Summary
For self-serve learners, the course recommender offers course suggestions, based on collective learner activities across the portal.
This feature is available to all customers who use the internal catalog.
Contact your Implementation Consultant or Customer Success Manager, to discuss how the course recommender works, and how it supports learners who pick their own courses.
Prerequisites
To use the course recommender, your portal needs:
- an active catalog of courses available - see Add courses to your catalog
- a volume of users who are enrolling, completing and rating courses, to provide data for the course recommender
Optional, to allow the course recommender to use ratings:
- ratings enabled on courses - see Courses: Additional Settings to manage learners' access to a course
Access the course recommender
When enabled, the course recommender is available to users who access courses through your catalog. You don't need to enable additional settings within the portal.
- From main navigation go to Catalog.
- From secondary navigation, choose Recommended Courses.
The following screenshot shows the feature as part of the Catalog page.
Course recommender in detail: data and patterns
The course recommender draws together 3 sources of portal data, to make suggestions of courses to users:
- course ratings
- patterns of association
- course popularity
The course recommender relies on information provided by learners as they:
- enroll and complete courses
- optionally, rate the courses
The feature uses all these data sources together: you can't select one source alone for recommendations for your portal.
As it runs on your portal, and gathers more data and patterns, the recommender's suggestions for each user improves.
The feature only draws data from courses which other users enrolled in and completed. It does not review a learner's browsing history or search history as part of the recommendations.
The course recommender works within a portal, using a single portal's data. The recommender includes data from your portal's first learner enrollments and completions onward - from when you started using LearnUpon.
If your users have access to more than 1 portal, their recommendations are different, depending on what data is available in each portal.
Course ratings
Ratings refer to the "out of 5 stars" assessment users give at the end of a course. Ratings represent other learners' experiences. The recommender tracks only the star ratings: it doesn't assess text comments.
As a base for its data, the recommender needs at least 3 users to rate a course. To appear as recommended, the course needs an average of 3 or more stars from the users who complete the course.
As users complete and rate courses, they provide more data for the course recommender to apply and use in its calculations. So: the more courses completed and rated, the better the recommendations get.
The following screenshot is the rating dialog users see at the end of a course.
Note: as admin, you control if ratings are available on a course, and if they are required. LearnUpon enables Allow course rating/reviews by default: you can disable this setting at any time.
If you don't allow ratings, the course recommender can't use them as data, but can use associations and popularity to recommend courses.
You can also enable Review is mandatory? to require users to leave ratings.
See Courses: Additional Settings to manage learners' access to a course
The following screenshot shows the ratings options in Courses > your course > Additional settings.
Patterns of association
The course recommender looks for patterns of association between courses other users completed together. This type of recommendation is familiar in online shopping as, "other people who bought X also bought Y" or "frequently bought together".
Example: If 10 users each enroll and complete Course A and Course B, they create a pattern of enrolling and completing the courses together.
For the next user who enrolls in Course A, the course recommender starts offering Course B to accompany it.
This data is separate from ratings, and is also separate from course popularity.
As users enroll in and complete courses, they create more associations, which provides more data for the course recommender. As individual users complete courses, they also provide data about what courses they tend to put together. The course recommender uses this new data to update its recommendations.
Course popularity
A course which has lots of enrollments and completions on a portal is offered to newest users. The course recommender has no data or patterns about new users, but it can offer courses which have high enrollments and completions.
This type of recommendation is common in online shopping, where a new shopper has no purchase history with a store, so the store offers popular items as a starting point.
Once a user enrolls and completes some courses, the recommender can refine its recommendations and offer courses closer to a user's needs.
Course recommender and groups
The course recommender works within groups. For group members, the algorithm considers ratings, patterns of association and course popularity of other group members, ahead of these data sources from outside a group.
The recommender won't override group restrictions on your catalog.
Example: if you restrict a set of courses to certain groups, the feature won't recommend the courses to users outside those groups.
See Add courses and learning paths to the catalog, under Restrict access to groups in the catalog.
Recommendations require data
If you enabled the course recommender and learners don't see any suggested courses, the typical reasons are:
- portal is new, with few courses and enrollments: the recommender needs a minimum of 4 learners and 5 courses completed to start
- learners need to complete courses to generate data
- if you use course ratings: the recommender needs more course ratings to start
- the learner has completed all available courses so far
See: