“What is a good course completion rate?” sounds precise, but it is incomplete. The answer changes when enrolment is free, when access never expires, when people buy for one reference module, or when the real goal is a result outside the platform.

Google Research’s work on large online courses is a useful warning: registered learners, active learners and people with different intended goals should not be treated as one uniform group. Their MOOC analysis explains the problem, and a related study examines skill and goal achievement rather than registration alone (paper summary).

Define the denominator

Write the formula next to every reported rate.

Completion rate = learners meeting your completion rule ÷ eligible learners

Now define “eligible”. Possible cohorts include purchasers, people who signed in at least once, people who started the first required lesson, or participants whose access period has ended. Each answers a different question.

Also define completion. It might mean viewing every required lesson, submitting a final assignment, passing an assessment or attending a required number of sessions. A course with optional reference modules should not force a misleading 100% content rule.

Use a small measurement ladder

Track five stages for each cohort:

  1. Enrolled: received access.
  2. Activated: signed in and started the first meaningful task.
  3. Reached the first useful result: completed the earliest action that creates value.
  4. Reached the core outcome: met the course-specific success rule.
  5. Retained the result: reported or demonstrated the outcome later, when appropriate.

This ladder tells you where to work. Low activation points to onboarding or expectation problems. A sharp drop at one lesson points to that lesson, its prerequisite or its workload. High platform completion but poor outcomes points to assessment or course-design problems.

Inspect the obstruction before adding reminders

For the biggest drop, review:

  • the exact promise made before purchase;
  • how long the task actually takes;
  • whether an example precedes the assignment;
  • mobile and caption accessibility;
  • unclear prerequisites or missing materials;
  • the delay before a learner receives help;
  • whether the next step is obvious after returning.

A reminder cannot repair an assignment that is ambiguous or disproportionately large.

Design for momentum

Make the first useful action small enough to complete in the first session. Split long modules around decisions or exercises rather than video length alone. Show required versus optional material. Save progress reliably. Give feedback close to the work it concerns.

Drip schedules can support a live cohort or prevent overload, but they can also block motivated learners. Use them because the teaching sequence needs timing, not because drip is a feature on the platform.

Evaluate changes one cohort at a time

Record a baseline, change one meaningful element and compare equivalent cohorts. Keep acquisition source, pricing and access rules in view; a cohort from a free promotion is not directly comparable to a small paid cohort.

Combine event data with five short learner interviews. Analytics shows where; a learner can explain why. Avoid claims that a tactic universally increases completion by a fixed percentage unless your own experiment and sample support it.

Idun Blue tracks learner progress and course activity in the current product. It is still in internal beta; join the waitlist if you want to be considered when external creator onboarding begins.