The L&D industry has been optimizing for completion rates for approximately two decades. The typical quarterly L&D report features a completion rate prominently — often as the headline metric, sometimes as the only metric. "87% of employees completed the Q3 required training." That number gets presented to HR leadership, logged in the HRIS, and used as evidence that the learning program is working.
Here is the problem: completion rates measure whether someone clicked through a module. They measure nothing else. They do not measure whether the learner understood the content, retained it a month later, or changed any behavior because of it. And yet they have become the primary metric by which L&D programs are evaluated and L&D teams are judged.
This post argues that optimizing for completion rates is not just unhelpful — it is actively misleading, and explains what to measure instead.
Why completion rates became the default metric
Completion tracking emerged from compliance training requirements. Regulators and auditors wanted documentation that employees had been exposed to required content: harassment prevention, data security, safety procedures. A completion stamp in the LMS provides that documentation. For that specific use case — compliance audit trail — completion is a valid and sufficient metric.
The problem is that this compliance-driven metric migrated into skill development contexts where it is not valid. When an L&D team builds a 12-module product knowledge curriculum and then measures its effectiveness by tracking how many employees completed all 12 modules, they are applying a compliance metric to a performance outcome. Those are not the same thing, and treating them as equivalent produces distorted decision-making.
What completion rates actually measure — and optimize for
If your primary performance metric is completion rates, you will optimize for completion rates. The behaviors this incentivizes in learners and L&D designers are worth being honest about.
In learners: Completion optimization produces click-through behavior — advancing through slides without engaging, leaving videos playing in the background, using multiple-choice answers to locate the correct answer through elimination. Learners know that completing the module is what triggers the checkmark, not demonstrating understanding. The rational response to a completion-based system is to complete, not to learn.
In L&D designers: Completion optimization produces shorter modules with lower cognitive demand, because shorter content has higher completion rates. It produces assessment questions that are easy to pass on the first attempt, because high pass rates make the program look good. These optimizations are directly opposed to the learning science principles — spaced retrieval, interleaved practice, desirable difficulty — that produce durable skill acquisition.
What to measure instead
The alternative is not to measure nothing. L&D programs need metrics to demonstrate value and guide improvement. The question is which metrics are proxies for skill acquisition and performance change, not proxies for mouse-click activity.
Pre/post skill assessment delta. If you measure skill proficiency before a learning intervention and again 4–6 weeks after, the delta is a genuine signal of whether the program produced skill change. This requires assessment infrastructure and willingness to measure the pre-state honestly, but it directly answers the question "did this program close the gap it was designed to close?"
Time-to-velocity by cohort. For onboarding programs specifically, comparing time-to-velocity across hire cohorts is a meaningful outcome metric. A cohort that went through program version A and reached team velocity in 14 weeks versus a cohort that went through version B and reached velocity in 10 weeks — that is signal about program effectiveness that completion rates can never provide.
Skill retention at 30 and 90 days. Learning science consistently shows that retention at 30 days is more predictive of behavioral change than immediate post-assessment scores. Measuring skill retention at intervals after program completion — not just at completion — tells you whether the learning design produced durable acquisition or short-term performance that fades. The gap between "right after training" scores and "30 days later" scores is often large enough to be actionable.
Manager-assessed behavioral change. For skill domains where behavioral observation is possible (customer communication, discovery questioning, presentation delivery), calibrated manager assessments of observable behavior change at 60–90 days post-program are high-signal outcome measures. They require calibration to be reliable, but calibrated manager ratings are a standard HR analytics methodology with well-understood validity properties.
The counterargument: completion still has a role
We want to be clear: we are not saying completion rates should be abandoned entirely. For compliance training, where the legal requirement is documented exposure, completion tracking is the right metric. For tracking whether learners are engaging with assigned paths at all — a prerequisite to any learning outcome — completion rates serve as a basic health check. If a path has 20% completion, something is wrong with the design or the assignment process, and that is useful to know.
What we are arguing against is using completion as an effectiveness metric for skill development programs. Completion is a necessary but not sufficient condition for skill acquisition. You cannot learn a skill without completing the relevant content, but completing the relevant content does not guarantee you have learned the skill.
The organizational challenge of moving to better metrics
Switching from completion metrics to skill outcome metrics is harder than it sounds, because completion rates are easy to produce and skill change metrics require more infrastructure and more patience. The reporting cycle for completion rates is days. The reporting cycle for time-to-velocity or skill retention data is months.
There is also a political challenge: if L&D teams have been reporting completion rates to HR leadership and business leaders, switching to harder-to-hit outcome metrics can feel like admitting that the previous reporting was misleading. The right framing is to present both in transition — keep completion as a health check, add outcome metrics as the effectiveness story — and gradually shift the headline metric to the one that actually answers the question leadership cares about: is the training improving performance?
A team that can show "our onboarding cohort this quarter reached velocity 3 weeks faster than last quarter's cohort" is in a fundamentally different position with leadership than a team that can show "87% completion." The first number is defensible. The second is noise dressed up as signal.