The Learning Transfer Company
Table of Contents
The Shift
Learning transfer is what happens when people take what they learned in training and use it in their work. It’s the difference between an employee who can explain a new skill and one who actually applies it, adapts it to real situations, and keeps using it over time.
The Kirkpatrick Model shows where transfer fits. Level 1 asks how people reacted to the training. Level 2 asks: did they learn it? Level 3 asks the question transfer depends on: are they applying it on the job? And Level 4 asks whether that application moves business results.
Transfer is the bridge between Level 2 and Level 3, and without it, Level 4 results rarely follow. That bridge can’t be built in the training room. It’s built afterward, in the workplace, which is why training alone can’t guarantee it.
Learning transfer depends on what happens after training. Give employees opportunities to apply new skills in their work, receive feedback, and improve through practice.
Designing for Transfer
If transfer happens after training, it has to be designed for. The 6Ds framework, developed by Roy Pollock, Andrew Jefferson, and Calhoun Wick, approaches learning from that angle: how can a learning experience be designed and supported so people actually apply what they learn?
The framework emphasizes the conditions that move learning into performance, including:
Structured peer review builds several of these conditions directly into real work. Employees use the skills they’re learning, get feedback against defined criteria, and keep improving through repeated practice, long after the training itself ends.
Go Deeper
Want to go deeper on designing for transfer?
We sat down with Roy Pollock, CLO and co-founder of The 6Ds Company and coauthor of The Six Disciplines of Breakthrough Learning, to talk through transfer climate, application, and measurement. Watch: Tackling Training’s Two Greatest Challenges, Lessons from the 6Ds® →
The Problem
Learning transfer is hard to measure because it happens after training ends, in real work. A survey or quiz can capture Kirkpatrick Levels 1 and 2 in the moment, but Level 3 means observing how people work, over time.
Source: ATD Research, Effective Evaluation: Measuring Learning Programs for Success (2019)
Fewer than 1 in 5 organizations measure at Level 3. Fewer than 1 in 10 measure at Level 4.
Because business results depend on behavior change, the gap at Level 3 means Level 4 remains invisible.
The Level 3 gap is becoming harder to ignore. As organizations move toward skills-based hiring, internal mobility, and AI-assisted work, more decisions depend on knowing what people can actually do. Course completions and self-reported skills show what people have finished or claimed, not what they can demonstrate.
For more on the difference between claimed and demonstrated skills, see Does Your Workforce Data Hold Up?
The Evidence Question
If Level 3 is hard to observe, the next question is what evidence is good enough. The Learning-Transfer Evaluation Model (LTEM) focuses on exactly that: the quality of the evidence used to evaluate learning.
LTEM separates weaker evidence, like attendance, learner reactions, and simple knowledge checks, from stronger evidence that shows what people can actually do. Put simply, it’s the difference between two requests:
Evidence from real work sits at the strong end of that scale. When employees apply a skill on the job and someone reviews it against clear criteria, the result is a record of what they can do, not just what they remember.
| Framework | Primary question | Where peer review fits |
|---|---|---|
| Kirkpatrick | Did learning change behavior and results? | Creates repeated opportunities to observe skills applied in real work |
| 6Ds | How can learning be designed and supported for transfer? | Builds application, feedback, and reinforcement into the work itself |
| LTEM | What evidence do we have that learning occurred? | Produces evidence of performance, not just knowledge |
Measuring Transfer
Measuring transfer means observing how people apply what they learned in the context where it matters. Traditional approaches rely on manager observation, periodic reviews, or self-report. They can offer some insight, but they run into the same barriers as Level 3 itself: limited manager time, inconsistent raters, long gaps between reviews, and high cost.
Structured peer review builds measurement into the work itself. Employees apply what they learned to real work or a scenario-based assessment, receive feedback from peers against defined criteria, and build a record of demonstrated capability over time.
Because the work itself becomes the assessment, every review cycle produces evidence of transfer, and Peerceptiv captures that evidence at enterprise scale.
For a step-by-step look at measuring each level, see Measuring L&D Impact with the Kirkpatrick Model, or score your own team with the free Kirkpatrick Assessment.
The Research
Peerceptiv is built on more than 20 years of research at the University of Pittsburgh, involving 20,879 learners across 76 institutions, into exactly this cycle: applying a skill, getting feedback, and revising. Peerceptiv wasn’t inspired by this research. It was built from it. For the full summary, see The Science Behind Peer Review as a Corporate Learning Strategy.
Reviewing a colleague’s work means evaluating it, diagnosing problems, and explaining your reasoning. That active work builds more durable learning than passively receiving critique.
What this means: the act of reviewing builds skill in the reviewer, independent of any benefit to the person reviewed.
Yu & Schunn (2023); Wu & Schunn (2021)
Meta-analyses find a 0.63 average correlation between peer and expert ratings, with well-designed studies reaching 0.70 to 0.91. Combining several peer reviews closely tracks expert judgment.
What this means: multi-peer review with structured rubrics is a credible, scalable alternative to expert review.
Xiong et al.; Cho, Schunn & Wilson (2006)
When several colleagues independently flag the same issue, people are far more likely to act on it. Employees revising from multi-peer feedback improved more than those coached by one expert.
What this means: routing each piece of work to multiple reviewers raises the odds that feedback changes behavior.
Wu & Schunn (2020); Patchan, Charney & Schunn (2009)
Multiple meta-analyses confirm that structured peer review produces measurable performance gains that carry over to new tasks, while strengthening critical thinking and feedback skills.
What this means: peer review isn’t a workaround for limited trainer capacity. It makes the whole workforce developers of each other’s capability.
Yu & Schunn (2023); Li et al. (2020, 2021); Double et al. (2020)
Twenty additional peer-reviewed studies back these findings. See the full publications list.
The Results
The research holds up in the workplace. Here’s what structured peer review looks like at enterprise scale.
Of 58 skills tracked across multiple assignments
Of those show measurable growth (32 of 43)
Skills tracked across multiple courses
Source: Peerceptiv customer data, Fortune 100 insurer deployment
CFA Institute uses Peerceptiv for peer review on its Climate and Investing Certificate program, applying the same methodology to real candidate deliverables, with structured peer review and timestamped skill evidence. The approach that works for entry-level onboarding also works at the highest levels of professional credentialing.
Make the Case
The financial value of transfer comes from what changes after training: faster time to proficiency, fewer mistakes, better execution, and skills that actually get used. Without evidence of those changes, it’s hard to show that a learning investment paid off.
Evidence of skill application shows where learning is translating into workplace performance, and where it isn’t.
Earlier evidence of capability helps you spot readiness gaps and focus support where it’s needed.
Unapplied skills, slow ramp time, and inconsistent performance all carry a business cost that varies by role, workforce, and program.
Getting Started
Getting started doesn’t require a redesign. Structured peer review layers on top of the programs you already run.
Consider starting with onboarding. It’s where people produce real work fastest, and where time to competency is easiest to prove.
Build a simple rubric around the skills that matter most to the business.
Have peers assess a handful of real tasks or deliverables. The strongest signal comes from multiple reviews, not one.
You’ll have Level 3 evidence from a single cohort. Expand to other programs from there.
Common Questions
Learning transfer is the application of knowledge, skills, or behaviors learned in training to real work. It’s visible when someone uses, adapts, and improves a skill in the context where they’re expected to perform.
Learning is whether someone acquired knowledge or a skill. Transfer is whether they apply it, and keep applying it, in real work. A program can succeed at the first and still fail at the second.
Most often because there’s no structured opportunity to apply the skill afterward, no feedback on that application, and no reinforcement from the work environment. The 6Ds framework identifies these as preconditions for transfer.
Level 3 is the behavior level of the Kirkpatrick Model: whether learners actually change what they do on the job after training. Fewer than 1 in 5 organizations measure it, largely because manager observation is slow, subjective, and expensive at scale.
By observing whether people apply a target skill in real work, not by testing whether they can recall it. In practice, that means having peers or managers assess actual deliverables against a rubric tied to the skill, repeated over time.
Structured peer review creates the conditions transfer depends on: real application, structured feedback, revision, and reinforcement. Feedback from multiple peers drives more revision than a single expert’s alone, and every review cycle generates Level 3 evidence as a byproduct of the work.
Key Takeaways
Sources
Research conducted at the Learning Research & Development Center, University of Pittsburgh, spanning 20,879 participants across 76 institutions. Includes work by Schunn, Yu, Wu, Cho, Patchan, Xiong, Zong, Godley, Nelson & Wooley. Business and cost benchmarks sourced from ATD, McKinsey, SHRM, Gallup, DDI, The Bridge Group, and IBM. Enterprise deployment results from Peerceptiv customer data, Fortune 100 insurer deployment. Full citation list: corporate.peerceptiv.com/research/publications
Related Reading