutes a further crucial phase in the construction process of a DW and marketing intelligence system. The significant efforts required and dedicated to the realization of the DW often distract attention from end-user tools and front-end applications necessary for searching and accessing information in the same databases, resulting in end-users being left with their traditional tools already in use, such as spreadsheets, very rigid custom applications, and not very user-friendly interfaces. Flexibility and ease of information production, as well as data access, are the keys to the success of the decision-making process and, therefore, the DW. Areas of Application The main application areas of Business Intelligence & Decision systems are: • Marketing & Sales Analysis • Customer & Marketing Database • Budgeting (cycle, formulation, etc.) • Financial Reporting • Financial Consolidation • Management Reporting (directional reporting) • Executive dashboards, Tableau de Bord (EIS), Balanced Scorecard • Profitability Analysis; Quality & Satisfaction Analysis • Clickstream Analysis • Geomarketing RESERVED REPRODUCTION