Why Most Corporate Training
Doesn't Stick

U.S. organizations spend over $100 billion annually on employee learning and development. A significant portion goes toward mandatory e-learning modules, full-day classroom sessions, and compliance training. These formats are organizationally convenient: they are schedulable, trackable, and auditable. Whether they produce lasting retention is a separate question from whether they are easy to administer.

The gap between common training formats and the research on memory consolidation has been documented since 1885. The formats in widest use today were largely established before that research existed, and organizational incentives have not strongly favored changing them.

The Forgetting Curve

In 1885, the German psychologist Hermann Ebbinghaus published Über das Gedächtnis ("On Memory"), a meticulous self-study in which he memorized thousands of nonsense syllables and tracked how quickly he forgot them. What emerged was one of the most replicated findings in cognitive science: the forgetting curve.

The curve is steep and it is fast. Without any reinforcement after initial learning, the average person retains roughly 58% of new information after 20 minutes, 44% after one hour, 33% after one day, and around 21% after one month. The most dramatic drop happens in the first hours after exposure. By the time most employees return to their desks after a training session, they have already lost a third of what they learned. By the end of the week, they have lost close to 90%.

This is not a failure of motivation or intelligence. Forgetting is a feature of how memory consolidation works, not a bug. The brain encodes new information into short-term storage first. Without a signal that the information is worth keeping, it is progressively deprioritized and eventually pruned. Sleep, repetition, and emotionally salient experiences are the mechanisms that move information from short-term to long-term memory. A single training session provides none of them at sufficient intensity to defeat the curve.

The research suggests that a single-session training event, regardless of content quality or facilitation, is unlikely to produce lasting behavior change on its own. The format, not the content, is the primary constraint.

Why the Traditional Format Persists Anyway

If the research on retention has been available since 1885, the question of why standard formats have not changed more substantially involves measurement and incentive structures as much as awareness of the science.

The traditional training format: a full-day classroom session, a recorded e-learning module, a compliance walkthrough followed by a multiple-choice quiz. All of these are exceptionally well suited to the things that organizations actually measure. Completion rates are trackable. Pass/fail thresholds are auditable. Seat time is billable and schedulable. An LMS (Learning Management System, the software platform that hosts and tracks corporate training) can report on all of these in real time.

What an LMS cannot easily report on is whether employees can demonstrate a new skill thirty days after completing the module. That metric requires a different kind of measurement: a follow-up assessment, a manager observation, a structured practice scenario. These are harder to standardize, harder to automate, and harder to fit into a quarterly compliance calendar. So they rarely get built.

The result is a system where what gets measured shapes what gets optimized. Completion rates and pass/fail thresholds are measurable and reportable. Retention at thirty or sixty days requires a different measurement approach: follow-up assessments, manager observations, structured practice scenarios. These are harder to standardize and automate, and are less commonly built into program design from the start.

The Spacing Effect

Ebbinghaus didn't only document the forgetting curve. He also identified one of the most powerful interventions against it: distributing learning over time rather than concentrating it in a single session. This phenomenon, known as the spacing effect, has since been replicated across hundreds of studies and dozens of domains. The finding is consistent: spaced practice produces dramatically better long-term retention than massed practice, even when the total time spent studying is identical.

The mechanism is well understood. Each time a learner retrieves and re-engages with material, the memory trace is reconsolidated. That is, the brain re-encodes the information, strengthening the neural connections associated with it and making retrieval faster and more reliable. When learning is concentrated in a single session, that reconsolidation process only happens once. When it is spread across multiple sessions with gaps in between, reconsolidation happens multiple times, and each repetition builds on a slightly stronger foundation than the one before.

The research by Cepeda and colleagues (2006) found that optimal spacing intervals depend on how far into the future the learner needs to retain the information, but even relatively simple spacing, revisiting material once or twice over the week following initial learning, produces substantially better retention than a single block of the same total duration. A 60-minute training session split into three 20-minute sessions across a week can produce two to three times better long-term retention than one 60-minute block.

The practical challenge is that spacing requires a different program structure than a single-day event. A spaced program involves multiple touchpoints scheduled across weeks, sustained engagement over a longer arc, and follow-through on material after an initial session has concluded. These are not insurmountable design constraints, but they are different from the logistics of a one-day training event.

Retrieval Practice

The spacing effect is powerful on its own. Combined with retrieval practice, the impact compounds significantly.

Retrieval practice refers to the act of actively recalling information from memory rather than simply re-reading or re-watching it. The distinction matters more than it intuitively seems. Rereading feels effective because the material looks familiar. Fluency is mistaken for knowledge. But familiarity with content and the ability to retrieve it under conditions that resemble actual use are very different cognitive states. Only retrieval builds the latter.

Roediger and Karpicke's landmark 2006 study demonstrated this precisely. Students who studied material once and then were tested on it remembered significantly more a week later than students who restudied the same material three times without any testing. The act of attempting to recall, even imperfectly and even when it felt difficult, produced stronger memory traces than repeated exposure.

This is what cognitive scientists call a "desirable difficulty," a term coined by Robert Bjork at UCLA. Desirable difficulties are conditions that slow down immediate performance but accelerate long-term retention. They feel harder. Learners often rate them as less effective, precisely because they are more effortful. But that effort is the mechanism. The struggle to retrieve information is what strengthens the memory trace.

Retrieval practice takes a range of forms: low-stakes quizzes at the start of a session on material from the previous session, flashcards that require active recall rather than recognition, and scenario-based exercises that ask learners to apply a concept rather than identify it in a list. What these have in common is that the learner must produce something from memory rather than passively receive information. Research consistently shows this production effort, even when imperfect, produces stronger long-term retention than repeated re-exposure to the same material.

The reason is partly structural. Building good retrieval practice requires knowing what learners need to be able to do, not just what they need to have seen. That design process takes longer, costs more upfront, and requires a clearer theory of what competence looks like six weeks after training. That clarity is rarely present when a training program is commissioned around a compliance deadline or a product launch.

Implications for Training Design

The research on spacing and retrieval practice points toward several design considerations that differ from those underlying most standard corporate training formats.

Distribution over time

A single session concentrates all learning into one block. A distributed program sequences the same content across multiple sessions with gaps between them, allowing reconsolidation to occur between touchpoints. The same total time, structured differently, produces measurably different retention outcomes in controlled studies. The practical question for any training program is whether the content and learning goals warrant distribution, and whether the organizational context makes that feasible.

Active recall over passive review

Lecture and e-learning place learners in a receiving role. The learner is exposed to information, recognizes it as familiar, and may perform adequately on an immediate assessment. Active formats require the learner to produce something from memory: an answer, a decision, a response to a scenario. The research on retrieval practice suggests that this production effort, even when effortful or imperfect, produces stronger long-term memory traces than re-exposure to the same material.

Measuring retention, not just completion

Completion metrics measure whether the training was delivered and received. Retention metrics measure whether learning persisted. The two are related but not equivalent. Programs designed with retention as the goal typically look different from those designed around completion: they include follow-up assessments, spaced retrieval opportunities, and application scenarios that require demonstrating competency rather than recognizing it. What gets measured in a program shapes what gets designed into it.