Every organization that funds training makes the same implicit bet: if people learn the material, performance will follow. The research record shows how often that bet is lost, and more importantly, why. The gap is not a mystery of motivation or a failure of content quality. It is a missing stage in the learning journey itself.
How big is the training transfer problem?
A widely cited measurement of transfer decay comes from Saks and Belcourt (2006), who surveyed training professionals across 150 organizations and published the results in Human Resource Management (study abstract). Their respondents estimated that the share of employees applying training material on the job was:
- 62% immediately after training
- 44% six months later
- 34% one year later
Read as a curve, the numbers describe steady erosion. By these estimates, roughly a third of employees never carry the training to the job at all, and among those who do, nearly half have stopped within a year. A program can post strong completion rates and assessment scores while two-thirds of its graduates are no longer applying it twelve months later.
Wilson Learning's research on learning transfer arrives at a compatible estimate from the performance side: most analyses suggest only about 15 to 20% of learning investments result in work performance change (Learning Transfer Model, Leimbach). For a budget holder, that is the uncomfortable arithmetic behind every course library and workshop calendar. Most of the spend produces knowledge that is never converted into behavior.
Saks and Belcourt also surfaced a design irony. Organizations concentrated their transfer-supporting activities during training, yet activities in the work environment before and after training were more strongly related to transfer. The effort is being spent where it is easiest to spend, not where it works.
Why doesn't knowledge become performance?
The transfer gap persists because knowledge and judgment are different abilities, built by different kinds of learning. A course can make a learner fluent in a framework: they can define it, pass a quiz on it, even explain it to a colleague. None of that requires the learner to recognize a live situation where the framework applies, choose among plausible options under ambiguity, or recover when a first attempt lands badly. Those are the abilities performance actually demands, and content consumption does not build them.
The broader evidence that doing beats watching is unusually strong. Freeman et al. (2014), in a meta-analysis of 225 studies of undergraduate STEM courses published in PNAS, found that active learning raised examination performance by 0.47 standard deviations, a medium effect size by convention (0.8 counts as large), while students in traditional lecture courses were 1.5 times more likely to fail (Freeman et al., 2014). Students who worked through problems themselves outperformed students who watched an expert present, across every discipline studied.
Worse, learners cannot feel the difference. Deslauriers et al. (2019), in a randomized experiment at Harvard using identical course materials, found that students in passive lectures felt they learned more, while students in active classrooms actually learned more (Deslauriers et al., 2019). The fluency of a polished presentation is misread as learning; the cognitive effort of applying material is misread as struggling. This perception gap helps explain why programs built on content delivery can earn strong satisfaction scores while producing weak transfer. Everyone in the room, including the learner, believes the training worked.
The practical consequence shows up on the job as a familiar pattern: they know it, but they do not do it. When the real moment arrives, the learner has no comfort level with acting on the knowledge, because the training never asked them to act. Old habits are rehearsed daily; the new approach was only ever described. Under pressure, the rehearsed behavior wins.
What does a complete learning journey look like?
The transfer problem becomes tractable when a learning program is designed as a journey with distinct stages, each building a different ability.
Stage 1
Knowledge acquisition
Builds: Facts, concepts, frameworks, vocabulary
Methods: E-learning courses, instructor-led training, videos, reading, documentation
Stage 2
Learning to apply knowledge
Builds: Judgment: what to do, when, and why in realistic situations
Methods: Realistic scenarios with decision points, consequences, and expert guidance
Where transfer is won or lost
Stage 3
Practicing communication delivery
Builds: Polish: how to say what has been decided
Methods: Role-play and conversational rehearsal tools
Stage 1 is where most training effort and budget already goes, and it is genuinely necessary. Nothing can be applied that was never learned. The failure mode is treating Stage 1 as the whole journey.
Stage 2 is where transfer is won or lost. This is the stage where learners learn to apply what they acquired: they face situations that resemble real work, make consequential decisions, see outcomes play out, and receive guidance at the moment of choice. Every learning program needs this stage, because application in realistic situations is what makes knowledge relevant, actionable, memorable, and useful. Without it, the Stage 1 investment is stranded.
Stage 3 addresses delivery, refining how a message is spoken once the learner already knows what the right move is. It is valuable for communication-heavy roles, but it presumes Stage 2 is complete. Rehearsing how to say something before learning what to do wastes the investment: polished delivery of a decision the learner never learned to make.
Most organizations do not skip Stage 2 knowingly. They cover it with proxies: a vignette question at the end of a module, a case discussion after a lecture, or a traditional branching simulation that routes learners down prewritten paths. These gesture at Stage 2 without meeting its requirements. In practice they tend to be shallow, difficult to scale, and hard to measure, and traditional branching simulations in particular grow brittle and costly as paths multiply. The stage appears in the program plan but not in what learners actually experience.
What does designing in the application step look like?
Wilson Learning's synthesis of 32 studies, covering 66 distinct transfer activities, offers one of the clearest maps of what moves transfer (Leimbach, Learning Transfer Model). Individually, most activities improved performance about 20% over training alone. Combined, the full set improved learning effectiveness by as much as 186%. The activities fall into three categories: learner readiness before the program, transfer-oriented instructional design during it, and organizational alignment after it, and all three categories carry real weight, so in-program design is one lever among several rather than the whole answer.
Instructional design is the lever designers control most directly, and the research finding there centers on fidelity: the closer the in-course experience replicates real working conditions, the greater the transfer. Building on that finding, this publication's view is that an application step earns its place in a program when it provides four properties:
- Realistic situations. Fidelity is the researched foundation: abstract exercises train abstraction.
- Consequential decisions. Learners choose among plausible options and see outcomes unfold, not select the obviously correct answer from a list of distractors.
- Expert guidance at the moment of choice. Feedback tied to a specific decision, delivered when the decision is made, converts a wrong turn into learning. End-of-module scores arrive too late to shape judgment.
- Observable evidence. The step should generate decision-level data, what learners chose before guidance and how their choices changed after it, so the program can demonstrate changed decisions rather than infer change from quiz scores.
The field is already moving this direction. As of ATD's 2021 study, 98% of surveyed organizations used scenario-based learning and 75% used technology-based simulations, up sharply over five years; better knowledge retention and application was the top reported driver for simulations and among the top drivers for scenario-based learning (ATD, 2021). The same research found high-performing organizations use these methods in a significantly greater share of their programs than the median organization. Adoption, however, is not the same as design quality; a scenario that lacks realistic decisions and moment-of-choice guidance inherits all the transfer problems of the content it was meant to fix.
This methodology is what AliveSim was built to implement. Its Guided Scenarios put learners inside realistic situations where they learn to apply what they have studied, with corrective mentoring at each decision point, and the platform captures decision-level analytics as learners work. Learners show a 2.5x improvement in decision-making performance within the scenarios, measured against their own baseline first choices (Syandus data): in-experience evidence of growing competence, the earliest signal a transfer-minded program can collect. The methodology stands on its own, though: any program that engineers realistic decisions, immediate expert guidance, and decision-level measurement into its application step is doing Stage 2 properly, whatever tools it uses.
How can you tell if your program has the application gap?
The gap rarely announces itself, because Stage 1 metrics look healthy right up until performance is measured. A short diagnostic:
- Where does the learner's last decision happen? If the final graded activity is a recall quiz, the program ends at Stage 1. If it is a realistic decision with consequences and feedback, Stage 2 exists.
- Could a learner complete the program without ever choosing badly? Programs that never let learners take a wrong path also never correct one. Judgment is built at the moment a default choice gets challenged.
- What evidence exists beyond scores and satisfaction? If the program cannot show what learners chose and how those choices changed, transfer is being assumed, not demonstrated.
- What happens in the ninety days after the program? Saks and Belcourt's decay curve is steepest early. If nothing in the design reinforces application after the event, the six-month number is being left to chance.
- Do managers describe graduates as knowing it but not doing it? That phrase, in any of its local variants, is the application gap reported in plain language.
Two or more weak answers indicate the same structural issue: a learning journey that invests in knowledge acquisition and stops. The remedy is not more content, better slides, or a longer course. It is completing the journey, designing in the stage where learners learn to apply what the rest of the program taught them. That is the step where training investments stop decaying and start compounding.
References
- ATD Research. (2021). Use of simulations and scenario-based learning is rising. Association for Talent Development.
- Deslauriers, L., McCarty, L. S., Miller, K., Callaghan, K., & Kestin, G. (2019). Measuring actual learning versus feeling of learning in response to being actively engaged in the classroom. Proceedings of the National Academy of Sciences, 116(39), 19251–19257.
- Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., & Wenderoth, M. P. (2014). Active learning increases student performance in science, engineering, and mathematics. Proceedings of the National Academy of Sciences, 111(23), 8410–8415.
- Leimbach, M. Learning transfer model: A research-driven approach to enhancing learning effectiveness. Wilson Learning Worldwide.
- Saks, A. M., & Belcourt, M. (2006). An investigation of training activities and transfer of training in organizations. Human Resource Management, 45(4), 629–648.
Related questions
What is training transfer?
Training transfer is the degree to which knowledge and skills acquired in a learning program are applied on the job and sustained over time. It is the difference between what learners can recall in a course and what they actually do in real work situations. Researchers distinguish transfer from learning itself: a learner can score well on an assessment and still fail to transfer, because recall and application are different abilities. Transfer is the outcome most training investments are ultimately funded to produce, which is why transfer decay, the steady decline in application after a program ends, is the central problem in learning design.
What percentage of training transfers to the job?
Saks and Belcourt (2006), surveying training professionals across 150 organizations, found that 62% of employees apply training material immediately after training, 44% six months later, and 34% one year later. By that estimate, two-thirds of trained employees are no longer applying what they learned within a year. Wilson Learning's synthesis of the transfer literature reaches a similar conclusion from a different angle, relaying the common estimate that only about 15 to 20% of learning investments result in measurable work performance change. Both figures describe typical programs built around knowledge acquisition alone, without a designed application step.
How do you improve training transfer?
Transfer improves most when it is treated as a design problem spanning the whole program, not an afterthought. Wilson Learning's review of 32 studies grouped effective transfer activities into three categories: learner readiness before the program, transfer-oriented instructional design during it, and organizational alignment after it, with a combined potential improvement of up to 186% over training alone. Instructional design is the lever designers control most directly, and the research points to fidelity: the more closely in-course experiences replicate real working conditions, the greater the transfer. Realistic scenarios where learners learn to apply knowledge and make decisions provide that fidelity, and manager coaching and peer support then sustain the new behavior on the job.
What is the difference between knowledge acquisition and learning to apply knowledge?
Knowledge acquisition is absorbing content: courses, lectures, videos, reading, and the assessments that verify recall. Learning to apply knowledge is a separate stage in which learners develop the judgment to use that content in realistic situations: recognizing which principle applies, choosing among plausible options, and adjusting when conditions shift. The two stages build different abilities, and one does not produce the other automatically. A learner can pass every quiz and still freeze at a real decision, because recall was trained and judgment was not. Complete learning programs design both stages deliberately.
Published February 18, 2025 · Updated July 10, 2026 · 9 min read