Data Science in Auditing professor: ‘Just a tiny part of the information that’s available is being used’

23 February 2018

Accountants and auditors who fail to familiarise themselves with the potential of data science in auditing, are both their own and their employer’s worst enemy, says Edo Roos Lindgreen of the University of Amsterdam (UvA). ‘You have to know how to find and analyse information. Otherwise, you miss out on a lot.’

Beyond internal structured data

Structured data from within the organisation, such as data from HR or sales management systems, provide just a fraction of the information auditors and accountants can draw upon to test and improve processes, Roos Lindgreen emphasises during an interview at ABS. Accountants should look well beyond internal structured data, he argues.
Focused analyses of internal and external unstructured data can yield useful patterns one can take advantage of. Roos Lindgreen gives the example of a non-traceable analysis of email correspondence within an organisation: ‘It might signal whether a particular business unit is susceptible to fraud. Based on that, one can decide whether or not to tighten up controls.’ External unstructured data such as comments on social media can provide helpful information as to how customers view a product or image; external structured data obtained from analysts’ reports, for example, can give insights into the competition and market developments.
It may seem easy. However in practice, a mid-sized company considering entering into a business relationship with a supplier from - say Kazakhstan -  wanting to find out everything about that company and its owners is unlikely to get very far. ‘You have to know how to find information and how to analyse it.‘

Level 1 swimming certificate

To many auditors, that combination remains a bridge too far, says Roos Lindgreen, who initially was appointed in his position as UvA professor for one day a week last April. It prompted UvA to develop a Data Science for Auditors programme that includes some solid programming in the statistical computing language R. ‘The accountant of the future has to know how easy and powerful it is to work with data. For that, one has to be able to do a bit of programming.’
The programme will be part of the regular curriculum of the Postmaster Accountancy, the Executive Master of Science in Internal Auditing and the Amsterdam IT-Audit Programmes, but it will also be offered as a four-day master class to external parties starting this spring. Roos Lindgreen counts on at least 100 participants a year. ‘This extra income will enable us to hire more researchers. And there will be an exchange of ideas between professionals and the university.’

Although the use of data science in auditing offers many opportunities, the programme is not meant to turn accountants into real data scientists. The master class will be 'like a level 1 swimming certificate in data science for auditors‘, says Roos Lindgreen. He has no choice: during the pilot, last fall, there turned out to be two groups of students. A small, very active group, making suggestions and asking for more statistics. These people were delighted. But to most participants, the master class was difficult, especially the programming and statistics. ‘Their wish was: less and simpler. In practice, that will apply to the majority. We don’t intend to set the bar too high for participants, we actually want to lower the barriers to using data science.’

Research and training

Roos Lindgreen expects his job as programme director of the Internal Auditing and IT-Auditing executive programmes to take up around two days a week. The bulk of his remaining two working days at UvA will be spent mentoring and supervising PhD students. Currently, he supervises various research areas, including research into ways for auditors to test the reliability of algorithms that are being used, and research on innovation programmes related to data science. Another major research topic concerns automating parts of auditors’ work, including the use of artificial intelligence.
The topics fit with earlier phases of Roos Lindgreen’s career. He studied informatics at UvA and obtained a PhD at Delft University of Technology on research into the practical applicability of information security methods. Until October 2016, he headed KPMG’s national innovation program.

Looking ahead

Working at KPMG was satisfying but the university offered a more inspiring long-term perspective. While the emphasis at KPMG was on the commercial aspects, at the university Roos Lindgreen experiences more scope for the content and for in-depth study.  ‘Here my mind is engaged in my work in a different way. It allows me to satisfy my curiosity .’

He is adamant: accountants and auditors should also ready themselves for the future. Looking ahead, they simply cannot afford to ignore the opportunities offered by data science. He points to the increasing demands in the field of data science made by institutions such as the Institute of Internal Auditors, the professional body for accountants in the Netherlands (NBA) and the commission responsible for accountancy professional competence certificates in the Netherlands (CEA): ‘For the time being, you’re not yet in trouble because you’re not the only one but eventually, you’ll lag behind. Relative to your competitors as well as in the labour market.’

By Christine Lucassen

Published by  Economics and Business