In Brief

Use of the term “big data” has become as pervasive as the phenomenon itself. The authors attempt to define the term as it relates to auditing, providing examples of the benefits and obstacles to integrating big data into the audit practice of the future. They also discuss how professional standards may need to change to accommodate the technology.

* * *

The discussion of “big data” has generated tremendous insight into business management and led companies to rethink their strategies, implementing intuitive and meaningful methods of utilizing the wealth of information available in the 21st century. A 2012 AICPA white paper pointed out that technology development and real-time information have posed challenges that warrant greater attention on the adoption of more up-to-date auditing approaches (Paul E. Byrnes, Abdullah Al-Awadhi, Benita Gullvist, Helen Brown-Liburd, Ryan Teeter, J. Donald Warren, Jr., and Miklos Vasarhelyi, “Evolution of Auditing: From the Traditional Approach to the Future Audit,” That is, a new set of resource databases and analytical methods may aid the audit profession by continuously providing high-quality assurance services.

Definition of Big Data

The definition of big data goes beyond the volume of information. For example, Juan Zhang, Xiongsheng Yang, and Deniz Appelbaum describe big data with four “Vs”: massive volume, high velocity, large variety, and uncertain veracity (“Toward Effective Big Data Analysis in Continuous Auditing,” Accounting Horizons, June 2015, What qualifies as “big” is also a relative concept. A data set can be considered big if the information system is either at its maximum capacity or cannot accomplish the task (Miklos Vasarhelyi, Alexander Kogan, and Brad M. Tuttle, “Big Data in Accounting: An Overview,” Accounting Horizons, June 2015, Big data also comes in various formats, such as text, images, voices, and videos. Big data can also be generated independently of human operations. Examples of big data include the number of clicks on ads, the phone call details of customer support, or the complete medical history of a patient.

The impact of big data has also penetrated into daily life, such as the customized Internet search results generated by big data feedback. An effective application of big data can allow organizations to identify common bottlenecks, understand customer behavior, and improve performance. Cisco Systems conjectures that by 2020, the number of intelligent devices connected to the Internet will be 37 billion (Murali Nemani, “Cisco and Verizon Showcase the Connected Athlete Experience,” Cisco Blogs, Jan. 8, 2013, Andrew Leonard has predicted that, “the companies that figure out how to generate intelligence from that data will know more about us than we know ourselves, and will be able to craft techniques that push us toward where they want us to go, rather than where we would go by ourselves if left to our own devices” (“How Netflix Is Turning Viewers into Puppets,” Slate, Feb. 1, 2013,

Big data has also been adopted in accounting practice. With radio frequency identification (RFID) technology, a company can track its products from assembly lines through stores. This allows immediate adjustment to inventory, as opposed to using the assumptions in traditional inventory methods (e.g., FIFO, LIFO). Another use of big data is the measurement of intangibles. If data on customer satisfaction can be continuously collected and analyzed from worldwide social media and platforms, a company can have more reliable and timely evidence of an intangible’s true market value, which remains a daunting task under current accounting measures (Vasarhelyi, Kogan, and Tuttle 2015).

Big data is anticipated to make important contributions in the audit field. It is useful to external auditors by enhancing the quality of audit evidence and facilitating fraud detecting (Kyunghee Yoon, Lucas Hoogduin, and Li Zhang, “Big Data as Complementary Audit Evidence,” Accounting Horizons, June 2015, One of the most useful potential uses of big data is its ability to provide a population-based audit, the results of which should generate more relevant audit evidence (Roshan Ramlukan, “How Big Data and Analytics are Transforming the Audit,” Financial Executives International Daily, Dec. 16, 2015, For example, if an auditor has access to a client’s complete record of accounts receivables, a thorough examination (e.g., existence, confirmation, collections) can be conducted to mitigate the bias from sampling. Furthermore, such high volume allows an audit firm to stratify the accounts receivables by difference variables (e.g., transaction amounts, time, location) and make comparisons across the stratified groups to find patterns and obtain more meaningful insight.

The shift from traditional audits to future audits will not happen because auditors choose to do so. The main driver of big data application by auditors is client-side demands.

Moreover, big data can improve the efficiency of overall data analytics, including descriptive, diagnostic, predictive, and prescriptive analytics. These analyses can provide descriptive statistics on the entire population, offer audit evidence on a larger and more complete scale, build connections between financial statements and actual business operations, and identify potential red flags. Internal audits can also benefit from big data by utilizing more unstructured and nonfinancial information to control risks. The actual integration of big data into future audits will require further consideration.

Traditional Audits versus Future Audits

The shift from traditional audits to future audits will not happen because auditors choose to do so. The main driver of big data application by auditors is client-side demands (Michael G. Alles, “Drivers of the Use and Facilitators and Obstacles of the Evolution of big data by the Audit Profession,” Accounting Horizons, June 2015, Alles shows the analogy between Enterprise Resource Planning systems (ERP) and big data in terms of their impact on audit practice. If ERPs can create the motivation for the audit profession to adopt IT-based audits, the same should apply to big data.

Imagine an environment filled with clients utilizing big data in their business operations. An auditor without sufficient knowledge and skills will have a hard time understanding the client’s business and providing assurance. The application of big data in a client’s day-to-day functions can affect auditing in two ways. One is through the use of accounting information contained in various formats (e.g., audio, image, video). For example, sales calls made to customers might be combined with revenue numbers to provide a complete record of sales activities. In this case, the auditor who examines the income statement would need to understand and analyze the integrated data. Similarly, surveillance videos can be merged with inventory data, which requires the auditor to use them effectively as complementary audit evidence. In addition, big data can generate accounting information on a real-time basis, increasing both the volume and speed of data collection. Therefore, auditors will need to be more familiar with both structured and unstructured data.

Another way big data can potentially change audit practice is the increased complexity in the client’s overall business environment. An auditor should always obtain a sufficient understanding of a client’s business in order to identify risks, such as suspicious transactions and potential fraud. For example, Netflix has been using big data for color and title analysis in its TV show covers. This data has allowed Netflix to thoroughly understand its customers’ viewing habits and develop shows that target specific demographics (Phil Simon, “Big Data Lessons from Netflix,” Wired, March 2014, Thus, activities that contradict the established customer preference can be recognized by the auditor as areas of risk.

In short, it seems that the application of big data on business practice has no limits. When companies start to conduct business and record accounting numbers in a way driven by big data, the audit profession must update its knowledge of big data.

Integrating Big Data into Audits

While the potential of big data might make it appealing to audit firms, its actual integration into audits is not yet mature. Several elements must be addressed. First, big data integration starts with the combination of traditional data and big data. These two sources are equally important to audit procedures, as they imply different types of information. While traditional accounting data is mostly quantitative and structured, big data also includes unstructured and semi-structured data that offer more supporting evidence and detailed information. Given the complex nature of modern business transactions, auditors often need to obtain various types of evidence. Yoon et al. argue that the addition of big data can enhance the sufficiency, reliability, and relevance of audit evidence, which further improves audit quality. For example, in verifying shipment information, traditional shipping documents are the primary proof of occur-rence. Additional big data, such as GPS data, can provide more solid verification. In short, auditors should first identify potentially relevant and useful big data, then collect and merge data.

Nonetheless, data aggregation at this level faces serious challenges, mostly due to data incompatibility; big data is unstructured and lacks a common identifier. Consider a scenario where an auditor, in an attempt to verify revenues of an energy company, wants to combine the phone call details of each service installation with the number of sales. Performing this task requires both a thorough understanding of the two data sets and sufficient competency in data programming, which points to the other two necessary components in big data integration: human resources and technology.

Another serious issue with the inclusion of big data is the security related to data storage. Because aggregate big data can include sensitive information, addressing confidentiality is important to both clients and regulators. It might also raise concerns about independence when external auditors know too much about their clients.

Data aggregation at this level faces serious challenges, mostly due to data incompatibility; big data is unstructured and lacks a common identifier.

The second element of big data integration is the talent training process. The end results of big data integration mostly depend on the competence of the people managing it.

Even with automated systems, it is questionable whether manual labor will be significantly lessened, as the integration of big data will demand a greater skill set. For example, an auditor who used to examine traditional audit evidence regarding inventory will now have to gather other relevant evidence supported by big data and analyze it. Audit professionals therefore might need to become experts in both audit and information technology (IT).

Meanwhile, the recruiting and training of future auditors already proficient with big data is a difficult task. Universities should design accounting courses with focus on data skills and encourage interactions between the accounting and computer fields. Both audit firms and their clients should plan continuous training sessions or workshops that improve auditors’ knowledge and skills in data management. They should also allow auditors to rotate through some positions and receive cross-departmental training. The role of regulators is important here as well, because new standards on professional exams can change the content of accounting education. For example, standard A7 of the Association to Advance Collegiate Schools of Business (AACSB) suggests that “accounting degree programs include learning experiences that develop skills and knowledge related to the integration of information technology in accounting and business” (

There is great potential for big data to be used in analytical procedures. Due to its large volume and real-time basis, big data can allow for population-based audits. This is perhaps its most significant contribution; if every analysis (e.g., trend, ratio, comparison) can be conducted on the population level, it leaves very little room for risks and mistakes. For example, each sales transaction can be compared to prior transactions, both from the same client and from other entities in the same period, to identify anomalies in revenue data. Trend and ratio analysis will also become possible for individual transactions.

Another use of big data is to enhance the degree of prediction accuracy. The relationship between two or more financial items can be more reliably determined from detailed, real-time information. The same also holds for predicting the relationship between a company’s financials and industry averages. Big data will also make fraud detection more effective by generating connections between financial and nonfinancial information. This is particularly relevant to monitoring management and boards of directors. For example, the emails, phone calls, and meetings of the audit committee can all be collected and analyzed to identify potential patterns or links with financial data. Overall, effective analytical models that can truly capture the essence of big data will need to be designed.

Finally, big data can also be integrated into auditing outside of financial statements. An important example would be the auditing of external business relationships (EBR). While a company’s relationships with external entities (e.g., suppliers, distributors, strategic partners) can create both tangible and intangible benefits, EBRs also carry risks. For example, reputation damage to a supplier can potentially harm the business itself, and any dispute over fees rendered can delay revenue generation. Big data allows an auditor to collect information on a client’s EBRs, especially in risky areas not captured by accounting data. Examples include online reviews or news reports.

When accounting data evolved from paper-based ledgers to automatically recorded transactions, the accompanying technology (e.g., Quickbooks, Oracle) also emerged quickly to facilitate the transition. Similarly, the shift to big data audits cannot be achieved without matching hardware and software, including storage devices, data design and programming software, and analytical tools.

The implementation of big data tools will encounter three obstacles. One is the level of user-friendliness (Alles 2015). This might seem unnecessary, but even experienced IT professionals, however, need time to adjust to new systems, not to mention regular audit staff. Big data software that requires strong data skills can be rather expensive if the audit firm has a high turnover rate, due to the costs associated with training new employees.

Another potential barrier is information overload. Prior studies have shown that too much information impedes an auditor’s predictive ability by limiting information processing (Roger Simnett, “The Effect of Information Selection, Information Processing, and Task Complexity on Predictive Accuracy of Auditors,” Accounting, Organizations, and Society, October 1996,; Helen Brown-Liburd, Hussein Issa, and Damielle Lombardi, “Behavioral Implications of Big Data’s Impact on Audit Judgment and Decision Making and Future Research Directions,” Accounting Horizons, June 2015, While an effective big data application can mitigate the negative impact by providing more accurate and relevant information, a suboptimal alternative will only aggravate the overload.

If the collection and merging of big data become feasible and relatively effortless, auditing standards should emphasize the importance of true population examination.

Finally, compatible software that can efficiently handle both the volume and analysis will be unavoidably costly. The mere price of external drives used for continuous data storage will deter many audit firms in integrating big data tools (Kumar Setty and Rohit Bakhshi, “What Is Big Data and What Does It Have to Do with IT Audit?” ISACA Journal, 2013, Using expensive big data applications in auditing work might seem less cost-effective than applying them to consulting services, which could harm audit quality (Tammy Whithouse, “Auditing in the Era of Big Data,” Compliance Week, April 2014,

Predicament in Big Data: Auditing Standards

When practice in a field changes, it usually leads to the discussion of potential regulatory changes. Accordingly, if audit procedures adjust to big data, auditing standards should move in a similar fashion. Based on the above concerns, there is room for the current auditing standards to be improved to address these issues.

First, if the collection and merging of big data become feasible and relatively effortless, auditing standards should emphasize the importance of true population examination. The PCAOB’s Auditing Standard (AS) 2315, Audit Sampling, states the following:

The size of a sample necessary to provide sufficient evidential matter depends on both the objectives and the efficiency of the sample. For a given objective, the efficiency of the sample relates to its design; one sample is more efficient than another if it can achieve the same objectives with a smaller sample size.

The current sampling standard lacks proof of effectiveness and consistency; too much is based on auditors’ professional judgment, and the sampling results cannot be fully verified to support the procedure. Thus, some believe that traditional sampling methods should be gradually abandoned and replaced with population-level audit work when the data are available (see John P. Krahel and William R. Titera, “Consequences of Big Data and Formalization on Accounting and Auditing Standards,” Accounting Horizons, June 2015, Furthermore, the standards should provide clarification on choosing between population and sampling audits. Thanks to big data, it is possible that a population-level audit will be more cost-effective than sampling, considering the planning work involved. If standards permit, the results might be audits that are more effective at lower cost.

Second, audit standards should provide more in-depth explanations regarding the qualification of auditing professionals, especially on the expected specific skill sets. For example, AS 1010, Training and Proficiency of the Independent Auditor, states the following:

The attainment of that proficiency begins with the auditor’s formal education and extends into his subsequent experience. The independent auditor must undergo training adequate to meet the requirements of a professional. This training must be adequate in technical scope and should include a commensurate measure of general education.

The current standards use the term “adequate training,” which is both vague and subjective. More specific content should be included to address training in technology-related skills. The standards should be updated to be more consistent with business trends.

Finally, standards should consider the effectiveness of analytical tools when big data are analyzed on a regular basis. The monitoring of a data system is at least as important as the monitoring of the data itself. Regulations analogous to the auditing of internal control systems should be established to audit future big data systems.

Further Consideration

The reliability of big data remains a concern; thus, the authors advocate studies on the techniques in big data extraction and its safekeeping. The authors also encourage research on the relevance of big data in generating audit evidence. More detailed descriptions and standards should be provided to clarify whether a specific piece of big data should be included in audit evidence. Finally, to enhance audit quality, the authors believe it is important to investigate how big data can facilitate the detection of accounting anomalies, material misstatements, and fraud.

Jiali (Jenna) Tang is a PhD student in accounting at University of Massachusetts–Lowell, Lowell, Mass.
Khondkar E. Karim, DBA, CPA is a chair of the accounting department and a professor of accounting at University of Massachusetts–Lowell.