Learning transfer: help people apply training in everyday work

Learning transfer is the application of learned knowledge or skills in work, including their use beyond the exact training exercise and their retention over time. Ford, Baldwin and Prasad’s research review discusses that generalisation and retention, alongside unresolved questions about how transfer should be investigated and measured.

A course can be well received without changing the task it was meant to support. Plan transfer before the training: define the work, provide practice and feedback, arrange an opportunity to use the learning and check what happens later. The learner, manager and people supporting the work all have a part.

This is an original practical planning process informed by research and institutional guidance. It offers a way to diagnose application problems; it does not supply a universal transfer rate or a validated rule for dividing learning time.

Define the task before choosing the course

Start with the difference that should become observable. “Improve dispatch skills” is too broad. “Prepare a mixed delivery crate against the current order, recognise a mismatch and use the agreed exception route” identifies something that a learner and coach can practise and inspect.

Describe the setting, available reference, required standard and authority. Someone who can complete a rehearsal with a coach standing beside them has demonstrated something useful, but different from independent performance during a busy dispatch period. Make both expectations explicit.

Use competency mapping when several roles need different capabilities. The course should address the required gap, rather than begin with a catalogue of available modules. Consider whether the task instead needs a clearer reference, a better tool or a decision about responsibility.

The distinction fits ADKAR’s knowledge and ability elements: knowing what to do and being able to do it are different questions. Neither label establishes the cause of a problem. Investigate the work before deciding that a colleague requires more instruction.

Design a practice opportunity before the course

Consider a hypothetical bakery changing how mixed delivery orders are checked at the packing bench. Workers attend a short session explaining a new sequence: compare the order, check the bread types, confirm quantities and route any mismatch before dispatch. The example concerns process learning, not a real food-safety procedure.

The manager initially plans a presentation at the end of a shift. A worker points out that their next three shifts are in another area, so they will not use the new routine immediately. Another says the order reference is normally kept away from the packing bench. Both are transfer conditions that need attention before training begins.

Choose a realistic rehearsal with fabricated order data. Demonstrate the sequence, then let the learner attempt a different order while explaining the checks. Include a plausible mismatch and an opportunity to ask for help. A repeated copy of the trainer’s example may conceal whether the person understands the task.

Arrange the first supported use in ordinary work. Name the coach, the expected time and the person covering other duties. Provide an accessible reference at the bench and decide which cases still require oversight. “Apply what you learned” is not a plan when the learner has no relevant assignment or safe support route.

Check the environment before attributing a skill gap

The COM-B framework considers capability, opportunity and motivation together. Applied to transfer, a worker may understand the new checking sequence yet lack the time, reference or authority needed to perform it. A capability intervention alone would leave those conditions unresolved.

For the fictional bakery, examine whether the bench has space for the order and crate, whether the current order can be retrieved and whether the dispatch target allows the check. Ask what happens when a colleague discovers a mismatch. If the supervisor tells staff to follow the new routine but rewards leaving immediately, the working instructions conflict.

Keep these issues with people able to alter them. The manager owns capacity and priorities; the process owner owns the checking and exception arrangement; the learner owns their participation and questions. The coach can report a barrier but cannot promise a staffing change they have no authority to make.

Use change readiness to bring those prerequisites into the implementation decision. If a new application changes where the order appears, coordinate the transfer plan with the software rollout. Otherwise, practice may teach a task that the live system does not yet support.

Use feedback that helps the next attempt

Give feedback on an observable action and its consequence. “You compared the quantity but did not check the bread type; the crate contains the wrong item” tells the learner what to examine. “Be more careful” provides little direction and can obscure a confusing order layout.

Ask the learner to explain their reasoning before supplying the answer. They may have used an old reference, misunderstood a term or correctly noticed a contradiction. Check the authoritative instruction where that matters. Coaching should help someone recognise the decision, not teach unquestioning imitation of a local workaround.

Sequence practice so that the learner can first perform the routine case, then recognise variation and ask for appropriate help. More difficult cases should not remove safeguards. For a consequential exception, competent performance may mean stopping and escalating rather than independently completing it.

When the learning comes from an experienced colleague, use the knowledge-transfer guide to capture the reasoning and boundaries behind the task. Keep materials current as practice reveals ambiguous instructions. An apparently complete reference often improves when another person tries to use it.

A hypothetical bakery transfer plan

The invented bakery agrees that a worker will complete supported checking at the next two suitable dispatch sessions, then attempt a different mixed order with a coach observing. The supervisor arranges coverage so that the first attempt is not squeezed between two unchanged priorities.

Moment	Learner’s task	Support and evidence
Before training	Explain the current checking routine and a difficult case	Coach identifies existing knowledge and barriers
During training	Rehearse a mixed order and identify a mismatch	Immediate task feedback using invented order data
First ordinary use	Perform the check with the current reference	Coach available; record support actually needed
Later review	Handle a different suitable order and explain escalation	Observer notes scope, conditions and unresolved gaps

The timings are illustrative choices, not evidence-based dosage. The relevant question is whether the learner has enough supported opportunity for the task and consequences involved. A complex judgement task may need a different sequence and a longer follow-up period.

Suppose the hypothetical first session reveals that a delivery change reaches the coordinator but does not reach the packing reference. The worker’s check is correct against an outdated order. The team should fix the information handoff and revisit affected cases, rather than label the learner’s performance a transfer failure.

After that correction, the coach may discover a genuine learning gap: the worker cannot explain which exception needs a supervisor’s approval. They use two contrasting scenarios and revisit the boundary. Record both issues separately, with their owners, so that the learning plan does not absorb every operational problem.

Treat 70/20/10 as a prompt, not a quota

The 70/20/10 shorthand associates learning with challenging experience, developmental relationships and formal coursework. A Center for Creative Leadership white paper presents the ratio in the context of management and executive development. It also discusses limits in specifying which experiences teach which lessons and in generalising earlier work across settings.

That context does not establish that every employee learns exactly 70% through work, 20% through other people and 10% through training. It does not justify allocating every course budget in those proportions or cutting formal instruction to a tenth of the programme. This is an inference about the source’s scope, rather than a claim that it experimentally disproves the ratio.

Use the three categories to ask practical questions. What must the person be taught explicitly? What can they practise in suitable work? Who can offer feedback or demonstrate judgement? The categories overlap: a coach can give formal instruction while the learner works on a real task.

For our bakery, choose explanation, supervised rehearsal, ordinary practice and a reachable reference because those activities serve the task. Calculating whether the minutes satisfy a 70/20/10 split would add an unsupported constraint. Some work also requires prescribed training; the shorthand cannot replace it.

Measure application separately from satisfaction and benefit

Decide what evidence will show the skill being used, and when it will be checked. A learner’s course rating tells you about their reaction. A quiz can test a limited set of knowledge. A task demonstration offers evidence of performance in its stated conditions. A business result answers another question again.

CIPD’s public summary of its learning-evaluation guidance links evaluation with identified performance gaps and business objectives. Applied here, define a task check that fits the intended performance rather than selecting a convenient completion statistic after the training.

Use several proportionate sources where possible: the learner’s account, a relevant work sample and observation by someone who understands the task. Record whether help was provided and whether the cases were comparable. Revisit later use if retention matters; an immediate successful rehearsal says little about application after a long gap.

Blume and colleagues’ meta-analysis of 89 studies reports positive relationships between transfer and factors including motivation and a supportive work environment. It also reports variation by skill type and measurement, including inflated relationships when predictor and outcome came from the same source and measurement context. These associations do not quantify the benefit of this bakery plan.

If dispatch errors decrease after training, examine other changes before attributing the result to training alone. A quieter week, revised order data or extra staffing could contribute. State the comparison and its limits. “The observed tasks used the checking routine” is a more defensible claim than “the training caused the entire improvement.”

Keep responsibility and follow-up visible

Make the learner’s next opportunity, available support and review date explicit. Ask what remains difficult and act on the answer. Keep support proportionate as independence develops, while retaining the normal escalation route.

Communicate any changed instruction through the change communication plan. A learner should not have to discover that the routine changed by being corrected for following the old version. Ensure the coach and supervisor are using the same current reference.

Protect the ability to ask for help. The conditions described in our psychological-safety guide matter when admitting uncertainty carries interpersonal consequences. Inviting questions while treating them as incompetence sends conflicting signals.

When the same barrier persists across several learners, examine the governing target or process using double-loop learning. More practice may be useful, but a queue design that makes the required check impractical needs an operational decision.

Evidence and limitations

The cited research supports a broader account of transfer, its relationships with individual and environmental factors and the importance of measurement. The 2010 and 2018 reviews were consulted through their publisher abstracts and available metadata; no effect size or detailed intervention result is inferred from inaccessible full text.

The bakery case, planning table and feedback sequence are hypothetical original guidance. They have not been validated as an intervention, and the suggested practice occasions are not a universal learning schedule. Different tasks, consequences and learners require different support.

Plan around a clear task, an opportunity to apply it and evidence from later work. Keep the diagnosis open: an application problem may concern knowledge, practice, conditions or authority, and each needs an owner who can actually address it.