Two Challenges of Applying RPA and Data Analytics in the Education Industry in China
Before 2020, my knowledge of robotic process automation (RPA) and data analytics was rather limited to its applications in the technology industry, such as user review analysis in e-commerce. Since COVID-19, I have been increasingly aware of the growing applications of RPA and data analytics in traditional industries such as accounting and automobiles, but I had hardly heard of such applications in the education industry where I worked. In order to explore the possibility of empowering business growth, I got my supervisor’s agreement on integrating resources and collecting data within the company, and I completed two projects, namely “Analysis of WeChat Records of 118 Students and Families” and “Features of Students Who Defer Application and Refund.” While we got exciting results from the projects, the process of turning insights into solutions was not very smooth due to the change of managers with different views and ideas on applying data. Despite the results, the takeaway of applying RPA and data analysis in the education industry (and many other traditional industries) is valuable. In this article, I will summarize two difficulties I observed in the process.
The first challenge is the difficulty of collecting accurate data. For technology companies, this is not a problem: people are using cell phones, computers, smart bracelets, and other intelligent devices all the time. They take the devices with them, and each time they use them, their behavior turns into structured data with time, location, and other details stored in the database of technology companies. In Europe, there are stricter rules about how to get user consent before collecting data. In China, on the other hand, many permission options are open by default, and there are many UI/UX design details that “trick” users into giving up their privacy.
However, conditions in traditional industries are far from the same. It would be easier if companies collected data from their own APPs, but many business data sets are not generated through the APP, which all need to be manually entered from various sources. For example, in education services, students’ details are required in order to be evaluated for suitable academic programs. Data sources include undergraduate academic transcripts provided by the students themselves, standard test scores provided by Cambridge or ETS, and graduate school admission results from different university systems that need to be checked and entered by the practitioners in education services themselves.
Such data collection turns into even more chaos in poor company management, in which departments collaborate with low efficiency and different departments have different criteria for examining data indicators and performing analysis. Since traditional companies often don’t have top programmers and system engineers like tech companies, their information systems and their codes often operate poorly in time and space complexity. So, instead of spending their time building their businesses and giving high-quality, personalized services, practitioners have to spend a lot of time entering the same data into different sheets or information systems, all of which are hard to use.
In order to make possible improvements, I tried to offer advice to managers in charge of information systems, such as adding new functions to collect data in multiple dimensions to facilitate analysis of accurate data collected in one single system instead of multiple systems. However, my attempt revealed a crucial question: interdepartmental cooperation was much more difficult than expected. Managers within the same department couldn’t agree on one decision, and obviously, managers who considered it to increase their own workload called a stop to the change of digitization. Besides, the different task schedules of each department (such as business, IT, etc.) and their managers’ opinions on such changes are also a vital factor in deciding whether such a project would be implemented smoothly. Moreover, in China, for a company that seems to be stable, the cost of making these system-level changes could likely be greater than the cost of inputting data manually.
In short, conflicts of interest within a department and between departments pose a critical challenge to applying data technology like RPA and data analytics, especially the first step, which is collecting accurate data from multiple resources. Although such conflicts would be fewer in smaller companies, they have their own challenges. If they aim at a smaller but specialized market share, they might consider whether it is worth it to spend a large amount of money on collecting a large amount of data and deriving insights.
In addition to the difficulty of collecting data, which can be seen rather obviously in business operations, the bigger challenge is the corporate culture that is not visible on the surface: is the business of the industry highly dependent on big data? Do managers believe that data is critical to the future of business, and do they recognize the importance of data analysis? For technology companies, this is not a problem either. For instance, social platforms rely on advertising to make profits, and effective advertising requires examining and analyzing user profiles, so insightful data analysis is naturally needed. Nevertheless, many traditional industries are different. Although some companies claim to be “XX (any traditional industry) technology companies,” technology is rarely present in corporate culture. Many managers believe that business goals can be accomplished merely through many years of experience or even micromanagement, and it would be fine for “AI/big data” to serve as nothing more than an eye-catching gimmick. Few of them care whether these big data are accurate or reflect real-world conditions, whether decisions based on these data are ethical, and whether such a business can be sustainable in the long run, which is a sad reality.
Overall, when applying RPA and data analytics in traditional industries, it can be a stressful experience for practitioners if they encounter these two major difficulties. Their enthusiasm and belief in data analytics might also be lost in the complex conflict of interests. However, we should remain optimistic about the unstoppable future trend: The financial industry always takes the initiative to embrace new changes, and Big Four has been applying RPA and developing AI accounting since 2017. Their continuous efforts in digitalization have brought them considerable profits. The education industry, with lower profit margins and higher requirements for personalized services, still has a long way to go.
