top of page
2 days ago
2 min read

People buy our course and never start. What should we change?

By Trayan Hristov · Founder, Thrivelia · 19 November 2026

Before rewriting the welcome sequence, I’d try to start the course as a new student. Did the access link work? Was the first task obvious? Could I complete it in the time the sales page led me to expect? Those questions separate a delivery problem from a motivation problem.

Define a first win a learner would recognise

For a guitar course, that might be completing a short exercise and choosing the next practice session. For a spreadsheet course, it could be submitting the first working model. Write down what counts and which event records it. A page open or an autoplayed video can be useful for troubleshooting but is weak evidence of learning.

Start the clock when access actually becomes available. Separate immediately available self-paced courses from scheduled cohorts, gift seats and corporate seat licences. A buyer who pays in March for an April intake hasn’t failed to activate in March. If a learner can work offline, acknowledge the activity your platform cannot see. Thinkific’s progress reporting explains its recorded measures.

Count the missing steps

Imagine 200 learners get access on Monday. Within seven days, 170 receive a usable invitation, 130 start lesson one and 75 complete the first exercise. The overall first-win rate is 37.5%. Thirty learners need access investigated. Fifty-five started and did not finish the exercise. They need a different intervention from the 40 people who received access and never began.

For access failures, give support an owner and a way to reissue a working link. For people who reach a long first exercise, test a smaller starting task or a worked example inside the course. For people who have access and haven’t started, offer a choice of practice time or a direct link to the first task. Stop the reminder when the event arrives.

Measure the next week too

Randomly test one onboarding change for comparable new learners. Count first wins per assigned learner, then check whether week-two practice or submissions persist. Keep refunds and support volume visible. A shorter first lesson can improve its completion rate without changing whether people learn.

The Guitar Institute case describes building around different learning paths. Use that distinction to decide what a first win means before asking a CRM team to drive one.

Ask ChatGPT or Claude about your own numbers

From these eligible learner counts, find the loss between access, first lesson and first completed exercise. Separate scheduled cohorts and gifted seats. Give me one test inside the product and one support fix.

Related reading

 
 
 

Comments


bottom of page