Amsterdam Business School
The experiences of external auditors about data analytics in the
audit process.
Name: Pritesh Jangbahadoer Sing Student number: 10871403
Thesis supervisor: prof. dr. B.G.D. (Brendan) O' Dwyer Date: 20 June 2016
Word count: 24052
MSc. Accountancy & Control, specialization [Accountancy] Faculty of Economics and Business, University of Amsterdam
Statement of Originality
This document is written by Pritesh Jangbahadoer Sing, a student from the University of Amsterdam, who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Preface
Hereby I present to you my thesis titled:” The experiences of external auditors about data analytics in the audit process”. This study has been conducted at a big-four audit firm in the Netherlands and is written as part of the Master curriculum Accountancy & Control at the University of Amsterdam. This part of my final internship of this curriculum and has been conducted between February and June 2016.
I have created this research question together with my thesis supervisor from the University of Amsterdam, Brendan O' Dwyer and during this process my supervisor from the university and my supervisor from the audit firm always answered any question.
Herewith, I want to thank Brendan O’ Dwyer and my thesis supervisor from the audit firm for their great effort to support me in this process. I also want to thank all my colleagues and all the respondents in this thesis. Without their cooperation I would not have been able to conduct this research. Furthermore, I would like to thank my family, friends, and especially my girlfriend for supporting me morally me at all times.
Finally, I would like to thank my parents for making this all possible for me. Their motivating words has led to eventually writing this thesis.
Enjoy reading!
Pritesh Jangbahadoer Sing Almere, 19 June 2016
Abstract
Studies regarding data analytics remains in the academic domain. This study analyses motivations and experiences of auditors in the audit practice about data analytics while using the four principles of the Formalization Framework, which are: repair, internal transparency, global transparency and flexibility (Adler & Borys, 1996). Semi-structured interviews have been conducted, where I find that data analytics are perceived as enabling by auditors in the audit field. I find that auditors have the required skills to resolve errors while carrying out data analytics and hence experience repair as enabling. Auditors can also select several tools to create multiple reports, making flexibility enabling. Moreover, I also find that internal transparency and global transparency are being perceived as enabling since data analytics enhances the guidance and monitoring of individual staff members and can develop their understanding of the audit process. At the same time, I found new issues regarding that data analytics in the audit practice whereby data analytics provide a more comprehensive understanding and use about how auditors deal with contingencies by conducting more test of details and more communicating with the client. This is contradicting to the research of Dowling and Leech (2014). They found that the system does not always support learning during the process, since auditors can over rely on the audit support system. Correspondingly, this could argue whether data analytics motivates employees to learn in the audit process.
Content
Preface ... 3 Abstract ... 4 1 Introduction ... 7 1.1 Contribution ... 9 1.2 Paper structure ... 10 2 Literature review ... 11 2.1 Data Analytics ... 11 2.1.1 Assessing data ... 112.1.2 Benefits data analytics in the audit field ... 12
2.1.3 Challenges data analytics in the audit field ... 15
2.2 Expertise auditors... 17 3 Theoretical framework ... 20 3.1 Formalization framework ... 20 3.1.1 Repair ... 21 3.1.2 Internal Transparency ... 22 3.1.3 Global transparency ... 23 3.1.4 Flexibility ... 24 4 Methodology ... 25 4.1 Research setting ... 25 4.2 Data Sample ... 26 4.3 Data analysis ... 27 5 Case context ... 29 6 Case analysis ... 30 6.1 Repair ... 30 6.2 Internal Transparency ... 32
6.3 Global Transparency... 37 6.4 Flexibility ... 41 7 Discussion ... 45 7.1 Repair ... 46 7.2 Internal Transparency ... 47 7.3 Global Transparency... 48 7.4 Flexibility ... 49 8 Conclusion ... 51 8.1 Conclusion... 51
8.2 Limitations and further research ... 53
9 Bibliography ... 55
10 Interview Guide ... 63
11 Appendix A ... 65
1 Introduction
Recent company scandals such as Enron, WorldCom and Xerox, arising from false financial reporting, caused serious financial damage for those who held the shares of these companies (Murcia, de Souza, Borba, & Ykeda, 2007). They stated that these scandals made it necessary that auditors use new (continuous) audit approaches. New audit approaches should include more information technology as proposed by the accounting world in order to redesign the confidence in the auditing process (Murcia, de Souza, Borba, & Ykeda, 2007).
The limitations of traditional approach on the reliability and timeliness of financial data generated by IT-based systems resulted in the emergence of new continuous approaches and has become more relevant for the automation and real-time monitoring of big data files that are incorporated eventually in the financial statements (Vasarhelyi, Alles, & Williams, 2010).
Public accounting firms are competing with each other to provide better and more comprehensive data analytical services to their clients, but the question still remains as to how they will actually accomplish this challenge (Earley, 2015). Data analytics is the process of using unstructured and structured data with help of applications of several analytic techniques to provide useful information to decision-makers (Davenport & Harris, 2007); (Sims & Sossei, 2012). Earley (2015) also explains that data analytics has been a significant area of investment for accounting firms, mainly in the advisory services, but also more in auditing lately. The availability of the nature and large amounts of data in companies has been constantly increasing over time.
Gartner (2013) defines big data as ‘‘high-volume, high velocity, and high-variety information assets that demand cost-effective, innovative forms of information processing for enhanced insight and decision-making’’. Marks (2001) shows that firms are rapidly installing new technologies, like computer assisted auditing tools to perform financial analysis in the audit.
This requires auditors to understand these systems and the risks associated with these technologies. Vasarhelyi and Halper (2002) believes that real-time information systems, that record all transactions, will impact the audit procedures and provides a more variety of assurance services to clients.
These systems are supported by new technologies that grant auditors with broader perspectives of both financial and non-financial information and hence improve the audit efficiency that derives from audit automation (Trompeter & Wright, 2010).
Cao et al. (2015) state that auditors are becoming more secure in the examination of audit evidence and in risk assessments. They should explain their findings more extensively and expand their knowledge of data analysis since much of this big data informs and affect decision that are made on corporate level (Cao, Chychyla, & Stewart, 2015). In addition, new auditing standards need to be accepted and implemented to match the relentless pace of technological change, since these technological changes require knowledge about different systems and this should be added to the skillset of an auditor (Vasarhelyi & Halper, 2002).
Auditors with skills in data analytics have greater opportunities to conduct an audit. If they are very skilled they could audit on a more frequently and therefore improve the quality of the audit process (Vasarhelyi, Alles, & Williams, 2010).
On the other hand this could also increase the costs since the collection and analyzation of big data require auditors to work more intensive to access a major part of the data sources (Yoon, Hoogduin, & Zhang, 2015). These are primary costs which are needed in the data processing effort to reach a certain audit statement (Pathak, Chaouch, & Sriram, 2005).
In the audit process there is a lot of room for professional judgment related to the type and amount of evidence collected from those audits, since auditing standards providing rules and guidelines to auditors (PCAOB, 2010).
Furthermore, software used to analyze large volumes of data, like data mining tools as well as more sophisticated tools, increases the skills of individual auditors to understand the story which the data is trying to tell them in more detail (Capriotti, 2014); (Whitehouse, 2014).
Given several concerns like judgment, interpretation problems and the highly regulated environment of auditing, logically firms would be more careful when incorporating data analytics in the audit practice. Many organizations collect massive amounts of data about clients, competitors and the external environment, but they often are not informed about taking the appropriate next step of analyzing the data (Earley, 2015).
The challenging part for the audit profession will be; creating value from big data and to understand how the behavior of auditors engage in the audit process (Brown-Liburd, Issa, & Lombardi, 2015); (Earley, 2015). It is important to understand their decisions, which are based on individual judgment (Brown-Liburd, Issa, & Lombardi, 2015).
The adoption of data analytics in the audit field has been slower than in other fields, such as in advisory practices (Katz, 2014); (Whitehouse, 2014). Yet, data analytics is being seen as the future of audit (Lombardi, Bloch, & Vasarhelyi, 2014). As Capriotti (2014, p. 38) stated: “It has the potential to be the most significant shift in how audits are performed since the adoption of
This paper therefore contributes to the existing literature by scrutinizing how data analytics is experienced in the audit practice. The research question is stated as follows; How do auditors
within a big four firm experience data analytics in practice? 1.1 Contribution
This study extends the research of Dowling & Leech (2014) by also using the formalization framework of Adler and Borys (1996) to gain a deeper understanding about the experiences of auditors with data analytics in the audit process of a big four audit firm.
In the research of Dowling & Leech (2014) they investigated how an audit support system is evolving, interpreted and can change the auditors’ behavior and how the audit team interacts. They used the formalization framework of Adler and Borys (1996) and found that many of the systems features could have been used as coercive, yet the auditors view the system as enabling.
This study has found new issues regarding that data analytics in the audit practice. Data analytics provide better understanding about how auditors deal with contingencies. They do this by conducting more test of details and more communicating with the client. The study shows that auditors use data analytics very extensively, with different tools and on very different ways, like conducting risk analysis or using business process mining tools in the audit process.
This means that whenever the system is used, the users are complying with the rules and regulation. This is contradicting to the research of Dowling and Leech (2014). They found that the choice of selecting a tool is constrained in how they apply the system and users of the system are always complying with the rules and regulation of the firm.
In Dowling and Leech (2014) they found that the repair is facilitated by the system is low, but still seems to be viewed as enabling. This study questions that finding by showing that the same data can be interpreted differently, yet auditors experience that they resolve errors about data analytical systems very well.
Auditors can develop themselves by courses which are provided by the firm. It shows that auditors get reviewed during the audit process which enhances the audit quality. It shows that auditors are working together much better. However, this is not always the case, since an important finding is that employees also experience to have their own role in the organization which does not stimulate teamwork.
Besides, this study provides deeper understanding about judgment and it agrees that auditors have to gain more knowledge about data analytics in the audit process. Changes in the audit process regarding supported systems, like newly implementing data analytical systems, could
lead to very different judgmental issues like overreliance which is also mentioned in the research of Dowling and Leech (2014).
Besides that, this study extends that finding by recognizing ambiguity, meaning that the same information can be interpreted in different ways. This depends on the individual judgment of the auditors. This responds to a call of Dowling and Leech (2014) explaining that there is future research needed to investigate how support systems can impact auditors judgment.
Surprisingly, Dowling and Leech (2014) does mention assurance in their research. By saying this, auditors can follow the rules and guidelines of the firm when working with the audit support system and this can be overruled by them. However, the study does not mention how assurance can be granted by using the audit support system. This study extends the assurance concept by discovering that assurance can be granted when confirmatory activities are executed with data analytics, like linking datasets to source documents.
Furthermore, in their research they do not mention materiality at all. Auditors act in service of the public and the society. Materiality ought to be minimized and financial statements should cope with the true and fair view principle of 95 percent. Data analytics give a total overview of the whole population which is a hundred percent. Important here is to question what society demands of the auditor in the future and whether mistakes are tolerated or not to assure audit quality.
1.2 Paper structure
The remainder of this study is built up as follows: paragraph 2 describes the prior research about data analytics, the benefits and challenges in the literature and will define the notion of data analytics. Also, it provides a deeper understanding about the importance of skilled auditors in the audit practice. Paragraph 3 illustrates the formalization framework of Adler and Borys (1996). This framework provides a theoretical structure for this study. Paragraph 4 explains the chosen research methodology by using semi-structured interviews. Paragraph 5 explains the case context in which this study has been conducted. Paragraph 6 describes the findings of this study. Paragraph 7 discusses the subject to prior literature and the theory. Finally, paragraph 8 reveals important conclusions including limitations and suggestions for further research.
2 Literature review
2.1 Data Analytics
2.1.1 Assessing data
The possibility to gather big data in companies progressively increases over time. However, the pace of new innovations in cloud computing and the use of social networks has changed the accessibility of large data sets and the nature of this data (Earley, 2015).
The gathering of huge amounts of data has been termed as “Big Data” and is often referred to a large population of datasets whose size is above the ability of typical software tools to capture, store, manage and analyze these datasets (McKinsey, 2011). Big data is defined as ‘‘high-volume, high velocity, and high-variety information assets that demand cost-effective, innovative forms of information processing for enhanced insight and decision-making’’ (Gartner, 2013). Big data derives from traditional systems that trace transactions, as well as new data sources such as emails, internet activities and social media (Zhang, Yang, & Appelbaum, 2015). According to Laney (2001) it is also possible to add a fourth term in extension to these three basic “Vs”. Veracity can be added too, as the integrity of accounting information becomes a concern without protecting the log files (Laney, 2001).
Big data must be analyzed in a future oriented way in order to be relevant, reliable and useful for the decision-making process (Gartner, 2013). With the rise of big data, the increased volumes and the number of transaction, continuous auditing is needed in the future to access, analyze and translate lots of information (Vasarhelyi, Alles, & Williams, 2010). Different organizations take advantage of big data to make more relevant and timely decisions about their business, such as risk assessment (King, 2013). In this way these organizations could improve their internal control processes, which eventually could lead to less work for the external auditors in the audit process (King, 2013).
The collection of big data is relatively easy, but it becomes difficult when processing and extracting useful information from huge amounts of data (Brown-Liburd, Issa, & Lombardi, 2015).
Accounting firms are competing each other to provide more and better comprehensive data analytical services to their clients, but there is still some uncertainty about the accomplishment of these services (Earley, 2015). Besides, there is limited information about how data analytics can help auditors in big-four audit firms to enhance their decision making (Alles M. , 2013).
As seen in the introduction, data analytics is the process of using unstructured and structured data with help of applications of several analytic techniques to provide useful information to decision-makers (Davenport & Harris, 2007); (Sims & Sossei, 2012).
Financial and non-financial data extracted from internet sites and mobile devices, social media and sensor-detected data are analyzed and used to make key business decisions (O’Leary, 2013); (Bertolucci, 2013). To include and process both structured and unstructured data to support decisions, auditors work with a new set of refined tools known as data analytics (ISACA, 2014)
Auditors should have the appropriate skills to interpret data analytics tasks since they are detail-oriented, trained to document all their work and have a lot of experience in making judgments in general (Schneider, Dai, Janvrin, Ajayi, & Raschke, 2015). Auditors with good skills in executing data analytics have greater opportunities to conduct an audit. If they are very well skilled they could audit more effectively and efficiently and hence, improve the quality of the audit process (Vasarhelyi, Alles, & Williams, 2010).
Early adopters of data analytics include marketing, social networks and other specific information systems that contains unstructured data. Today, data analytics creates the opportunity for auditors to generate value (Chen, Chiang, & Storey, 2012). Data analytics includes tools that make use of current technologies to extract and analyze information (Jordan, 2013). As the systems and technologies change and the nature and volume of data increases, data analytic tools will evolve as well (Bertolucci, 2013).
Data analytics can also help predict future sales demand or stock performance and thus aim key decision-makers in an organization. Finally, data analytics can be used in assurance tasks such as continuous monitoring or auditing (Vasarhelyi, Alles, & Williams, 2010); (Yoon, Hoogduin, & Zhang, 2015); (Zhang, Yang, & Appelbaum, 2015).
As discussed next, use of data Analytics in the audit presents significant opportunities, but also significant challenges to firms.
2.1.2 Benefits data analytics in the audit field
Financial products in combination with data analytics can give a good overview of the future needs of customers and a good estimation of a breach in the system (King, 2013). With the arrival of big data, the quantity and diversity of information has been increased. This results in big data that provides auditors with enormous potential to boost the efficiency and effectiveness of an audit engagement (Brown-Liburd, Issa, & Lombardi, 2015).
2.1.2.1 Improve audit effectiveness
The ability to fully acknowledge the benefits of big data lays in the use of more advanced data analytics techniques which potentially improve audit effectiveness (Brown-Liburd, Issa, & Lombardi, 2015).
Moreover, data analytic tools used to analyze big data give auditors the possibility to incorporate and use both structured data and unstructured data, for example the general ledger or transaction data.
Firms are increasingly exposed to new risks such as, fraud schemes, operational inefficiencies and errors that could lead to reputational damage or financial loss (Littley, 2012). This means that these organizations have to adopt innovative solutions to assess and manage risks to enhance the performance.
If the processes of data analytics and are implemented correctly, it helps the external auditors to clarify and boost the audit process by increasing the detection of potential fraud schemes, increasing operational efficiencies and reducing costs (Littley, 2012). The implementation of data analytical processes are also becoming a way for companies to create value (Littley, 2012).
Audit approaches can be fundamentally improved and transformed when data analytics tools and techniques are used (Littley, 2012). Moreover, he states that traditional audit approaches are based on periodic processes that includes manually analyzing control objectives, assessing and testing controls, performing tests and sampling. These approaches only have a small population to measure the operational performance and control effectiveness (Littley, 2012). Furthermore, according to Littley (2012), a continuous audit approach uses repeatable and sustainable data analytics and this approach turns into more risk biased and comprehensive overview of relevant data. By using data analytics, firms have the capability to check every transactions which facilitates a more efficient analysis on a greater scope (Littley, 2012).
These technological changes also gives auditors the possibility to incorporate unstructured data like, Wi-Fi sensors, RFID-tags and social media to identify unauthorized transactions, patterns in behaviors or suspicious trends (Brown-Liburd, Issa, & Lombardi, 2015).
2.1.2.2 Detect fraud risks
Conventional audits have not always been efficacious in revealing fraudulent activities since, these checklists and standard planning programs are linked with less effective analysis and identification to test these fraud risks (Asare & Wright, 2004). On the contrary, technological advanced tools, like data mining and data analytics, have been found to be effective and can be used in analyzing and evaluating big data in detecting of fraudulent risks (Humpherys, Moffitt, Burns, Burgoon, & Felix, 2011); (Bochkay & Levine, 2013).
One example from Cao et al. (2015) is about the SEC that is investing in big data analytics applications to seek out financial statement fraud and trying to identify audit failure. The SEC rolled out a system in 2013 that collected one billion records a day from private feeds of 13 national equity exchanges, time-stamped to the microsecond.
This data is extremely large, which is challenging the process to give the right output while requiring specialized expertise of data processing. Also, data analytics make it possible to analyze all data rather than just a sample. This can lead to models which are more reliable and robust.
Another example could be when the auditor wants to seek what attributes of journal entries are indicators of fraud or errors. In this case it is possible to analyze the entire amount of journal entries and use this information to analyze which of the current journal entries are actually unusual. Without data analytics auditors had to work very carefully to remove the infected data (Cao, Chychyla, & Stewart, 2015).
Auditors are also capable of performing high level analytics and can monitor an audit trail of the continuous auditing system to identify inconsistencies of fraud by higher management (Chan & Vasarhelyi, 2011). For example, fraud risk assessment is a highly complex and subjective judgment that auditors are required to make on every audit.
When analyzing transaction detailed information, auditors have used data analytics to identify unusual data movement to spot unusual data flows, such as unexpected volumes of data, high-frequency transactions or duplicate transactions that could indicate fraudulent operations (Dutta & Tavawala, 2013). The common agreement was data analytics should be promising for the detection of fraudulent activities since software tools enable auditors to process large amounts of data in an efficient way while being implemented at low cost in audit firms (AICPA, 2013); (Earley, 2015).
2.1.2.3 More use of non-financial data
Earley (2015) mentions that another benefit can be the possibility to use non-financial data and external data to better inform audit planning and to more effectively conduct audits on those areas that require professional judgment.
Furthermore, auditors can predict future events so they should be able to inform their clients in a better way regarding making strategic decision making about their businesses (Earley, 2015). These are called predictive events. Non-financial data includes data that the company gathers internally, such as personnel data, customer data, marketing data and even data from social media. These types of data go beyond the types of financial statement evidence that auditors typically analyze. As stated by Alles and Gray (2014, p. 16), ‘‘The vast majority of data in big data is non-financial data”.
Auditors, in general, are skilled to collect financial data and using this data to identify relevant information and filter out the noise (Brown-Liburd, Issa, & Lombardi, 2015).
This causes a distinction between useful and useless data that becomes more challenging to identify due to the auditors’ lack of knowledge in analyzing unstructured data (Brown-Liburd, Issa, & Lombardi, 2015). They can improve their core skills to support the interpretation of nonfinancial big data smaller and more efficient (Brown-Liburd, Issa, & Lombardi, 2015).
Since there is no commonly accepted formalized audit data standard, audit firms may either develop their own quality control standard or adopt the quality standards which are currently promoted by the American Institute of Certified Public Accountants (AICPA), about the use of big data (Brown-Liburd, Issa, & Lombardi, 2015).
2.1.2.4 Audit quality
Nowadays there is a lot of complexity presented in the audit scope and processes which means that audit firms being constantly challenged to improve the audit quality while being more cost-efficient (Mat Zain, Zaman, & Mohamed, 2015). Large implementation of data analytical tools has been slowed down due to the lack of efficient technological solutions and problems with data captures. Moreover, even though the accounting profession has already recognized the impact of data analysis on improving the quality and the relevance of audits, there are still other concerns about the privacy of data (Ramlukan, 2015).
The auditors’ judgment and own interpretation of the auditor are involved in the audit services conducted in the process (Brown-Liburd, Issa, & Lombardi, 2015). They also show that data analytic tools can be used to streamline data and assist in making decisions. However, without the auditors organizing and appropriately applying the information that is not covered, analysis would not be advantageous and audit quality would not be improved (Brown-Liburd, Issa, & Lombardi, 2015). These advantages can be created by operational efficiencies, reducing costs and detecting potential errors while improving the quality of the audit process (Littley, 2012).
Moreover, the actual presence of a continuous audit system connected with effective controls should translate into lower audit risk and should lead to higher audit quality (Malaescu & Sutton, 2015).
2.1.3 Challenges data analytics in the audit field
2.1.3.1 Information overload
Information overload means that the auditor simply receives too much information (Eppler & Mengis, 2004). This could be both financial information like, large data files with balance sheet items, and non-financial information which causes this information overload (Eppler & Mengis, 2004).
In that case, decision makers do not have the ability to process a huge amount of information. On top of that the existing research is not capable enough to analyze multiple sources consistently which shows that the outcomes will be not eligible (Benbasat & Taylor, 1982); (Iselin, 1988); (Kleinmuntz, 1990). Auditors can feel pressurized by regulators to demonstrate audit quality through compliance with the firm’s methodology, policies, and auditing standards (Bamber & Lyer, 2002) (Francis, 1994).
Information overload shows how the performance of an individual differ with the amount of information the auditor is exposed to regarding adequate decision making (Eppler & Mengis, 2004). Eppler and Mengis (2004) found that the quality of decisions of an individual auditor is positively associated with the amount of information they receive up to a certain point. If the auditor receives more information than this point the performance and ability to make rational decisions rapidly declines (Eppler & Mengis, 2004).
Big data can be reliable since it is often generated externally (within the clients firm) and acquired by auditors directly (Schneider, Dai, Janvrin, Ajayi, & Raschke, 2015). However, they found that noise in big data can cause an overload of false positives, which could lead to a lower reliability. Also according to them, it could be critical to eliminate this noise (veracity) for the audit objective. Evidence gathered by analytics regarding big data provide more unique and timelier information than traditional sources (Russom, 2011).
2.1.3.2 Information relevance
Increased amount of information could make it difficult to accurately identify relevant indications or useful information and may result in a lower performance (O’Reilly, 1980). Moreover, information overload makes it more difficult and more complicated to make good judgment in the audit process since this could lead to difficulties in identifying relevant information (Russom, 2011). This could influence the decision-making capabilities of the auditor (Brown-Liburd, Issa, & Lombardi, 2015).
Another negative aspect could be the level of irrelevant information. Higher levels of irrelevant information has been shown to reduce the decision makers’ ability to identify relevant information which also leads to a reduction of their overall decision-making performance (Hodge & Reid, 1971); (Streufert, 1973); (Well, 1971).
2.1.3.3 Pattern recognition
Big data provides the decision maker with the ability to search for particular patterns in large populations of data that would contrarily be undetectable in samples or smaller data sets (Alles M,
While the output is generated from the analysis of big data is a large advantage, prior research has shown that auditors do not adapt recognizing patterns in both nonfinancial and financial data (Bedard & Biggs, 1991). Auditors may be incapable to recognize inconsistencies in financial data and may evaluate on an individual basis as opposed to evaluating these inconsistencies combined (Bedard & Biggs, 1991). In the audit environment it could be that the audit risk assessment process involves recognizing patterns in the data that may suggest errors or fraud generally, for example with inconsistencies or complex data files (Gray & Debreceny, 2014).
Next up, Ambiguity can also be an issue, leading to different outcomes using the same information.
2.1.3.4 Ambiguity
Another aspect of big data is the unstructured nature of the data that comes in many formats, for example: text, images or videos. A lot of businesses are overpowered by the quantity of raw unstructured nature of data and they do not know how to get value out of it (Brown-Liburd, Issa, & Lombardi, 2015). It could be challenging to determine and extract relevant information that has to be addressed for the audit process (Brown-Liburd, Issa, & Lombardi, 2015).
Ambiguity arises from variations in the amount and type of information that is available, as well as from the different sources or the lack of knowledge of events that took place (Einhorn & Hogarth, 1986). Auditors need to develop a higher bound for ambiguity when using semi-structured and unsemi-structured big data since there could be different interpretations of information that has the same meaning (Brown-Liburd, Issa, & Lombardi, 2015).
2.2 Expertise auditors
The AICPA encourages universities to adopt new programs in the accounting curriculum. This ensures accountancy students to become acquainted with new technologies and get a better view of different auditor insights. This new accounting program should include, data sharing, data mining, data reporting, data creation, data analytics and storage in and throughout organizations (AICPA, 2013).
Technological innovations like, electronic commerce and online transactions have led to a huge increase in the volume and complexity of accounting transactions, making it more challenging for auditors to evaluate those transactions (Brown-Liburd, Issa, & Lombardi, 2015).
The increased complexity with big data translates into increased costs for companies. As the volume of the data rises, the costs of this technology also moves upwards and putting pressure on firms to hire more data analysts and invest in more data software (Golia, 2013);(Inbar, 2013).
Auditors that are well-educated regarding data analytics will have greater opportunities to conduct and successfully execute an audit (Brown-Liburd, Issa, & Lombardi, 2015). As seen with the use of non-financial data, auditors are trained to collect, organize, and analyze financial information, and they can improve their skills to work more structured with big data (Brown-Liburd, Issa, & Lombardi, 2015).
In fact, adequate training and skills play a critical role in adopting analytical tools. When auditors are skilled and have the knowledge to use data analytics, they could audit on a more frequent basis and that should improve the audit quality (Vasarhelyi, Alles, & Williams, 2010). Big data has the possibility to change the decision and collection of audit evidence in a serious way (Moffitt & Vasarhelyi, 2013); (Issa, 2013); (Kogan, Vasarhelyi, & Wu, 2010); (Vasarhelyi & Halper, 1991). For example, an auditor could use data analytical techniques to examine all client transactions for trends, exceptions and outliers (Brown-Liburd, Issa, & Lombardi, 2015). Furthermore, auditors now use data analytics to understand and evaluate security and controls over clients’ large datasets with help of data analytical tools, like data mining (Cao, Chychyla, & Stewart, 2015) (Vasarhelyi, Kogan, & Tuttle, 2015).
Ramlukan (2015) states: when the auditor gets to this point, they need to find the appropriate balance between applying judgment of the auditor and relying on the results of these data analytics. This will require them to develop new skills concentrated on the ability to know what questions to ask about the data along with the ability to use analytics to produce and gather audit evidence (Ramlukan, 2015). It requires an initiative built from the ground-up to improve the understanding and influence that the education has on students at their universities. It should enhance the learning and development progress and the establishments of appropriate implementation to support audit teams to integrate big data and analytics into the audit effectively (Ramlukan, 2015).
Research has shown that the use of technology does not always lead to better performance since auditors sometimes already have certain biases (Arnold, Collier, Leech, & Sutton, 2004). Auditors who are making decisions based on data analytics could develop a “computer is correct” attitude while using technology which could lead to a biased image of the true and fair view (Tang, Hess, Valacich, & Sweeney, 2014). One example is that the financial statement analysis can be increased by interactive reporting yet this can also lead to overconfidence and user-misunderstanding (Tang, Hess, Valacich, & Sweeney, 2014). The results of the audit process are of good quality when the initial input is right (Schneider, Dai, Janvrin, Ajayi, & Raschke, 2015). This can even occur when data analytics includes analyzing large volumes of data (Selby, 2011).
Selby (2011) also found that auditors are better able to interpret risk patterns in the continuous auditing environment when they have more knowledge about the automated control procedures of their clients.
Data analytical tools are currently being used to mitigate difficulties that auditors encounter with the recognition of these risk patterns (Thiprungsri & Vasarhelyi, 2011). An example of this could be a cluster analysis, which is an explorative analysis that tries to identify structures within the data. This has been extensively used in marketing and in the insurance industry and now in the auditing field (Thiprungsri & Vasarhelyi, 2011). Opportunities to incorporate these tools in the audit environment have upside potential, even though these tools have not been incorporated widely.
Cluster analysis can be used in large populations of transactions to identify specific cases where managers give authorization to exceed the limit of payments. The increased complexity with big data translates into increased costs for companies. As the volume of the data rises, the costs of this technology also moves upwards and is putting pressure on firms to hire more data analysts and invest in more data software(Golia, 2013); (Inbar, 2013).
In sum, standard setters are trying to inspire universities to adopt new audit programs (AICPA, 2013). Huge increases in the complexity of accounting evidence is the result of technological innovations (Brown-Liburd, Issa, & Lombardi, 2015).
Selby (2011) found that auditors are better able to interpret audited activities when they have better skills about IT-procedures. However, other researchers have shown that using technology not always lead to better results as they could have their own bias (Arnold, Collier, Leech, & Sutton, 2004). Moreover, auditors that make decisions based on technology could lead to overreliance of the system, creating an “computer is always right” attitude (Tang, Hess, Valacich, & Sweeney, 2014).
Yet, auditors who have the appropriate knowledge about data analytics will have better perspectives to carry out and conduct an successful audit (Brown-Liburd, Issa, & Lombardi, 2015). Furthermore, Auditors are working structured with big data and this is the result of improving their skills. In the end, auditors are trained to collect, organize and analyze financial information where training has a critical role (Brown-Liburd, Issa, & Lombardi, 2015). Auditors need to find a middle point between applying judgment and relying on the results of data analytics (Ramlukan, 2015).
3 Theoretical framework
3.1 Formalization framework
In this research I will use the framework of Adler and Borys (1996), the so-called Formalization framework. This framework provides a general approach to understand how a structured system, like data analytics, can take an enabling or coercive form (Free, 2007). Formalization framework incorporates an organization’s rules, policies and procedures. How formalization affects the attitude of employees, in this case the selected interviewees, which takes an enabling or coercive form by assessing the experiences with data analytics in the audit process. These attitude characteristics can take an enabling or coerce form of behavior (Adler & Borys, 1996). The framework of Alder and Borys (1996) provides a theoretical structure for interpreting the experiences of data analytics as enabling or coercive in the audit process.
In this research I would like to analyze the reactions of individual auditors and auditors in training about their experiences with data analytics in the audit process. In addition, I will conduct them to think about how to improve this process and to think about how they can gain knowledge to make better decisions in the audit process. In this research I will classify each category of the formalization framework into one of the four design principles that determine if formalization takes an enabling or coercive form in terms of behavior. These categories are: repair, internal transparency, global transparency or flexibility.
Enabling formalization stimulates employees’ skills and knowledge and provides guidance to help employees to do their jobs more effectively (Adler & Borys, 1996). This can be achieved by mastering their audit tasks, dealing with inherent contingencies and identifying systems and rules within the organization. Auditors want to explore the most efficient and effective way to execute data analytics in the audit process. An audit support system like data analytics express the rules, systems and believes of the audit firm (Dowling & Leech, 2014).
Coercive formalization generates “a foolproof system” (Ahrens & Chapman, 2004) that demonstrates how employees should perform their tasks and restricts their self-regulation (Adler & Borys, 1996). This means that auditors are striving to diminish as much as rules as they can (Dowling & Leech, 2014). Moreover, they found that these rules present the manner and extent to the audit process and how this gets formalized.
3.1.1 Repair
The assumption underlying repair is that certain events and activities are not easy to program and to standardize (Dowling & Leech, 2014). The difference between an enabling and coercive form of repair is how data analytics support actors (auditors) to resolve contingencies, uncertainties, or breakdowns of the data analytics system (Dowling & Leech, 2014).
It is assumed that these auditors do not have the required skills or knowledge to identify or resolve data analytical issues which are not supported or examined during the audit process in the coercive system (Dowling & Leech, 2014). Enabling systems provide the possibilities to facilitate actors while resolving unexpected concerns. Data analysis is becoming more important and more popular for audit firms through the years, as seen in the introduction (Murcia, de Souza, Borba, & Ykeda, 2007).
One aspect of highly usable systems is the convenience to which users can repair the system themselves, rather than allowing a breakdown to force the audit process to stop (Adler & Borys, 1996). In addition, it would be more preferable to use data analytical tools in an audit system with built-in online “help” facilities, rather to interrupt the flow of work with problems by forcing users to examine a manual or a supervisor (Adler & Borys, 1996).
3.1.1.1 Coercive repair
In the coercive repair philosophy, any inconsistency from the standard audit procedures is seen as suspicious. In this philosophy, data analytics are constructed to inform audit leaders when subordinates’ behavior and activities are in compliance.
Data analytics are not created to support team members determining whether the process is operational. Also, it is not developed to support users to explore the inevitable contingencies of the work process, nor to support them to identify and seek for opportunities for improvement within the audit process. Data analytics is created and developed to clarify superiors whether staff members are in compliance with the rules and regulation of the audit firm.
3.1.1.2 Enabling repair
The enabling repair philosophy creates an environment of procedures that simplify reactions to actual work contingencies. System breakdowns and repairs indicates the firm about issues with the formal procedures regarding data analytics and become opportunities to learn from that process and improve it. In this research repair is defined as how data analytics supports auditors to identify and cope with issues related to the behavior and documentation of audit activities.
- Coercive data analytics will necessitate auditors from dealing with unforeseen issues and would not provide tools to support the identification, analyzation or resolution of issues. - Enabling data analytics will help auditors to identify, analyze and resolve unexpected
issues and therefore provide tools to enable auditors to resolve these issues.
3.1.2 Internal Transparency
Internal transparency enables auditors and auditors in training, to develop an understanding of the functioning of data analytics in the audit process (Ahrens and Chapman 2004). Data analytics can be designed to minimize the reliance on the skills of individual auditors since the system does all the work. This can result in unnecessary explanation about the internal functioning of the data analytical system.
3.1.2.1 Coercive Internal Transparency
Coercive internal transparency is formulated as a list of vague guidelines to users of data analytics in the audit process. These guidelines are not developed to support employees as much as their supervisors. They do aim to help employees to be as effective as possible, but these explanations are only created to punish employees when they do not follow the data analytical process within the audit process. They do not support employee’s input as much as their supervisors. In this system the process is designed to sanction employees in case of deviations.
3.1.2.2 Enabling Internal Transparency
Enabling internal transparency foresee auditors with the sight and knowledge of local processes they are regulated by like, data analytics. This can be achieved by highlighting key parts, components and rules of the audit process. In this way, users will be provided with an explanation of the basic theory of data analytics in the audit process. The enabling internal transparency also provides users with feedback on their execution by providing metrics that users can help to compare their activities based on results from the past. Furthermore, enabling transparency enables the members of the firm to intensify their teamwork with other specialist. We define the data analytical process as the local process. Likewise, one of the conditions, if you want to work with data analytics is that the user should understand how data analytics work.
- Coercive data analytics will not facilitate transparency. A highly coercive system would divide the work of each participant and only appraisers would have a comprehensive view of the process of data analytics, which means that the activities are not trackable by staff members
- Enabling data analytics supports the understanding of auditors about their tasks
3.1.3 Global transparency
Global transparency “refers to the understandability for employees to comprehend the broader system in which they are working” (Adler and Borys 1996). The scope of the broader system is defined by the audit process whereby data analytics is a part of it. Internal transparency address to the internal workings of data analytics and the process as used by the employees. Global transparency address to the rules, guidelines and procedures of data analytics referred to the employees within the audit process.
3.1.3.1 Coercive global transparency
In the coercive approach of process design, global transparency for users of the system is a risk that should be minimized. Audit activities are allocated among employees and if employees do activities which are beyond their responsibilities they would be called back. The procedure itself is quite vague for staff members. Employees in a coercive global transparency has no plain idea about the evaluation of the process, criteria, and if the process or certain tasks are rejected, why they are rejected. They do not receive feedback on their audit activities.
3.1.3.2 Enabling global transparency
In the enabling approach of process design, employees are accommodated with a broad range of contingent information designed to support them to interact with an immense organization and environment. The process is designed to allow them to understand what their own tasks are and stimulate learning. This should encourage numerous suggestion that are brought up by employees. This results in added value which is clarified in great detail to all users of data analytics in the audit process. Explaining the evaluation criteria of employees is seen as an opportunity to improve the understanding of the audit process and to improve the audits effectiveness.
This study defines the whole system of interests as the firm’s policies and methodologies that all audit engagements should comply with.
- A coercive system would force compliance about data analytic without facilitating understanding the use of it in the audit process.
- An enabling data analytics system facilitates global transparency by promoting the use of data analytics and the auditor’s knowledge about the firm’s audit process.
3.1.4 Flexibility
Flexibility is the extent to which the data analytics provides actors with options and possibilities in how they use data analytics in the audit process. Flexibility provides choice for predictable issues that actors could address. Auditors can overrule the system’s guidance by documenting their motivations and reasons in the areas which can be used for that in the data analytics system. And also by gathering additional audit evidence.
A systems flexibility boosts the users to use the system in different ways. Employees can adapt the interface and add functionalities of several tools to fit their demands.
3.1.4.1 Coercive Flexibility
Coercive flexibility defines in extensive detail what the specific steps are in the audit process and how these need to be followed throughout the audit process. This creates a compulsory approval which needs to be asked for by superiors. Every process is unique in their way, which means that if steps needs to be skipped in the process, approval is needed. The assumption within this type of flexibility is a three-way approach.
The manual has a description that has to be executed by employees which again, should ask for authorization by their supervisor. Within this type of flexibility there are boundaries established to minimize the deviation of the audit process.
3.1.4.2 Enabling Flexibility
Enabling flexibility beholds the process more as a learning process whereby employees that deviate from the regular process are not only at risk, but also learning from these deviations to find opportunities to improve their effectiveness (Borys, 1992). Overreliance by auditors on data analytics is a common response when the system inflicts controls that are developed and implemented before usage (Majchrzak, Rice, Malhorta, and King 2000).
4 Methodology
The research methodology of this study will be described in this section. There are two types of research namely, qualitative and quantitative research. The objective of a qualitative research is to create a deeper understanding about the underlying reasons, beliefs, habits and motivations and to figure out what is happening in the field. The objective of quantitative research is to quantify data. The selected theory and literature are used to deduce the hypothesis. In a qualitative study the data can be collected with the use of semi-structured measuring instruments that are customized to the research subject and this can be clarified throughout the research process.
This research will be analyzed by using a qualitative research with a case study. To answer the main research questions, interviews are conducted to gather data. The structure of this paragraph is as follows: In the first paragraph the research setting is discussed. Next, the justification of the chosen method will be explained as well as the data sample. The last paragraph describes the data analysis.
4.1 Research setting
Bryman (2008) defines qualitative research as: “A research strategy that usually emphasizes words rather than quantification in the collection and analysis of data” (Bryman, 2008).
As mentioned earlier, the purpose of the qualitative research method is to gain a deeper understanding of the underlying reasons regarding habits, motivation, reasons and behavior of actors (auditors) in the field. The main objective of this study is to gain a profound understanding about how auditors use data analytics in the audit process and how they base their decisions on these data analytics gathered in the audit process.
In this study, I investigate the different perspectives on the notion of data analytics in the audit practice. By employing qualitative research, the experiences of data analytics can be explored intensively by performing interviews. In this way it is possible to gather insights of the audit process and experience about the use of judgment, benefits and positive aspects regarding data analytics by auditors in the audit field.
In the previous parts of this study you can find the introduction, research question, literature review and theoretical framework. To conduct a field research, a case study has been conducted. According to Yin (2009) a case study grants the opportunity to answer how and why questions of a research question in more depth and detail. This case study will handle various objectives.
The main objective is to investigate the experiences auditors have with data analytics and how they use this to make decisions based on evidence gathered with data analytics.
The case study is performed within a big four audit firm in the Netherlands. The studied company will stay anonymous in this research and will be called Firm X for the remainder of this study. Firm X has been chosen as they have the proper (financial) resources to implement data analytics (Liddy, 2015). An internship has been conducted with the primary goal to write a thesis gives me the opportunity to get access to several auditors. The goal of these interviews is to gain a deeper understanding about the use, interpretation and experiences of data analytics in the audit process.
4.2 Data Sample
As mentioned earlier, this study will contain a qualitative case study as a research method since semi-structured interviews will be conducted. There are several reasons for these methods. First qualitative interviews have more potential to get insights into the perspectives of the interviewee. Quantitative interview reflects more the concerns of the researchers. Also, qualitative interviews can be conducted in a more flexible way.
Interviews can differ significantly from with the interviews which are following-up. Moreover, after interviews the researcher can ask new questions and can obtain authorization to get follow-ups via email. Therefore, interviews have more promising details conducted in their answers than from quantitative interviews (Bryman, 2008). The data for this research is created from the interviews with several auditors. The interview provides series of questions which can be asked. This does not have to be in a specific order, since these interviews questions will be semi-structured. This also means that some questions will be removed changed during the interview process. In conducting my interviews, I do not make any separation between male and female interviewees.
However, the auditors must have experience with data analytics in the audit process. The auditors interviewed were at the level of senior staff, managers and director; with three to fifteen years of experience in the audit practice. Interview duration will be around one hour. All interviews were taken in Dutch. As the interviews continue, other (new) questions came across which were addressed in the following interviews. All the interviews are tape recorded and transcribed later.
Before interviewing all the participants has received an email where is asked if they were willing to participate in the interview. Firm X has employees with different job grades. This can
Also, the questions which are formulated before the interview were focused on the aspects: repair, internal transparency, global transparency and flexibility. These themes made it possible to ask questions that are related to the so-called Formalization framework on which this study is based.
At the start of each interview, a plain explanation of the interview objective has been discussed with the interviewees to arrange them with basic information about different interview topics. These topics are: traditional auditing, continuous auditing, big data and data analytics.
I have conducted all interviews until no new or other issues were arising from the interviews when they were conducted. Ten individual interviews were conducted with professionals which are part of the audit firm. The functions of these professionals vary between several levels. In addition, since data analytics is a novel concept in the audit field, this is not fully incorporated into firm X.
Therefore, all participants who are interviewed are part of the project which is called Project-W. The project is an initiative of the audit firm to increase the demand of their clients and convince them that data analytics will be the future. Furthermore, the project is a pilot to implement data analytics in the future during every audit. A briefer explanation of can be found in the case context paragraph. The participants which are interviewed are from different levels in the audit firm. These levels will be referred in numbers from number one until number six.
This has resulted in ten individual interviews across different locations of the audit firm, all taken up among 35 and 80 minutes. Nine out of ten interviews were taken face-to-face. One interview took place via a telephone line since the interviewee was full time stationed at an external location. The language of these interviews is Dutch, however the transcriptions are fully written in English. All interviews were recorded after permission of the interviewee. An overview of the interviewees can be found in Appendix A.
4.3 Data analysis
Data analysis has been conducted manually. The transcribed documents were fully written by me. It resulted in a document of 105 pages. The transcriptions were made completely anonymous before printing them out and analyzing them by eliding the names and specific names of certain tooling. Also by coding any references that were aimed to Firm X were removed. The coding was done manually with several colors of markers and a pen. The decision was made to do it manually resulting in a better feeling about the depth of the data collected. The transcriptions were consulted two times to get a good in-depth thought about several aspects.
The data is analyzed by the process of analysis which embraces three sub processes: Data reduction, data display and conclusion drawing (O’Dwyer, 2004).
Data reduction was conducted in several steps, including: read relevant interview notes, revisiting the interview notes, developing an open coding scheme and preparing critical summaries. The coding process also includes; managing, identifying, sorting, arranging, transforming and grouping the data that was retrieved from the interviews. During the initial analysis it was important to gain deeper and broader insights of the subject and the reactions of the respondents while trying to create an understanding of the bigger picture. Thereafter, in order to focus more clearly on the interview data, a profound and detailed coding process was conducted.
Data display includes grouping several codes into themes and these are used for the conclusion drawing steps. This process consists of creating, organizing and arranging themes to support the analyzation and conclusion drawing. During this process of data displaying some themes were merged and others were removed, since they fit other themes. An overview of this can be found in Appendix B.
In the process of conclusion drawing there will be searched for patterns, contradictions, relations, explanations and complements to find results which can support this study.
5 Case context
The case study was conducted at a big four audit firm, which is referred as Firm X. The data enabled program started last year, and this year is their second year of the pilot. According to the data enabled program manager:
“How I see the future of auditing. That consists of an era where annual report as a report physically does not exist anymore. On the website of our customers the annual report of the customer will be presented real-time and where you have a continuous update which is constantly checked by a system. At all real-times there will be a certification with: verified by the audit firm. That should also mean that we need to create algorithms which continuously audit and validate the bookings of the clients. And when there are unusual patterns in those transactions, the system should recognize those and send them to the auditor for validation”. (I9)
The audit firm created a data enabled audit pilot which contains several clients that are selected by the firm to participate in the pilot. This is called Project-W. This project explores the implementation of data analytics in the audit process and how this is perceived by auditors. There are several coaches and audit teams participating in this project. These coaches are selected by the program leader and focus on supporting audit teams to carry out data analytics in the audit process. Furthermore, these coaches monitor the usage of data analytics by audit teams in the audit process and receive feedback from them about their experiences with data analytics.
A huge amount of auditors has joined this program which combines several auditors of the firm with different expertise. This is allocated over several special departments like, the methodology department, tax department, valuation department, audit department, information technology department and the profession department.
Two years ago the data analytics in Firm X was nearly unknown. With the help of several researches and initiatives of the data enabled audit leader the usage of data analytics is being expanded by the years. It is expected that data analytics will be used more and more extensively throughout the coming years.
The focus of the upcoming years with the data enabled audit program is to educate the personnel of the audit process to introduce them to a new way of auditing.
“It is not only about: here you have the tool, you can access it here and good luck with it. No, the program is more skills driven, it is all about the skills of your audit personnel. So we need to educate our people to improve their analytical thinking and also to teach people to recognize patterns out of a lot of data”. (I10)
6 Case analysis
In this paragraph we uncover the findings of this study. These are organized as follows. Paragraph 6.1 describes the repair section of the formalization framework. Paragraph 6.2 outlines the internal transparency section of the formalization framework. Paragraph 6.3 illustrates the global transparency section of the formalization framework. Paragraph 6.4 clarifies the flexibility section of the formalization framework.
During the process of analyzing and coding the interviews it became clear that there is a wide range of experiences of data analytics in the audit process. Most of them are positive, however there were also some negative aspects which are mentioned in this case analysis and which should be taken into account.
6.1 Repair
Data analytics is analyzing data using IT-systems and different tools. In some cases, these tools can also give a lot of issues, since there are a lot of bugs or the system is too complex. So if there are issues, these need to be resolved. Often it can be resolved by the auditor himself, however there are some situation whereby the auditor has to assess a specialist to work with the tool or system. Besides, auditors often need to ask the client for clarification about certain data. This can be the case when the data is corrupt or is not interpretable. Because data analytics is a novel concept there are still a lot of bugs when carrying out audit activities. If the data is not interpretable or there are issues with the data, then the auditors falls back on their old methods, like the test of details and test of controls.
“That depends on the complexity of the system. Sometimes we do it ourselves and sometimes we bring in to the IT-Audit department to obtain the data translated in readable form and eventually for the analysis. These will be prepared by, for instance, a by colleague IT-Audit or we will be preparing it ourselves”. (I1)
Moreover, it could happen that the administration does not match the data. Auditors could question the workings of the processes then. On the other hand, when there are issues which need to be resolved, it also means that we can clarify it for the client. Another interviewee substantiated this, but acknowledged that it is not only the tool that needs to help Firm X to
“Well, there is also the link: the tool is not often the issue, it is more about how are you working with the tool, how do you handle it. Well, if I think about the typical issues that I encounter. If I involve the tool, we are unfortunately dealing, since the tool is actually quite new, quite often with a lot of bugs and errors. But on the other hand there are not much more uncertainties.
The only thing is that there are bugs, and that the system could get stuck, that kind of issues. But the use of the tool, also what does the user think, against what does the tool actually does. That is a little bit of the perception of the user. Does the user also know audit technical how to use at one hand the analyzing of the data and at the other hand the translation of the results of that analysis in the audit tool. Sometimes auditors define certain risks and then these risks are sorted by data specialists.
Often we learn to know our clients much better with the help of data analysis. Because suddenly we can see everything of the clients in those systems. But often we also learn a lot of the system. That we suddenly identify and recognize wait, their sales system is supported for years by different applications. So then you can go more in-depth”.(I10)
To summarize, Firm X is still developing data analytics, and therefore they have a pilot which examines the workings and experience with it in practice. In about two years they want to standardize data analytics in the audit process with every audit. Since it is a novel approach, auditors still experience a lot of bugs. A lot of these bugs they can be resolved by themselves. Yet, there has been issues which they could not resolve and therefore, they had to consult IT-specialists.
They do have the skills to resolve several contingencies, however if there is a system breakdown auditors prefer to call an IT-specialist. This saves a lot of time, since specialized people have more knowledge about those systems. The direction with data analytics with Firm X is to improve the skills and knowledge of auditors. This could mean that the perception of the users is crucial in the execution of the audit regarding data analytics.
Every step in the audit process is documented in the audit file. This means that not only supervisors can see what is done, but also staff members themselves. If the responses of the respondents are reflected to the theoretical framework, we see that in this case repair takes an enabling form. If the data cannot be interpreted, then auditors use their test of details and test of controls.
6.2 Internal Transparency
To improve the knowledge of staff members, they are trained to work with data analytics and get supported by several teams that help them achieving this. There are several coaches who are helping them to conduct an audit by using data analytics. These coaches are: The W-Coach, M-Coach, I-M-Coach, and audit manager. All have different roles in the audit process and also in supporting employees with this novel concept to execute it successfully. There are several meetings which are organized frequently. Also when there are questions about specific aspects they can ask it to them. In order to successfully implement data analytics, employees should clearly understand the purpose and functionalities of these techniques. The importance of the fundamental audit knowledge becomes clear when we ask several interviewees how data analytics can support their understanding of the audit process. One of the interviewees responded:
“Data analytics has different aspects. You can use it for the understanding of the clients’ environment and use that for your risk analysis. For example, how are the journal entries flowing through the organization? It can help with the test of details to get comfort on which auditors can rely. For example, journal entry testing to explore the exceptions and to identify and scan transactions on irregularities”. (I1)
However, not every respondent has the same view on the novelty that data analytics has in concept. It does not back the statement of understanding of the audit process. The respondent underpinned this in the following statement:
“In design, data analytics is nothing new. Data analytics facilitates to create and design the audit on a more effective and efficient way. Also, your understanding of the client is better, but that didn’t help me to understand the audit process better since that is not changed”. (I3)
Yet, the importance of basic knowledge becomes evident when analyzing the interviews. The importance of the understanding of the audit process and the fundamental knowledge that you need to have about executing data analytics, is underpinned in the following statement.
“I think that you need the technical fundament of the audit profession really hard at the moment that you are going to work with data analytics. We have a lot of auditors in this audit firm, that you are trying to learn something else than they already have learned in the past. So, at this moment, we are busy with a transition process to help them to learn new skills and also to work on a whole new way. And that is only possible when the fundament, the basics, are sufficient enough”. (I10)
