We celebrate high completion rates in our training programs. They feel like success, clean percentages that stakeholders understand and LMS systems track easily. However, do completion rates reveal anything meaningful about learning?
People click through entire modules without absorbing the content, just checking compliance boxes. Think about the last time you completed that annual mandatory training. Then ask your colleagues about the last time they finished theirs. In other instances, learners dive deep into the content, apply it immediately, but never officially “complete” the course because they got what they needed and moved on to solve real problems. Unless, of course, completion is required.
The gap between completion and learning remains. Are we measuring the right things?
Why We Love Completion Rates
Completion rates are seductive because they’re measurable. Your LMS generates clean reports with metrics that allow leadership to track progress, and everyone feels confident about their training investments. They solve the “how do we know people did the training” question. But completion rates measure compliance, not capability. They tell us someone reached the end of content, not whether they can do anything different because of it.
The actual evidence of learning is evident in job performance, as evidenced by someone making better decisions, solving problems they couldn’t handle before, or improving their work as a result of training. However, measuring performance change is complex, time-consuming, and often depends on factors beyond the training itself. So, we default to completion rates because they’re available, maybe not because they’re meaningful.
Signs That Learning Happened
If completion rates don’t tell us about learning, what does? Learning science points to several indicators that someone has genuinely absorbed and internalized new knowledge or skills:
Voluntary re-engagement with content. When people return to specific resources or bookmark particular sections, they’ve identified something valuable enough to revisit. This self-directed behavior suggests the content addressed real needs.
Evolution in questioning. Early in learning, people ask procedural questions: “How do I do this?” As understanding develops, questions become more sophisticated: “What if I did this differently?” or “How does this connect to that other concept?” The complexity of questions often reflects the depth of learning.
Cross-contextual application. Real learning transfers across situations. Someone who applies a conflict resolution technique from leadership training to a family situation has moved beyond surface-level memorization to genuine understanding.
Spontaneous knowledge-seeking. When learners voluntarily explore related resources or seek out additional learning opportunities, it indicates that the initial learning has created momentum for continued growth.
Teaching or explaining to others. Perhaps the strongest indicator of learning is when someone can introduce the concept to others or reference it naturally in conversations. This demonstrates both understanding and integration.
The Middle Ground: Better Indicators
These engagement patterns aren’t proof of learning either – they’re more like conditions that make learning transfer more likely. Someone who never engages deeply with content is unlikely to apply it. But deep engagement doesn’t guarantee application – it just creates better odds.
Return visits, time spent, and bookmarking behaviors are still analytics, not evidence. They tell us more about content quality and learner intent than completion rates do, but they’re not the same as watching someone solve problems they couldn’t solve before.
Think of engagement analytics as leading indicators rather than evidence of learning. They might predict whether learning transfer is possible, but they can’t tell us whether it actually happened. There’s more to explore about what these signals predict and what they miss entirely. More on that next week.
Questions I Ponder
What would happen if we designed learning experiences assuming most people won’t “complete” them traditionally? What if we optimized for immediate usefulness rather than comprehensive coverage?
How might we track performance improvements without creating burdensome measurement systems? And what would it mean to define training success based on application rather than completion?
I don’t have neat answers yet. However, I believe that our current metrics may be addressing the wrong problem.
Small Experiments Worth Trying
Instead of abandoning completion tracking entirely, what if we supplemented it with human-sized indicators?
Track what people actually use. Most LMS platforms already capture return visits and content engagement – we might just need to look at this data differently.
Ask simple application questions. Not immediately after training, but 30 days later: “Did you use this?” The answers might be revealing.
Notice skill demonstrations in real work. Look for evidence that people are doing things differently, rather than relying solely on assessment scores.
Pay attention to what people share. When someone references training content in a meeting or forwards a resource to a colleague, that’s application happening naturally.
The goal isn’t perfect measurement – it’s a better understanding of whether our learning experiences actually help people perform their jobs more effectively.
What patterns have you noticed between completion and actual learning? I invite you to share the gaps you’ve observed and what signals you trust instead. Converse with me at the Coffee Corner or on LinkedIn.
