Learning & Development (L&D) programs typically have two high-level goals:
1. Successfully imparting knowledge and skills to traineesWhile these goals are simple in theory, they are often complex in practice, and it is not always clear which programs or approaches are most effective at achieving them. This ambiguity results in a large amount of scrap learning – training that is delivered, but does not affect job performance, whether due to a failure of learning or development.
According to CEB, scrap learning is a serious problem, making up 45% of learning investments in the average organization. In an effort to minimize this wastage, L&D departments are increasingly turning to sophisticated learning analytics solutions to help them better understand their training data.
These efforts have certainly helped organizations make great strides in their understanding of existing training initiatives. However, they are also limited in the sense that they are mostly backward-looking, and tend to emphasize the past instead of the present and future. That’s a missed opportunity, because obstacles to L&D are best dealt with by anticipating them beforehand, or dealing with them when they arise.
Enter Predictive Learning Analytics (PLA) – a set of methods and technologies used to model future learner outcomes. By identifying patterns and trends in historical data, organizations can make predictions about how learners are likely to behave in future training programs and on the job. Those predictions are then leveraged in various ways to reduce scrap learning.
Let’s take a look at four uses for PLA in your organization:
In any given course, trainees will vary in how quickly and easily they are able to progress through the material. And often, trainees who lag behind experience a negative “snowball effect”, where slow progress leads to discouragement and further difficulty in learning advanced concepts.
The best way of dealing with this problem is to provide appropriate coaching and support when trainees start to fall behind. Unfortunately, that is often easier said than done, as it can be hard to know who is struggling with the course material. In some cases, trainees are unaware of their own “knowledge gaps”, and in others, they might simply be uncomfortable asking for assistance.
With PLA, instructors can keep a close eye on trainee progress by comparing specific metrics against what they typically mean for course performance. For instance, consistently low quiz scores combined with a lack of forum participation might indicate a trainee who’s not actively engaged. This will allow instructors to identify opportunities for intervention with those specific trainees.
Another way to increase the success rate of training is to provide the trainees themselves with direct feedback about their performance. Most trainees are keen to maximize their learning, but have “blind spots” when it comes to evaluating their own progress. In many cases, they might simply not realize where they are underperforming or lagging behind.
Organizations can use PLA to identify actionable metrics for trainees, and display these metrics in a way that’s appealing and easy to understand. This can help trainees chart the best course for their own learning.
An example in the education field is Austin Peay State University’s Degree Compass system, which is inspired by recommendation systems from Netflix, Amazon and Pandora. Leveraging general and personal course data, the system identifies future courses that would fit well into the student’s program of study, and that the student is likely to do well in. It then presents the most promising recommendations through an intuitive dashboard, accessible through web and mobile.
Timely individual interventions are not the only way to reduce scrap learning. Sometimes, there can be more fundamental problems with the course itself. By providing a clearer picture of how specific course elements influence learning results, PLA can be a powerful tool for improving course design.
An effective implementation of this is found in the Curtin Challenge platform. Developed by Curtin University, this application was designed to facilitate self-organized activity and open-ended inquiry among students. Given this hands-off approach to learning, course designers required a strong analytics program to evaluate the effectiveness of various course elements.
Curtin Challenge’s administration dashboards include information on drop-off rates, student ratings of various modules and weekly participation metrics. Consistent activity drop-off could indicate a lack of engaging material, or an issue with the level of difficulty. In this way, designers were able to monitor learning performance, and refine future iterations of the course.
In a knowledge-based economy, the importance of personnel development extends beyond the L&D department. Managers and administrators often need to understand the process and expected results of training to make better decisions.
An example of a tool that facilitates this is Lambda Solution’s own Zoola Analytics platform. Zoola makes it easy to create intuitive, visually appealing reports for decision-makers across your organization. You can make a custom report for each set of decision-makers, and automate delivery with a click of a button.
Armed with this information, your managers will be better able to plan for onboarding time, allocate training investments, and nominate the most suitable trainees for each course.
Depending only on backwards-looking analytics is like driving while looking at the rearview mirror – comforting, but dangerous. To increase the effectiveness of your training programs, it is essential to consider the future, and the best way to do that is by implementing predictive learning analytics.