The problem
Every learner received the same linear course regardless of prior knowledge or pace, so advanced learners disengaged and struggling learners dropped out.
What we built
- An adaptive engine that re-sequences modules based on assessment performance
- Learning analytics dashboards for instructors to spot at-risk cohorts
- Automated remediation content triggered by repeated misses on a concept
How we got there
- 1
Analysed historical completion data to find the drop-off points
- 2
Piloted adaptive sequencing on two courses against a control group
- 3
Scaled to the wider catalogue once completion gains were confirmed
Results
- •65% improvement in course completion rates
- •40% better learning outcomes
Technologies
Machine Learning
Adaptive Algorithms
Learning Analytics
What’s next
Extending the engine to recommend the next course, not just the next module.