Business analytics holds tremendous potential to create value for the industry and has been doing so for around two decades. Considering this, most of the business schools have introduced courses related to business analytics in their program curriculum. Some business schools have also started MBA or PG programs in Business Analytics.
As a faculty who teaches analytics courses, I have been learning from my experience and also interacting with the instructors at many business schools on how they handle the analytics courses, and how the students tend to learn these courses. I could find a few learnings in this journey and am sharing them with you through this write-up.
- Help students develop a strong foundation in mathematics, statistics, and operations research
First and foremost, it needs to be reinforced to the students that the body of knowledge in the three courses of mathematics, statistics, and operations research, is the mother of business analytics. So, the students need to have a very strong foundation in these courses to be able to appreciate business analytics in a true sense. Contrary to this, I have found many students to be thinking that business analytics is all about the programming and the tool; while it is actually about solving business problems by integrating the knowledge of business and mathematics.
- Emphasize on Solution Approach
Second, the instructors need to make the students appreciate that code is just the solution approach converted into a language that can be understood by the computer. So, the importance of understanding the solution approach cannot be undermined. This gets facilitated if the major part of the contact time with the students is spent on formulating and visualizing the problem than that spent on solving it.
- Teach analytics courses only after covering foundation in functional areas
Third, it needs to be taken care that the foundation courses in functional areas of management are thoroughly taught to the students before introducing analytics courses into the curriculum so that the students can appreciate the scope of analytics in solving business problems related to the functional areas. But, it is time that a little bit of analytics flavor is given to each course since analytics has touchpoints with every discipline. For example, in the HR course, at least a small exercise can be done on predicting employee attrition. In the corporate finance course, some exercises on risk-return modeling can be done. Similarly, in the marketing course, a small exercise on clustering or market basket analysis can be done; while in the operations management course, some exercise on supply chain design analytics can be carried out
- Focus on Business Applications, and cover a wide range of functional areas
Fourth, there exists a need to deliver analytics courses in a manner that focuses on the specific business applications rather than teaching the courses as pure mathematical courses. This can be done with the help of case studies. It is also important to cover the applications of business analytics in all the functional areas while delivering a generic course on business analytics. This is important for a wider variety of students to be able to appreciate what business analytics has to offer in the area of their interest.
- Adopt Hands-on Pedagogy
Fifth, hands-on pedagogy is the best approach to teaching analytics courses. There needs to be a focus on learning while doing here. When the students solve the problems themselves, they not only absorb more but are also likely to feel more confident and cool when faced with the varying versions of the problem in the industry. Here, the instructor also needs to take care that he goes to the class with a good amount of preparation, and preferably with a written code so that the syntax errors do not consume the precious class time or distract the class flow.
- Use contemporary datasets
Sixth, most of the time, the stale datasets are used in the classroom. It is obvious that if the latest datasets are taken in the classroom, the students would be able to relate more to the topic. Or, if the contemporary developments are converted into case studies, the engagement between the students and the instructor increases. The instructors need to be ready to take out time for this exercise as it requires them to prepare more for the class. Also, if the dataset is taken from a public source in the classroom, and processing is done in front of the students, the students can appreciate the end-to-end cycle of data analytics. Although there is a caveat that this may take a little more of the valuable classroom time and that the first-time data may not yield the desired learnings in the classroom.
- Encourage diverse opinions in the class
Seventh, it is important to realize that business analytics is not pure science but also an art. Therefore, diverse ways of approaching a problem, or interpreting an output generated by the software, need to be encouraged in the classroom. The students also need to be taught that there exists a trade-off between finding the perfect answer and finding the close-to-perfect answer.
- Eliminate the coding phobia among the students
Eight, given that many students are mentally averse to learning how to code, it is the responsibility of the instructors to make the students realize the importance of the same, and help them start developing codes for solving business applications. Today, we are in a world where the boundary between the technical role and management role has been vanishing. In such a scenario, the students who are averse to appreciating the programming codes will be at a disadvantage when they join the industry.
- Cover prescriptive analytics as well
Ninth, while most the business schools are covering descriptive analytics and prescriptive analytics in the course curriculum adequately, it has generally been observed that prescriptive analytics is an area that is being given lower attention. But prescriptive analytics is very important for businesses to make the best use of their scarce resources. Therefore, the students need to be given substantial exposure to prescriptive analytics at least in the functional area of their interest. For example, some portfolio optimization analytics can be done for finance students, or optimization of the marketing expenditure can be taught to marketing students. Similarly, optimization of workforce scheduling can be taught to HR students; or vehicle routing problems can be done for the operations domain. And in case the students can run a few simulations or virtual games in each of these domains before taking the analytics route, that can help them appreciate the concept of trade-offs in the prescriptive analytics much better.
- Set a strong technology flavor
Tenth, since analytics cuts across the technologies, the students need to be exposed to the basics of emerging technologies such as IoT, NLP, Cloud Computing, deep learning, blockchain, quantum computing, etc. The students also need to appreciate data mining, and there should be a good focus on teaching them to analyze the different forms of data- text and non-text, structured and unstructured.
- Leverage what industry practitioners have to offer
Eleventh, the industry sessions also need to be organized in the analytics courses. This is because the practical challenges related to problem identification, data collection, data cleaning, data processing, and drawing insights can be explained in the classroom best by the industry practitioners who are working as data scientists, business analysts, business strategists, or management consultants in the field. Also, their help needs to be taken in the design and review of the curricula. There is a relatively stronger need to appoint Professors of Practice, (as the management institutions call it), in the area of business analytics than in other areas.
- Handhold your students when they do it themselves
Twelfth, it has been observed that the students face many challenges when they try out the stuff. They may get errors due to improper syntax, or because some packages have not been installed. The instructor or a co-instructor needs to be available to them at that time, to facilitate the learning journey of the students. The teaching assistants may help the students in debugging their codes, helping them with the necessary libraries, or explaining the code at a micro level to them.
- Try using engineered datasets where possible
Thirteenth, the well-engineered datasets that have been designed to make learning more spontaneous and can generate multiple insights when minor changes are done in the data, can be very helpful. So, the instructor should make some pre-planned changes to the dataset and show to the students the impact on the solution, which can trigger a discussion on what that change in the data led to that impact.
- Evaluate the students on application and skill more than anything else
Fourteenth, the evaluation components in an analytics course should be far from theory, and should rather check the ability of the students to apply the concepts. Although there can be some theory questions as part of a broader numerical question so that they complement the quantitative problem solving
- Help the students visualize the data, and teach aesthetics of data visualization and communication
Fifteenth, visualization of data is an important element in the classroom. Before the students start working on a piece of data, if they can be asked to visualize that data through some graphical or pictorial representations, that can be very helpful for them to understand the problem-solving approach. Also, many a time, telling the storyline through the data can be very useful, and catchy if one has to deliver a business presentation to one’s client or manager. Therefore, the students need to be taught how to create dashboards, and how to draw smart visualizations, that can make the absorption by the audience easier.
- Expose your students to a variety of software tools, but do not emphasize the tool too much
Sixteenth, the instructors should try to use a variety of tools in the classroom, so that the students can become comfortable with multiple tools. For example, if some exercises can be done on R Studio, some on Python, and some on SPSS in a course, that can make the students comfortable with a wider variety of tools, which is necessary. However, at the same time, the tool need not be given as much attention as the process of framing insights from the data.
- Do not limit your delivery to structured data or numeric data only
Seventeenth, every analytics course should cover some content on unstructured data or text data as well. This is because, in today’s context, unstructured data is enormous due to the digitization of business processes. So, in any area of management, unstructured data can be worked upon to generate useful insights.
- Start the classroom session with a problem statement
Eighteen, the problem-statement-based approach has been found to work well in the analytics courses. It has been observed that if the class starts with a practical problem statement, followed by an explanation of the theory in the process of solving that problem, the absorption level of the students is higher than that in the case when the class starts with theoretical concepts.
- Ignite the scholarly passion among the students
Nineteen, it is the responsibility of the instructors to ignite the scholarly passion among the students. For this, the depth of the topics covered should be significant. For example, while teaching regression, the students can be taught all the aspects of it, including the assumptions of regression, detecting and treating the outliers, handling multi-collinearity, handling heteroscedasticity, regression using machine learning, L2 normalization, panel regression, etc. The students also need to be encouraged to try some innovations to the model taught by the instructor in the classroom.
- Develop ethics and integrity among the budding data analysts
Last but not the least, modern-day data scientists need to exercise ethics to the highest level while doing their studies and publishing the findings. Therefore, the students need to know how they can maintain objectivity in their study and not let the results or findings be influenced by their personal biases and opinions.
Thus, it can be seen that the institutions have an important role to play in the process of ensuring that business analytics as a course is taught not only in letters but also in the spirit. If this role is played adequately, the fresh graduates coming out of the business schools can be good data-driven decision-makers by the time they graduate and start working in the industry.
Dr. Gaurav Nagpal, Faculty of Decision Sciences, Business Analytics and Operations, BITS Pilani WILP