Data-Driven Control
Lecture 1; Introduction:
The world is changing into a more digital, technological world. Therefore, both the IT in Control
course and the Data-Driven Control course are relevant for future accountants. This changing world
brings an increased complexity; people have become more demanding; organizations respond and
react; all leading to growth, globalization, and interdependency in this changing world.
Following Moore’s law, the number of transistors on microchips doubles every two years with
minimal rise in cost. This also shows that the world becomes more digital and technological.
Other important technological developments in recent years are:
- A cheaper, faster, and higher capacity storage has been developed.
- An improved connectivity, through networking, internet, 4G/5G and Internet of Things (IoT).
- The development of programming languages.
These technological developments bring other problems with them for companies. Many companies
have difficulties with adapting to these technological developments, leading to digital disruption;
“research shows that since 2000, 52 percent of companies in the S&P500 have either gone
bankrupt, been acquired, or ceased to exist as a result of digital disruption.”
Increased Value of Data:
The value of data has seen a significant increase in recent years due to its critical role in driving
innovation, improving decision-making, and creating competitive advantages across industries.
The value of data is high when it possesses certain characteristic and is used in contexts that
maximize its utility and impact, meaning that its relevant and reliable. When the data is relevant
and reliable, it becomes useful and can support in decision-making processes. Data on itself is
nothing, but when you add for example a meaning to the data, it gets value. Data with a meaning
can become information.
The DIKW Pyramid model illustrates how raw data progresses through stages to become wisdom.
Raw data turns into meaningful information; extracting insights from this information can form it
into knowledge; and applying knowledge to make decisions turns it into wisdom. This model
emphasizes the role of analytics in providing context and understanding at each state, ultimately
increasing the understanding.
Context in this models shows that as you move upward, data is placed in increasingly broader and
richer contexts, slowly transforming it into meaningful insights and wisdom.
Understanding in this model shows that as you progress horizontally, understanding grows
through activities like researching, absorbing, doing, interacting, and reflecting.
Data = raw facts or figures without context or
meaning (e.g. a set of numbers, like “10,20,30”).
The key activity of data is researching (e.g.
gathering parts).
Information = data has been processed,
organized, or structured to provide meaning (e.g.
10 customer purchased 20 products on 30 days).
The key activity is absorbing data and beginning
to extract meaning from it (e.g. connection of
parts).
Knowledge = insights gained by analyzing
information to understand patterns, implications,
and causes (e.g. recognizing that customers
prefer products with discounts, which increases
purchases). The key activity of knowledge is
doing and interacting to test and refine insights
(e.g. formation of a whole).
Wisdom = the ability to apply knowledge in new
and novel contexts, making forward-looking
decisions (e.g. creating a strategy to implement
personalized discounts based on customer
preferences to maximize sales). The key activity
1
, is reflecting on past knowledge to generate
innovative solutions (e.g. joining of wholes).
The value of data can reduce when the meaning is wrong; so the information is wrong. In that
case the data is still correct, but the meaning is wrong, leading to a reduction in value of the data.
The value of data is also dependent on several drivers:
- Availability/Scarcity: a high availability of data drives immediate usability, leading to a
higher value of the data; especially in fast-moving industries (e.g. stock trading, logistics).
The scarcity of data provides a competitive advantage, as unique data can unlock insights
others cannot access, leading to a higher value of the data.
- Quality: high-quality data is more trustworthy and useful for critical decisions, enhancing
its overall value.
- Relevancy: relevant data enables targeted actions and avoids the noise of extraneous
information, making it more actionable and valuable.
- Timeliness: data loses value as it becomes outdated, especially in dynamic environments
where rapid changes occur.
- Reliability: reliable data provides confidence in its use, making it a cornerstone for critical
applications, resulting in a higher value of data. On the other hand, unreliable data can lead
to mistrust, incorrect decisions, or system failures, reducing its value drastically.
Improvements in technology have made data analytics (e.g. storing data and capturing data) more
accessible and affordable. As the cost to capture and store data decreases, the return on
investment (ROI) from analytics increases, driving greater adoption of data-driven strategies.
Data is part of the triangle that shows the relationship between people, process, and technology.
Data is placed in the middle of this triangle, meaning that its part of both people, process, and
technology. Having an understanding and adding context to the data eventually makes the data
more valuable.
The availability of data becomes more every year. Therefore, storing data is very expensive, but
storing the data also adds value to it. In the end it depends on how much you use the stored data to
determine whether storing it gets more expensive every year.
This simplified model has been made to show how decisions
regarding data can be made. It shares similarities with the empirical
cycle and the PDCA cycle.
In short, data can be used to make evaluations and eventually
make decisions regarding the data. These decisions lead to actions
and in the end to new data; resulting in the cycle starting all over
again.
2
,Data often moves in a process flow from data extraction to data mining. It often starts with
raw data extraction and querying for receiving the basic things regarding the data. After that,
data analysis starts to analyse for meaningful patterns and understanding the data. Ultimately, it
advances to data mining for predictive insights and strategic decision-making.
To move towards data mining, the process
also goes from simple tools to complex
advanced tools, which indicates the
increasing need for specialized skills and
resources as you move to data mining.
Moreover, you also move from a basic
understanding of past events to predicting
future outcomes.
Article 1; Big Data in Accounting: An Overview; Vasarhelyi et al.
(2015):
The article Big Data in Accounting: An Overview examines the transformative impact of Big Data on
accounting and auditing. It provides a comprehensive framework for understanding how the vast,
complex datasets now available are reshaping traditional practices, offering new opportunities, and
presenting challenges that need to be addressed.
Big Data is defined by its "4 Vs" — Volume, Velocity, Variety, and Veracity — each of which plays a
critical role in shaping its application:
- Volume: refers to the huge amount of data being generated, often too large for traditional
systems to store or manage efficiently.
- Velocity: the speed at which data is created and processed, like real-time updates from
sensors or transactions. Continuous auditing, for instance, leverages high-velocity data to
enable anomaly detection and exception-based auditing.
- Variety: the different types of data available, including numbers, text, images, videos, and
social media posts. This diversity expands the scope of accounting and auditing by
integrating traditional and nontraditional data sources.
- Veracity: the trustworthiness and accuracy of the data, which can sometimes be uncertain
or inconsistent. Ensuring reliability and addressing potential inaccuracies in large and
diverse datasets is a significant challenge for accountants and auditors.
A perfect example of Big Data is all the data regarding cars (weather impact, Tesla information,
emission output etc.). This is a lot of data, a lot of variety and velocity, but it also meets the
requirement of veracity, because you do not know if you have all data; it is uncertain and
unstructured.
In accounting, this Big Data enables a shift from traditional financial reporting to dynamic and
detailed business measurements. The integration of ERP systems and automated data capture
technologies has made it possible to trace economic activities earlier and more precisely than
before. Predictive analytics and real-time insights offer organizations the ability to enhance
decision-making processes, transforming how economic activities are monitored and reported.
Auditing is also experiencing profound changes with the adoption of Big Data. Continuous auditing
frameworks, supported by high-volume and high-velocity data, facilitate real-time monitoring and
the identification of anomalies(=afwijkingen/dingen die niet overeenkomen in de jaarrekeningen ).
The inclusion of unstructured data such as video, audio, and text streams further enriches the
assurance process. For example, RFID and GPS technologies can automate inventory counts and
provide robust verification of transactions.
Despite its potential, the integration of Big Data into accounting and auditing is not without its
challenges:
3
, - The computational demands of processing large datasets, coupled with the probabilistic
nature of nontraditional data sources, necessitate new methods for evidence collection and
decision-making; new methods for these are necessary for accountants and auditors.
- Ethical concerns, such as data privacy and security, must be addressed.
The article also highlights the need for changes in accounting education and accounting research to
fully leverage the benefits of Big Data. Accounting studies must be updated to include IT and
statistical skills, while accounting research must explore new methodologies to integrate Big Data
into accounting and auditing practices effectively.
In conclusion, Big Data is revolutionizing accounting and auditing by enabling real-time analysis,
predictive insights, and enriched assurance mechanisms. Its transformative potential is immense,
but it requires overcoming technical, ethical, and educational challenges to create a data-driven
future in these fields. The article underscores the importance of understanding and leveraging the
"4 Vs" to harness the power of Big Data in advancing the practice of accounting and auditing.
Article 2; A Design Science Research Methodology and its
Application to Accounting Information Systems Research; Guido L.
Geerts (2011):
The article A Design Science Research Methodology and Its Application to Accounting Information
Systems Research introduces the Design Science Research Methodology (DSRM) and its
relevance in the field of Accounting Information Systems (AIS).
DSRM = is a structured approach to conducting design science research in the field of Accounting
Information Systems.
Design science research differs from natural science research in its purpose and approach:
- Natural science seeks to understand and explain phenomena.
- Design science focuses on creating artifacts (e.g. concepts, models, methods,
frameworks, or systems) to solve real-world problems and improve existing conditions. For
an artifact to be valuable in design science, it must meet two key criteria:
o Relevance: solving significant problems.
o Novelty: providing unique or more efficient solutions.
The DSRM framework is a structured approach consisting of six steps:
1. Problem identification and motivation: define a specific problem and justify the need
for a solution.
2. Defining objectives: specify what a successful solution would achieve.
3. Design and development: create the artifact (e.g. a new model or method).
4. Demonstration: show how the artifact solves the problem in practice.
5. Evaluation: assess how well the artifact meets its objectives using metrics or tests.
6. Communication: share the findings and utility of the artifact with relevant audiences.
These steps ensure a systematic process for producing, presenting, and evaluating design science
research. The iterative nature of the DSRM allows researchers to refine their artifacts based on
feedback and evaluation, enhancing both rigor and utility.
Geerts highlights the application of DSRM in AIS by conducting a retroactive analysis of six
significant design science papers from the field. For instance, McCarthy’s seminal 1982 work on the
REA accounting model proposed a shared data framework to address limitations in conventional
accounting systems. This research followed several DSRM steps, including problem identification,
4
