Research shows that class schedules have a significant impact on students’ academic performance, graduation rates, and time to graduation, as well as on institutional effectiveness and financial health. Nevertheless, few colleges and universities invest time or resources into analyzing and optimizing course offerings. The American Association of Collegiate Registrars and Admissions Officers found that nearly 85% of undergraduate institutions rely on established time blocks to schedule classes – in other words, they repeat the same schedule year after year.

There are multiple advantages to fostering a more comprehensive understanding of the effects of scheduling. Operations departments can optimize the use of classrooms and classroom materials (e.g., projectors, lab equipment). Schedules can be optimized to ensure that students across programs are able to take the required courses to graduate on time. Registrars can drill down into class fill rates and understand why some courses are chronically under-filled or maxed out.

With all of these potential benefits in mind, why do so few institutions take a data-driven approach to class scheduling?

• Accessing and consolidating data is too challenging: An array of data, likely from many different sources, must be brought together to uncover the most impactful insights.
• Limited cross-domain knowledge: Even a basic analysis requires understanding a variety of data points, including student demand, program requirements for specific classes, class enrollment by time slot, faculty availability, etc.
• Scarce time and resources: If the data sets were very small, it might be possible to attempt this manually, but few institutions have the time or resources to consistently commit to this process – even at the department level.

The answer is advanced analytics, specifically through the application of machine learning (ML) algorithms. ML can help institutions quickly understand scheduling trends, patterns, and outcomes, allowing them to answer questions like:

• Are we offering the right classes at the right time to meet student needs?
• When is the best time to offer a capstone course?
• Are students performing better or worse in classes offered at a specific time of day?

While advanced analytics can help institutions take a more data-informed approach to designing their course schedules, we understand that this is potentially a politically charged conversation with a variety of competing interests. However, the potential value to an institution in increased retention, completions, and efficiency warrants a data informed strategy rather than simply repeating the same schedule year after year.  Otherwise, aren’t we just mirroring that quote about insanity, repetition, and expecting different results that is often misattributed to Albert Einstein?