1. PLANTEAMIENTO DEL PROBLEMA
7.6. Modelo y método
The rapid development of IT has resulted in various IS tools that enterprises can adopt to enhance their transaction-based systems with knowledge processing capability to support more strategic and complex decisions. The data warehouse has surfaced as a key source of information for knowledge workers and managers. Its well-publicised value in offering high query-response performance and increased information accessibility, as well as being an integrated source of data, is creating an extremely popular environment for decision support in firms (Watson et al. 2004; Watson & Haley 1997).
Many firms have turned to data warehouse to assist in making decisions about changes needed (Little Jr. 1998). Data warehouse has been cited in the literature as being one of the most powerful strategic weapons (Park 1999). Data warehouse emerged in response to the problem encountered in providing information for use by DSS, with the main limitation being the lack of separation of operational data suitable for decision support (McFadden 1996)
Unlike numerous operational systems, these are not designed to support strategic decision- making. Operational systems are designed to support and maximise the day-to-day, value creating work (Connolly & Begg 2002; Connolly & Begg 2005). There are several reasons
49 why existing operational systems could not meet these needs. Singh (1998, p.16) mentioned the following: 1) the lack on-line historical data; 2) the data required for analysis resides in different operational systems; 3) the query performance is extremely poor which in turn impacts performance of operational systems; 4) the operational database designs are inadequate for decision support.
Operational systems are optimised to automate business operations and must be efficient with transactions that are predictable, repetitive, and update intensive. These systems are
organised around business functions and the processes building up these functions (Barquin & Edelstein 1997). Conversely, data warehouse systems are designed to support efficient
processing and presentation for analytical and decision-making purposes and to provide the decision makers with suitable data and information (Poe, Klauer & Brobst 1998). A data warehouse holds data that is current, historical, detailed, and summarised to various levels. Apart from being supplemented with new data, the data in a data warehouse is seldom subject to change (Connolly & Begg 2002). The number of users served by a data warehouse is smaller than for operational systems (Barquin & Edelstein 1997).
On the other hand, the data warehouse is designed to support relatively low numbers of unpredictable transactions that require answers to queries that are unstructured, heuristics, and ad hoc. The data in a data warehouse is organised according to the requirements of potential queries and supports strategic decisions of managerial users (Connolly & Begg 2002). Moreover, the time horizon for holding the data in a data warehouse is importantly extended compared to operational systems. Generally, the time horizon for a data warehouse is five to ten years, whereas an operational system holds its data 60 to 90 days (Inmon 2005). The
50 following table shows the comparison of operational information systems and data warehouse (see Table 2-1).
Operational Systems Data Warehouse Systems
Application-oriented Subject-oriented
Transaction-driven Analysis-driven
Hold current data Hold current and historical data Store detail data Store summarised and detailed data
Repetitive processing Ad hoc, unstructured, and heuristic processing Predictable pattern for usage Unpredictable pattern usage
Support day-to-day decisions Support strategic decisions
Serve large number of operational users Serve low number of managerial users
Data is dynamic Data is static
Source: Connolly & Begg (2002)
Table 2-1: Comparing operational systems (OLTP: on-line transaction processing) and data warehouse systems
However, in today’s competitive age, as data is becoming an increasingly significant resource in supporting organisational procedures, the quality of the data that executives use becomes critical (Paradice & Fuerst 1991). Steiger (1998) suggested that the data warehouse has presented decision-makers with far more information, in a far more flexible form than has been true in the past. Accordingly, data warehouse applications have become an essential component of BI and decision support applications.
For example, previous study indicated that the introduction of data warehousing12
12
Data warehousing is defined as a relational database specifically organised to provide data for easy access (Turban, Aronson & Liang 2005).
technology and OLAP techniques has greatly improved traditional EIS (Chen 1995) and has led to a new EIS architecture that is sometimes referred to as contemporary EIS architecture (Fernandez & Schnedier 1996). In this architecture, the centralised database is replaced by a data
51 warehouse, and OLAP techniques are adopted for multidimensional data analysis and
information presentation (Hammer et al. 1995).
As data warehouses provide the data infrastructure for management support systems, that include many decision support applications: DSS, EIS, OLAP, SCM, CRM, BI etc., data warehousing supports these applications by providing a collection of tools which: 1) collect data from a set of distributed heterogeneous source; 2) clean and integrate this data into a uniform representation; 3) aggregate and organise this data into multidimensional structures which are suitable for decision-making; 4) refresh it periodically to maintain the data up to date and accurate.
There are many benefits that a data warehouse can provide. For example, the data warehouse can improve performance in better-targeted products, improved customer relation
management, and produce greater operational efficiency (Cooper et al. 2000; Moore & Wells 1999). Srivastava & Chen (1999) pointed out that it also results in reengineering of business processes. For instance, automated and integrated information delivered from the data warehouse may substantially free up managers’ time and efforts, thereby increasing their availability for other tasks.
The importance of data warehouse in supporting and improving decision-making is
recognised as major (Ghoshal & Kim 1986; Martinsons 1994; Rouibah & Ould-ali 2002), but the data warehouse does not provide adequate support for knowledge intensive queries in an organisation. The emerging heterogeneity of the decision environment is stimulating the need to explore more effective techniques for mining and presenting data in a meaningful format
52 (Rundensteiner, Koeller & Zhang 2000). Thus, the data warehouse can support multiple beneficial applications rather than being an independent application.
Specific techniques such as OLAP (Datta & Thomas 1999) and data mining (Drew et al. 2001) to produce information can improve performance. OLAP is an enabling technology that allows manipulation of enterprise aggregate data across many dimensions such as product, time, and location, etc. (Codd, Codd & Salley 1993). For example, by using OLAP and data mining tools, firms are able to exploit insights gained from their data warehouse to significantly increase sales (Cooper et al. 2000; Heun 2000; Whiting 1999), reduce costs (Watson & Wixom 1998), and offer new and better products or services (Cooper et al. 2000; Watson & Wixom 1998). Moreover, these techniques enable an organisation to detect weaknesses (e.g., customer dissatisfaction) as well as hidden opportunities (e.g. customer segmentation). Organisations can use these applications to simply provide resources to end- users or to guide end users in making a better decision (Silver 1990).
According to the combination of core technologies, enabling technologies, and BI application solutions, (Brackett 2001) and (Hill & Scott 2004) stated that these decision support
applications aim at the development of an accurate understanding of business dynamics. They also enable the organisation to monitor its environment and observe business trends, to detect new opportunities and avoid threats, by analysing the complex business environment in order to make decisions (Lönnqvist & Pirttimäki 2006). These processes are consistent with the most important components of the BI infrastructure of (Kalakota & Robinson 1999), which consists of 1) key IT (e.g. data warehouse); 2) IT potential (e.g. OLAP, data mining); and 3) decision support applications (e.g. BI, CRM, SCM).
53 An important role of decision support applications is to provide information for users to analyse situations and make decisions. Organisations that are interested to improve the quality of decision-making, image, or quality of partner service should incline towards the development of information technology infrastructure that will represent a holistic approach to business operations, customers, suppliers, etc. (Wells & Hess 2004).
Theory and practice show that the above-mentioned requirements are largely met by
“Business Intelligence” (BI) systems (Liautaud & Hammond 2001; Olszak & Ziemba 2004; Turban & Aronson 1998). Huber (1990) and Leidner & Elam (1995) proposed that the use of computer assisted information storage and acquisition technologies leads to organisational intelligence that is more accurate, comprehensive, timely and available. Olszak & Ziemba (2006) have concluded that the success of many decision support applications affects decision support.
In summary, then, based on the above information, the infrastructure of a successful decision support application that can meet decision makers’ needs should include core technologies, enabling technologies, and application solutions in terms of quality of information, system reliability, ease of use, and speed (Gray 1993; Liautaud & Hammond 2001; Olszak & Ziemba 2004). Thus, organisations and their decision support applications must embrace procedures that can deal with complexity and go beyond the technical orientation of previous decision support characteristics.