Business Intelligence Techniques

In this chapter, we provide a broad overview of accounting history commencing from 8000 BC, when simple tokens recorded evidence on transactions, through the ancient civilizations, where clay and papyrus were used, to the invention of the first printing press in the fifteenth century to modern times. We focus on how accounting philosophy developed to take into account legal, competitive, and especially technological changes in the environment. From the development of the Abacus in around 3000 BC to present day sophisticated accounting software, we discuss how accounting has changed and adapted to environmental needs.

The ability to make effective decisions is crucial to an organization’s survival in today’s tumultuous business environment. In order for firms to evaluate alternatives and make informed choices they must have reliable and timely data upon which to make their decisions. Consequently, the development of effective data management techniques is of central importance to an organization. Yet, many firms are learning that this is no easy task as they find themselves inundated with nearly overwhelming amounts of data. Assessing the specific data management issues firms face and the development of an effective methodology to address these issues is a central focus of this chapter. Specifically, this chapter explores data management from a cybernetic approach and focuses on methods of transforming various forms of structured and semi-structured data into structured, useful data that an organization can utilize to make effective, informed decisions.

The value of an Accounting Data Warehouse to an enterprise is obvious to most business managers. A central place for decision support where the numbers tie to the books of the company is a great boon to the analysts in the enterprise. However, what are not obvious are the complexities and challenges that must be overcome to actually populate the Accounting Data Warehouse with useful information. This chapter reviews the common challenges that will be encountered in the creation of the so-called Extract Transform Load (ETL) process; the software plumbing that carries data from the source systems to the Accounting Data Warehouse. An overview of the basic components of an Accounting Data Warehouse is first presented to provide necessary background for the discussion. The common challenges encountered in extracting data, transforming it and then loading it into the final warehouse data structures are then described. All discussion is presented in conceptual terms without falling to low-level technical-speak. At the conclusion of the chapter, the reader will have a solid understanding of the relationship between the capability or usefulness of the Accounting Data Warehouse and the complexity and associated cost of the ETL process required to deliver that capability.

At the heart of every organization, regardless of location or industry, is the accounting function. The accounting function is responsible for gathering, recording, and reporting on all of the financial transactions consummated by the enterprise. Using sophisticated transaction applications, accountants aggregate financial activity from disparate functions of the business and record a summary of the transactions in the general ledger (GL). Within the GL, data is organized in the chart of accounts, which represents the physical organization of the business. Business units, departments, products, and accounts are examples of chart of account segments. Accounting then produces the financial statements (i.e., balance sheet and income statement) from the data in the GL. The financial statements are distributed to internal and external stakeholders to evaluate the performance of the enterprise.

This method of capturing and organizing data works very well; however when it comes to reporting and analysis, limitations exist. Transaction systems are not designed to support dynamic analysis. Complicated data models make querying the database an arduous task. In addition, reporting from the GL is limited to the intelligence captured in the chart of accounts. Therefore, to improve financial reporting capabilities, a different approach is used. This approach is an Accounting Centric Data Warehouse ^TM (ACDW). To build an ACDW, the data in the GL is extracted and joined in a separate repository with supporting data captured in the operational subsidiary ledgers (sub-ledgers). The sub-ledgers are the applications used to capture the operational transactions of the business. Sales orders, loan servicing, accounts payable, accounts receivable, and inventory are all examples of sub-ledgers. By creating a platform that brings the GL and sub-ledger data together, a deeper level of financial analysis is possible. The ACDW contains summary level balances from the GL down to the supporting transaction detail from the sub-ledgers. With an ACDW, top down or bottom up analysis is possible. Over time sub-ledgers can be incrementally integrated into the ACDW to create a single enterprise decision support platform.

eXtensible Business Reporting Language (XBRL) is a computer markup language for the purpose of corporate business reporting. XBRL is based on XML, which is the universal format for structuring documents and data on the web. In an XBRL compliant report, each financial and non-financial item is enclosed by a pair of XBRL tags, which describes the meaning of the item. These tags provide semantic information to the reports and make the financial reports not only human readable but also computer comprehensible. Leveraging the power of computers and the Internet, XBRL provides the financial community a standard to electronically and automatically prepare, publish, exchange, and extract financial statements. It is expected that XBRL will have widespread effects on financial reporting. First, preparing financial reports will be easier with XBRL. Financial information needs to be keyed into the computer only once. XBRL-ready accounting software can generate financial reports in different formats, such as for SEC filing, loan application, or corporate web reporting. It greatly reduces the manual input burden and entry errors. Second, the publication and exchange of financial reports can be facilitated. Because XBRL-compliant reports use standardized tags, the reports can be conveniently exchanged between different software and computers, independent of the software formats and computer platforms. The financial statement users can view an XBRL report as easily as browse a web page. Third, since XBRL-compliant reports are computer comprehensible, the financial information contained in these reports can be reliably and automatically extracted through XBRL-enabled applications for financial analyses. The cost of financial analyses can be largely reduced. Decision makers, investors, creditors and other financial statement users can be alleviated from the burden of manually analyzing the financial reports. Small- and mid-sized companies will have an equal opportunity to compete with industrial giants for the capital market. XBRL is undergoing fast development. Since its inception in 1998, over 170 companies and organizations including Big 4 accounting firms, FDIC, Microsoft, and SAP have joined the XBRL consortium. XBRL will become a key standard for electronic financial reporting in the near future.

In this chapter we ascertain the significance of OLAP (On-line Analytical Processing) and its various tools useful for Accounting. We start with a brief history of OLAP. Then we define OLAP. We move on to analyze various types of OLAP- the Relational OLAP (ROLAP), the Multi-dimensional OLAP (MOLAP), the Hybrid OLAP (HOLAP) and the Desktop OLAP (DOLAP). We discuss some applications of OLAP tools in various areas of accounting. Finally, through a series of comparisons, we evaluate the different types of OLAP and select the type that could suit your needs.

An accountant commonly faces questions such as, “What is the profitability for fourth quarter across the X region for main products A and B?” This type of question requires multiple dimensions or perspectives on the data, such as time, region, and products. Multidimensional analysis is the process of analyzing data across multiple dimensions, such as sales per month for each product in each region, product performance by city by store, etc. Compared to the spreadsheet, which allows analysis of only two dimensions at a time, multidimensional analysis allows users to analyze data from an infinite number of viewpoints. An OLAP system would allow multidimensional analysis enabling you to gain insight into data through fast, consistent, interactive access to a wide variety of possible views of information. It allows you to view your data in the same way you think of your business and to see what is ultimately driving your business.

This study is an extension of prior studies that have used artificial neural networks to predict bankruptcy. The incremental contribution of this study is threefold. First, we use only financially stressed firms in our control sample. This sampling enables the models to more closely approximate the actual decision processes of auditors and other interested parties. Second, we develop a more parsimonious model using qualitative “bad news” variables that prior research indicates measure financial distress. Past research has focused on the “usefulness” of accounting numbers and therefore often ignored non-accounting variables that may contribute to the classification accuracy of the distress prediction models. In addition, rather than use multiple financial ratios, we include a single variable of financial distress using the Zmijewski distress score that incorporates ratios measuring profitability, liquidity, and solvency. Finally, we develop and test a genetic algorithm neural network model. We compare its predictive ability to that of a backpropagation neural network and a model using multiple discriminant analysis.

Neural networks are data driven methods. They provide additional information to the decision process as might be left hidden otherwise. Neural networks have already been applied in many different business areas; and they can be used for prediction, classifying, and clustering. They can learn, remember, and compare complex patterns. This chapter shows how a neural network, especially Kohonen’s self-organizing map (SOM), can be used in visualization of complex accounting data. The SOM is used for clustering ten years of monthly income statements of a manufacturing firm. The purpose is to show how the data sets of various accounts and years form their own groups. We found that the SOM can be a visual aid for classifying and clustering data sets, and that it reveals if some cluster contains data that a priori should not be in it. Hence, it can be used for signaling unexpected fluctuations in data. Furthermore, the SOM is a possible technique embedded in the continuous monitoring and controlling tool.

Modern accounting information systems provide decision-makers with such a large volume of accounting data, it can overwhelm even the most sophisticated accounting decision-maker. The result may be underutilization of relevant information. Advances in information visualization technologies provide an effective alternative to address the current and future volume of accounting information. This chapter addresses the theoretical background of visualizations, describing the literature that supports the use of visualizations. The application of information visualization to accounting is discussed, followed by a description of variety multidimensional visualizations techniques.

An important issue in the fields of accounting and management decision-making concerns the alignment of the accounting information system with an organization’s needs for conducting business intelligence activities. The present research identifies sources of requirements for business intelligence activities that are contingent on the degree of organizational formalization, information interdependence among functional areas, and dependence in interorganizational information sharing. Results of the empirical study indicated that, as hypothesized, the alignment between the accounting system design and the requirements for business intelligence resulted in a more successful system.

Recent years have seen significant advances in systems development methodologies. Structured systems analysis and design approaches have been complemented and often substituted by a variety of new approaches such as prototyping, Object-Oriented Analysis and Design methodologies (OOADM), and Rapid Application Development (RAD), among others. System development methodologies and methods have always reflected the available toolsets, e.g., Fourth Generation languages and CASE tools, which enabled rapid application development. Organizational focus has also shifted over the years from Transaction systems to decision support and competitive intelligence. The frequent, expensive occurrences of implementation failures are a stark reminder of the need for appropriate methodological approaches to implementing BI systems. This chapter examines the peculiar methodological needs of BI systems and contrasts those systems with earlier transactional and reporting systems. Based on this comparison and analysis, a methodological framework and approach is proposed for effectively developing and implementing BI information systems.

While standard accounting needs tend to be relatively static, the demand for corporate managers to be constantly aware of changing business dynamics is ever present. Consequently, support systems, technical or otherwise, must be up to the challenge of responding to the demands of their users by incorporating available accounting data. This chapter explains how organizations can use object-oriented analysis and design techniques to more effectively create systems to respond to escalating business intelligence needs. Towards this end, we discuss the differences between business intelligence systems and operational accounting systems. We explain basic object-oriented software principles and how data and procedures are viewed within the object-oriented paradigm. We demonstrate how object-oriented systems can evolve more quickly in response to changing business intelligence needs than systems developed using more traditional (structured) methods. Finally, we provide an example demonstrating the application of these principles. The example demonstrates how the collection of retail sales data and the separate recording of advertising expenditures can be combined to address specific business intelligence questions. This allows us to demonstrate how data collected for specific accounting purposes can be rapidly manipulated and combined, to be used in a decision support, as opposed to the regulatory reporting, role.

Adoption of new information technology often has an indirect effect on an organization’s bottom line; however, many organizations require some minimum financial return on investment before undertaking a major investment. The return on investment in business intelligence (BI) such as data warehouses, data marts, reporting and query tools, and analytic applications are even more challenging to measure since many of the benefits are intangible. How can a company objectively measure the value of improved communication and information for decision-making?

It has been noted that traditional financial performance measures do not measure the increase in value when companies improve their capabilities through the use of new technology. Further, traditional financial measures such as return on investment (ROI) appear to show improvement even when the technology is not being used effectively. In this chapter, we discuss the nature of the Balanced Scorecard and illustrate how it can be used as a measurement tool to evaluate the return on an investment in technology such as BI.

The 21st century organization is evolving from a bureaucratic form based upon hierarchy to a new-form based on knowledge and networks. This chapter explores the role of business intelligence in this new-form organization. It develops a model that positions business intelligence as the primary source of explicit knowledge within a stakeholder perspective, integrating with human and social capital (tacit knowledge sources). Stakeholder theory provides a useful theoretical basis for this model as it offers a comprehensive way to depict business intelligence as a pathway, which enables and captures knowledge. The stakeholder model of business intelligence helps to predict value creation in organizations characterized by indistinct organizational boundaries, information overabundance, and fast-paced change.