Most manager training is measured by the wrong numbers. Attendance, completion rates and smile sheets tell you what happened in the room. They say nothing about what happens on Monday. That gap is where training budgets go to die. Most training is bought as an event and measured by satisfaction scores, so nobody can say what changed.
Transfer is the share of trained behavior that shows up in the work: the degree to which trainees apply the knowledge, skills and attitudes from training to the job (Baldwin & Ford, 1988). It is not satisfaction, completion or knowledge gain. It is not a trait of the learner or the course. It is a property of the system around the work.
This article defines transfer and separates it from the metrics buyers usually accept. We will cover maintenance versus generalization, near versus far transfer, typical versus maximum transfer, what transfer is not, why it belongs to the system, why open skills depend on the work environment, how to measure it without fooling yourself, and what the definition changes in a program design brief.
Transfer Is the Share of Trained Behavior That Shows Up in the Work
Training transfer is the degree to which trainees apply what they learned in a training room to the job, the definition Baldwin and Ford set in 1988. The buyer’s version is plainer. Transfer is the share of trained behavior that appears in the work, not the share of people who liked the class.
Learning happens inside the room. Transfer happens in the weeks and months after it, and the two need separate measurement (Blume, Ford, Baldwin and Huang, 2010). Most manager training is bought as an event, so satisfaction and completion stand in for transfer even though neither describes behavior on the job.
This piece is written for HR, L&D and executives at mid-market and enterprise companies who buy or run manager training. What you write into the program brief follows from how precisely you define transfer, starting with what a new manager training curriculum should include. Define it loosely and you will buy an event. Define it as behavior and you can specify what changes.
What Is Training Transfer? The Definition to Write Down
Write the definition down before you write the brief. Transfer, as Baldwin and Ford (1988) established, is the degree to which trainees effectively apply the knowledge, skills and attitudes gained in a training context to the job.
Blume and colleagues add the consequence: original learning in a training experience is rarely enough to render that training effective (Blume et al., 2010). Transfer is what turns learning into changed work performance.
That makes the definition a measurement claim. A program that never names the behavior that should appear on the job has nothing to measure, which is why supervisor training that actually changes behavior starts by naming the behavior.
Transfer is not one thing, either. It has dimensions, and a program should state which one it targets before it promises an outcome.
Nothing in the definition makes the learner the problem. That matters for everything that follows.
Maintenance and Generalization: Does the Behavior Last or Travel?
Maintenance and generalization are the two dimensions the transfer research names, and they fail differently.
Maintenance is the trained behavior persisting over time. A manager runs one-on-ones in week one and still runs them in month six. A manager who ran one and stopped has maintained nothing.
Generalization is the trained behavior showing up in situations the training never covered. The same manager holds a hard conversation with a peer, not only with a direct report. This is where manager training earns its money, because leadership and management are not the same job, and the situation across the desk keeps changing.
A check the week after the workshop tests neither dimension. It captures the day-after spike.
Ask a vendor which of the two it measures, and over what period, before you compare two programs. Persistence at ninety days and application to an unscripted conversation are different claims.
Near and Far Transfer: Same Task or an Unfamiliar One?
Generalization raises a second question: how far does the behavior have to travel? Near transfer applies the skill to a task that closely matches the practice case from the session. Far transfer applies it to a situation that looks nothing like the practice case, which is common in variable managerial situations.
Leadership conversations tend toward far transfer, since the situations are open and variable rather than fixed procedures, though the research does not map skill types to the near/far spectrum directly. A delegation framework rehearsed on one scripted scenario must survive a real handoff to a skeptical team member with a deadline attached, which is why delegation is a skill managers need trained against real variability, not a personality trait.
A program that rehearses one script is testing near transfer and calling it capability. Ask which one the design targets.
Typical and Maximum Transfer: Unprompted or Pushed?
Setting aside where the behavior shows up, the next dimension is about effort. Typical transfer is what appears unprompted, with no reminders, no measurement pressure and no extra support. Maximum transfer is what appears when people are told to try as hard as they can, and the prompt inflates the number.
Most reported results sit between the two and rarely name the condition that produced them. That gap matters when you compare online leadership courses for managers: a figure gathered in a proctored week is not a figure gathered in a normal one.
Name the condition in your request for proposal. Ask which transfer condition produced any number a vendor shows you. Without that, a one-week spike gets sold to you as behavior change, and the gap between a spike and a habit stays hidden.
What Transfer Is Not: Satisfaction, Completion and Knowledge Gain
Typical or maximum, the number still has to be about behavior. Satisfaction, completion and knowledge gain are not behavior.
Satisfaction is a reaction measure. It tracks how people felt about the room, the facilitator and the materials. Completion is an attendance measure. It tracks whether people turned up and finished. Knowledge gain is a learning measure. It tracks what people know at the end of the session.
Blume and colleagues note that leadership programs can be beloved by participants while their effect on on-the-job behavior remains the open question at the center of the transfer issue.
Take a hypothetical manager. She rates the workshop 4.8 out of 5, finishes every module and passes the quiz. She has produced satisfaction, completion and knowledge gain. None of the three tells you whether her team now gets clearer expectations.
That distinction is not pedantry. A program can score well on all three and still move nothing on the job, which is why the definition of transfer has to stay separate from them. The five behaviors that actually predict team performance are observable in the work. A reaction score is not.
None of the three touches what surrounds the manager back at work, and that is where transfer is won or lost.
Why Transfer Belongs to the System, Not the Learner
Transfer does not belong to the learner, and it does not belong to the course. Baldwin and Ford (1988) group the inputs that shape it into three sets: trainee characteristics, training design, and work environment.
The work environment covers supervisor support and the opportunity to perform the trained behavior on the job. Those are conditions the buyer controls, so transfer rises or falls on them. Transfer research commonly ranks opportunity to perform, supervisor support, content relevance, training design strategies, and instructor training as the variables that matter most.
Blaming the learner or the course skips the parts a buyer can change: reinforcement, supervisor support, and the chance to use the skill at work. Those belong in a management development plan that actually changes behavior, not in the hope that a workshop sticks.
Open Skills Need the Work Environment More Than Software Does
The work environment matters more for some skills than others. Blume and colleagues separate open skills, like leadership conversations, from closed skills, like a software procedure. Across their meta-analysis of 89 studies, motivation and a supportive work environment showed stronger relationships with transfer for open skills than for closed ones.
That finding points the design brief in one direction. A manager who is coached after the session gets more chances to perform the behavior than one handed a certificate and a follow-up email. The same logic explains why most training does not build the skills it promises: What Management Skills Actually Are, and Why Most Training Does Not Build Them.
Which makes the next question measurement: who sees the behavior, and when.
How to Measure Transfer Without Fooling Yourself
If the work environment carries more weight for open skills, measurement has to be designed for it. The trap is same-source inflation: Blume and colleagues found transfer relationships were consistently inflated when the same source rated the training and the transfer in the same context. A facilitator’s end-of-course rating of the same managers they taught is not evidence of behavior change.
Use an independent rater who sees the behavior at work. For manager training, that is usually the manager’s own boss. Measure at three points, before the program, at the end, and ninety days later. A spike at week one and a change that holds look identical in a single end-point survey.
Rate named behaviors rather than general impressions, so two raters watching the same manager would agree on what they saw. “Runs a one-on-one with a clear agenda” is rateable. “Shows better leadership” is not.
Who rates, which behaviors, and when: those three answers belong in the brief before you review a proposal.
What the Definition Changes in the Program Design Brief
Those three questions, who rates, which behaviors, and when, become four lines in the program design brief. Write them before you review a proposal.
Behavior target. Name the observable behavior, not the topic. “Clearer expectations” or “a coaching conversation” is a target. “Communication skills” or a course title is not.
Reinforcement. Training alone rarely carries transfer, so decide before the first session what continues afterward and who owns it.
Rater and timeframe. An independent rater who sees the work, measured before the program and again at least ninety days after it. Anything shorter reads a spike.
Conditions. Supervisor support and opportunity to perform are environment variables. Write them into the plan, with named owners, instead of hoping they appear.
Tandem Solutions runs the brief this way. The Manager Effectiveness Index scores fifteen manager behaviors across five dimensions, rated by each manager’s own boss before, at the end, and ninety days later. The enterprise cohort takes one capability, up to seven managers, virtual, eight to twelve weeks, measured three times, with the Ninety-Day Behavior Change Guarantee included. The guarantee covers enterprise cohorts only.
Satisfaction, completion, and knowledge gain do not belong in these four lines.
Verdict: Name Transfer Precisely and You Can Buy It Precisely
Fund the reinforcement, not just the event. For open skills like leadership, supervisor support and opportunity to perform carry more weight than the session itself. A one-week spike is not transfer. Treat satisfaction, completion and knowledge gain as inputs to a decision, not as evidence that behavior changed.
A buyer who can name transfer can specify it. A program that is specified can be measured. That is what you are buying.
Conclusion
The next step is one page. Specify an independent rater, named behaviors and a ninety-day follow-up. Fund reinforcement, not just the event. Fund the ninety days.