Big data in Business: Auditors' Challenges
In recent times the term big data has become significant for the global organizations to modify the business information structure in terms of quality, velocity, and usage. The information structure modification has the great impact on the auditing profession. According to historical data, the term data was once described as structured and human-generated but technical advances have expanded the description and today means both structure, human-generated and unstructured machine-generated data (Sookhak et al, 2017). In this context, it is important that the ‘Internet of Things' has increased the significance of data outside the business boundary in the corporate environment. In simple terms, the Big Data can be described as the mass volume of exponentially growing data. So, it is obvious that this mass volume has the huge influence on the productivity, profit and risk management.
According to the CPA journal the hybrid and merged data culture opens various opportunity in auditors approaches while conducting internal and external audits. As an example, the sampling process is redecorated as the large volume of data can be stratified with different variables to produce important pattern and anomalies. Moreover, the video and inventory data hybrid can be used as audit evidence while GPS data can be useful to verify major shopping information and business transactions (Alles, 2015). However, the auditors face challenges in terms of usage, data availability, cross-platform analysis and understanding of the structured and unstructured data. Still, the opportunities like high quality, relevant insights and fraud identification in business operations are making the Big data inseparable from auditing.
Big Data Definition:
The definition of Big Data changes according to the aspects of applications. However, the most important definition describes the term Big data with four Vs and they are volume, velocity, veracity and variety. Moreover, the term Big is also highly relative in the implementation context. According to "Accounting Horizons," a dataset is big if the information system is not capable of taking the dataset in terms of capacity and task accomplishment. Apart from this the data can be of different types and various formats and usually is human operation independent today. The most common example is the medical history of a patient where different types of data (qualitative & quantitative) is generated after inputting the resources in the machine. In case of organization and industry the Big Data can be generated inside the business environment. According to the data sources the data types and format can change. As an example, the organization can produce customer complaint pattern while industry produces production based consumption rate considering the society related economic data (Yoon et al, 2015). Again, the external sources can make data purchasable or free to use. In this context, the structured and unstructured data are important. The traditional systems define structured data as relational and hierarchical in a sequential database while the unstructured data have no typical structure and constraints making the same a management concern. The structured data can be organized with queries in terms of ownership and vendor supported solutions. But the evolving and unpredictable nature of unstructured data (text, video, audio, graphs, images) is asking new solutions and analytic skills for effective management.
In terms of practical applications, Big data is used by business organizations in creative and innovative ways. Two major examples are given below:
- Retail: The retail business is producing advanced customer profiles using purchase history, geographical location, demographics, feedback aiming the data-driven customer experiences. In this context, the sales department activities are significant as its access to the purchase database identifies the customer profile and media posting. Moreover, the accessed data is combined with the current fashion trend to produce hybrid unstructured data focusing on the customer interests (Zhang et al, 2015). Moreover, the smart information system can deal with availability issue of resources in real-time. As an example, if a customer identifies the non-availability of product (preferred size and color) in the store then customer support system can immediately identify its presence in other outlets or inventory and arranges the home delivery for the customer. The in-store mapping, e-commerce optimization are other functions of Big data analytics (RFID, HADOOP).
- Financial institutions: According to IDC report the Big Data based revenue will be more than $203 billion dollars by 2020. In case of financial institutions, the Big Data technology can produce new revenue streams, secure customer solutions etc. In-depth business and industry analysis via analytics can forecast and predict business trend and outcomes in terms of model options and customer base. According to transaction history and demographic data, the financial institutions can offer personalized services to the international customers while organizing external data. One of the big four Australian bank ANZ has recently invested Big Data startup called Data republic to encourage secure data sharing process (Vasarhelyi et al, 2015). The cloud-based platform will help both governing and auditing structure using open banking reforms and new data right.
Auditors’ aspect in Big Data:
To identify the auditors' aspect of understanding the term ‘analytics' is important. The term refers to a process of data analysis and management to get precise and meaningful solutions. Considering the recent unstable and complex global economic framework the role and responsibilities of auditors are tougher in the financial context. The public interest theory suggests more audit firm intervention in the financial statements to encourage transparency and quality. From the organizational perspective, the major demand is the improvement of auditing dialogue and insights (Cao et al, 2015). It is observed that the emerging analytics are in the market for a long time but the organizations are not ready to use these solutions considering the data capture and privacy difficulties. However, it is significant that the transformation from traditional to the present technology based audit is not dependent on the auditors' choices. The enterprise-wide data repository is specially designed to encourage the data warehousing.
However, the data warehousing is significant for using multiple sources based relational databases where the data transaction (extract, load, transform) from traditional and cloud operating systems is encouraging the data lakes considering large data holding capacity and privacy. Today the changed auditing structure is moved from the sample testing techniques and introducing analytics to get more customized results. If the audit relevant data means the transaction activities and the key business process then intelligent analytics can provide good quality audit evidence with standard insights. Again, the auditors can identify better ways of financial reporting in terms of business risks, frauds and external environment. As an example, the multinational companies sometimes use the alternate audit entity DVD-by-mail for business information transactions. Moreover, the smart audit appliances are visible in the company data center and they help audit data streaming among the key audit teams (external and internal audit teams). In the financial and accounting sector the Big Data analytics like RFID can track products in the assembly lines while changing the traditional inventory methods. Again, the intangible business entity measurement uses the analytics (like Hadoop, Pentaho) to get rack organization, complete report generation combing virtual servers, customer ratings, market value etc.
In terms of the external auditor, the fraud detection and audit evidence generation depends on the population-based audit to generate more accurate and evidence-based outputs. As an example, the conventional sampling process was dependent on the existence, confirmation and collection schemes of data collection to mitigate the bias while the Big Data analytics use the larger volume of data for more complex personalised auditing process using different pattern variables like time, location and transaction amounts (Liu et al, 2014). Again, the internal auditors use the Big Data Analytics for organizational planning and hierarchical risk assessment. The significant risk determination process can help auditors produce appropriate plan and controls. While going through the process the auditors can train the board with Big Data initiatives and benefits. Moreover, this kind of audit decreases the number of different audits and opens the opportunity for a singular stand-alone efficient audit (Big Data).
Both formal and informal assessments can try Big Data programs in advisory projects as well as continuous improvement of board and controls. The data consumption process can help the acquired, integrated and consolidated solutions targeting many source systems. Furthermore, the organizational usage policy can vary the Big Data analytics. As an example, internal auditors can help program governance while aligning the major information with enterprise strategies. The same is also helpful in value creation of internal leadership and management. Introduction of disparate sources provides the complete views of audit components which means the unauthorized modification is limited in terms of substantive testing and reporting quality.
However, there are some integration barriers in terms of technology which increase the risk in the audit environment. Firstly, the efficient and cost-effective data capture schemes of organizations indicate data provision approval form consumer as time-consuming in auditing context where multilayer approval process (for security) increases difficulties. Secondly, auditors have to deal with diverse accounting systems for the same and different organizations which mean data extraction and capture become tiresome processes even using the most competent technologies. In terms of the general ledger, the Big Data Analytics provide sub-ledger information considering the patterns and components which increases the overall volume of data in the key business process (Sookhak et al, 2017). Although the descriptive analytics efficiently identify the major risk areas but it raises the audit evidence issue for every single risk. As an example, the use of the algorithm in data visualization asks the auditor to identify the balance between judgment and report generation which is difficult due to multiple source integration.
In case of integration, the auditors still lack knowledge about the new analytics which is disturbing for the major audit frameworks. In terms of analytics, the substantive procedures usually establish results without auditors expectations where data validation rules cover the underlying databases. That means the system generated report can create conflict between auditors' patterns and digital patterns. Moreover, the material misstatement faces precision issue while dealing with diverse revenue sources. In case of the audit engagement, the major challenges are performance risk, data management and program governance (Yoon et al, 2015). The funding model, related database, EUC application, the third-party interruption can change the business scenario where internal skill and knowledge are limited. Moreover, ineffective technology can create issues. As an example, the data lifecycle policy mismanagement and isolated Big Data system in maintenance strategy can degrade the audit performance. Finally, the reporting accuracy can be compromised via inaccurate procedure and reporting tool selection.
According to the given information, Big Data analytics can integrate the accounting and finance information of multiple sources into the audit framework of organizations. The multivariate statistical structure of the analytics can help distress modeling in finance as well as qualitative and financial fraud modeling. However, the main barrier between the auditors and the emerging technology is the potential knowledge gap. Apart from this, the auditors face evidence, procedure and validation issues considering the multi-source integration policy. But Big Data analytics can open new auditing opportunities in terms of judgment and audit techniques with the help of sentiment analysis and natural language processing.