Why should companies better profile their learners
The era of Data has led to a lot of new possibilities in the fields of human resources and learning and development. Indeed, companies seem to be increasingly interested in profiling their employees with regards to learning or, in other words, in learner modeling. We talk about Learning Analytics: the analysis of learning-related data to improve a learning system.
PROFILING IS NOT A RECENT PHENOMENON
The benefits of identifying learning profiles within a group of people were detected a while ago, and it has nothing to do with data computing. Already in education, teachers seek to identify who the students in their classroom are, what their level of proficiency is, which students are struggling and, on the contrary, which ones are progressing well in autonomy. For teachers, the main goal is to know each of their students to conclude the most adapted ways of making the classroom progress. Sometimes, this analysis also allows them to group students with similar or complementary profiles and make them work together.
In corporate education, we find the same principle but at a much higher scale. Employees are often profiled according to their position and the department which they belong to, and possibly according to their experience and seniority within the company. In that case, the objective is to identify learning programs which will presumably be more relevant for an employee: for instance, a manager who has just been recruited will follow the onboarding path about the company’s values and basic management skills. On the other hand, a technician who has worked in the company for twenty years will tend to follow a course when a new tool which he will be using is implemented. Therefore, profiling is already used by companies and it aims at personalizing learning and development. But do two employees with the same position and the same seniority in the same company truly have the same profile and the same learning needs?
HOW DATA BOOSTS LEARNER MODELING
If identifying learning profiles sometimes lacks accuracy, it is mostly due to the lack of data to analyze. In some cases, the data exists but it is simply not used. More precisely, we particularly look to the following data (in addition to the elements mentioned before): job requirements, skills repositories, employees’ learning history, results to quizzes (or assessments), assessment interviews, interactions with the learning management system, peer feedback, etc.
Once combined, this information allows to profile an employee according to:
- Learning objectives: from the employee’s position and the company’s repositories, the expected skills and level of proficiency can be deduced
- Level of proficiency: all the (qualified) data related to assessment allow to measure the employee’s knowledge and competency state
- Forgetting curve: when there is a continuous assessment of the acquired knowledge and skills, the employee’s forgetting pace can be measured
- Learning preferences: from the learning history, the formats or methodologies which have had the best impact on the employee’s upskilling can be identified
- Learning habits: the interactions with the learning management system(s) give information on the time and lengths during which learning is most effective
ADAPTIVE LEARNING FOR AN AUTOMATED INDIVIDUALIZATION
With artificial intelligence in addition to that, some behaviors can be predicted. In other words, analyzing the data of a group of people leads to identifying the one course which will be the most efficient to upskill a person from this group. To do so, the historical data is used to implement predictive approaches: if employees who have a similar profile to yours have better progressed with a course X than with a course Y, then this course X is the one you should follow to progress. Learner modeling allows to define the skill to develop, the level of difficulty adapted to you, the content format which suits you and the right time to notify you to learn. The only thing left for the learning system to do is recommending you the course which matches all these criteria (or most of them): this is Adaptive Learning. Once you have completed the course, the system will then be able to send you reminders at the right time to guarantee that knowledge and skills you have acquired are maintained over the long term.
As a conclusion, learner modeling within a company has two objectives. The first one, described above, is about individualizing learning. The more accurate the learners’ profiles are, the more adapted to their needs the learning paths will be. Adaptive Learning can automate the process, from data analysis to learning paths recommendations. For the company, a more optimal path means less time spent learning (to acquire the same proficiency on the same target skills) and, therefore, a faster upskilling process for employees. The second objective is reporting: while we often reproach artificial intelligence systems for the “black box” effect, Learning Analytics have the advantage of being accessible in dashboards for the learning players (learner, manager, L&D team, HR) to give them visibility on the entire learning process.
Our latest articles about the topics: learning, science and Domoscio's news.