Abstract:

The implementation of information systems (IS) in Higher Education Institutions (HEIs) is a highly complex and critical process. Its deployment, fundamental for modernizing management, is often fraught with errors that directly impact operational efficiency, generating bottlenecks, user dissatisfaction, and an overload on support units. This study aimed to identify, classify, and analyze the errors that occurred during the implementation phase of the Academic Information System (AIS) at the University of Quindío (Colombia). The goal was to transform incident data into strategic information to propose specific and prioritized improvement strategies.

To achieve this, a quantitative and exploratory research study was conducted. The primary data source consisted of incident reports submitted by end users (students, faculty, and administrative staff) to the institutional help desk. The statistical technique of cluster analysis was applied, using the K-means algorithm, to group the errors according to their frequency, criticality and relationship with the system processes.

The analysis identified five error clusters: ranging from frequent, low-criticality failures, such as grade entry, to complex errors related to academic program management. Most problems were concentrated on the most frequently used functionalities, highlighting the relationship between user training and the occurrence of errors.

The study concludes that cluster analysis is a useful tool for quality management, transforming support data into strategic information. It is recommended that training and redesign efforts be focused on the modules associated with the high-criticality clusters, to reduce operational risk and strengthen institutional efficiency.

A subsequent causal analysis of these clusters allowed the root causes of the problems to be classified into three main groups: operational (attributed to a lack of user training and understanding); technological (configuration failures, software bugs, and update problems); and managerial (originating from deficiencies in coordination between departments and institutional workflows).

It is further concluded that cluster analysis is a valuable tool for software quality management, as it allows for going beyond simple descriptive analysis. It reveals that the causes of failures are not only technical, but also organizational and human. It is recommended that training and redesign efforts be focused on modules associated with high-criticality and high-frequency clusters in order to reduce operational risk. Finally, the importance of incorporating knowledge management as a support tool is highlighted, transforming detected errors into opportunities for learning and continuous optimization for the institution.

Keywords:

Cluster analysis, multivariate data analysis, information classification, knowledge management, error detection, information systems, academic management, universities.