A manager asks you to create a communication skills course. It sounds like a straightforward request. You find a trainer, prepare the material, choose a date, invite employees, and track attendance.

The course receives positive feedback. Most employees complete it. The manager seems pleased. But a few months later, the communication problem is still there.

This happens because developing training programs is often treated as a content project. In reality, it is a performance improvement process.

A successful training program does not begin with slides, videos, or an LMS. It begins with a business problem, a clear understanding of what employees need to do differently, and a plan for helping them apply the new skill at work.

This guide walks through the full process of developing training programs that employees find useful, and organizations can connect to real results.

Understand the performance problem before choosing training

A training request is a starting point for investigation. Before designing the program, establish what needs to improve and whether learning can address the cause.

Connect the request to a business need

“We need leadership training” could mean several things. New managers may struggle to delegate, give feedback, or support employees through a change.

Each problem calls for different learning. Ask the business leader what is happening, why it matters now, and what successful performance would look like.

The pressure to build capability is substantial. In the World Economic Forum’s Future of Jobs Report 2025, employers expect 39% of existing skill sets to change or become outdated by 2030. Among those surveyed, 63% identify skills gaps as a major barrier to transformation, and 85% plan to prioritize upskilling.[1]

Those findings provide context. The organization still needs to identify its own priorities rather than turn every emerging skill into another course.

Check whether the barrier is a skill gap

Employees may know what to do but lack the authority, information, tools, or time to do it. Conflicting instructions and incentives can also prevent the expected behavior.

Consider an illustrative customer service team receiving more complaints. The cause may be weak complaint-handling skills, but it could also be a product defect or a policy customers find unreasonable.

Training might improve how employees respond. It cannot remove the underlying defect or change their decision authority.

Ask whether better knowledge or skill would make the task achievable. Where the answer is only partly yes, agree on the operational changes needed alongside development.

Compare organizational needs, role requirements, and employee capability

Needs analysis should connect the business objective with the work required and the employee’s starting point. CIPD describes learning needs analysis as an ongoing process and distinguishes it from simply collecting requests for training.[2]

Review actual work. A proposal, recorded handover, quality review, or observed conversation can reveal where performance falls short.

Combine evidence where appropriate. Employee interviews can explain difficulties; observations can show what happens; operational data can reveal how often the problem occurs.

Also identify the intended audience. First-time managers, experienced leaders, frontline employees, and remote teams may face different tasks and constraints, even when the training topic is the same.

Define the outcome and the evidence needed to assess it

Clear outcomes keep the content, activities, and assessments connected. Establish the target and the starting point before deciding how much material to produce.

Describe what employees should do differently

“Improve communication” is too broad to guide design. Specify whether the employee needs to write a clearer handover, ask better discovery questions, or conduct a useful feedback conversation.

For a feedback program, managers might need to describe an observed behavior, explain its effect, ask for the employee’s perspective, and agree on an action and follow-up.

These behaviors provide a practical destination. They also help the learner understand why the program matters.

A learning objective should state the task, relevant conditions, and expected standard. For example, an employee might complete a standard refund in the training system without assistance and in line with the approved procedure.

Record a baseline linked to the original problem

Choose measures that reflect the difficulty being addressed. An onboarding program might track time to independent work; a quality program might track a defined error rate.

Clarify how the measure is calculated and whose performance it covers. Comparisons become less useful if the task, standard, or participant group changes without explanation.

A baseline does not prove what caused later improvement. It gives the organization a starting point against which to examine change.

Where historical operational data is unavailable, an initial assessment or work sample can still establish participants’ current capability.

Design assessment before filling the course

The assessment should test the intended outcome. A knowledge test can check recall; a role-play or work sample can examine performance.

If employees need to negotiate, let them respond to a realistic objection. If they need to use a system, ask them to complete the relevant workflow.

Define criteria in advance and make them understandable to learners. The assessment should gather useful evidence rather than surprise employees with unrelated difficulty.

Workplace follow-up serves a separate purpose. Performing a task during training shows capability under those conditions; later observation helps establish whether the skill is being used at work.

Build learning around the task and opportunities to practice

Once the outcome is clear, select the content and methods that support it. The program should give employees enough explanation to begin and enough practice to improve.

Include what employees need to perform

Subject experts often have more knowledge than a learner needs for the immediate task. Work with them to identify the information essential for correct performance.

Keep required procedures and decisions in the core learning. Provide supporting detail through references, job aids, or follow-up resources.

Background information can be useful, but it should earn its place. Ask what the learner would struggle to do if it were removed.

A training program may combine courses, coaching, assignments, and assessment. The structure comes from how those activities prepare employees for the outcome.

Match the method to the skill

A demonstration can show a process. Guided practice allows employees to attempt it. A checklist can support the same task after training.

Conversation skills need interaction and feedback. Technical skills may require a lab or supervised practice. Decision-making can be explored through cases and scenarios.

CIPD’s guidance considers a range of learning methods and the factors that should inform their selection.[3]

The practical choice should follow the task, learner needs, and working environment.

Language, accessibility, available devices, and learning time also matter. A format that works for office employees may be difficult to use during a busy frontline shift.

Make practice realistic and feedback specific

Use the tools, information, constraints, and decisions employees will encounter. Start with manageable tasks and increase difficulty as learners become more capable.

A procurement employee might practice responding to a supplier who rejects a proposal. A multiple-choice question about negotiation principles would test a different capability.

Feedback should explain what worked, what needs to change, and what to try next. “You clarified the customer’s concern before offering a solution” gives the employee a behavior to repeat.

Provide another attempt where the skill requires it. A single activity may reveal a difficulty without giving the learner enough opportunity to address it.

Plan workplace application before the program begins

Learning needs a place to be used. Agree on the application task and manager support while designing the programme, rather than leaving them until the course ends.

Give managers a manageable role

Before training, managers can explain expectations, identify an application opportunity, and arrange the time needed for learning and practice.

Afterwards, they can observe the task, discuss progress, provide feedback, and remove barriers. Where necessary, they can arrange further practice.

Make these expectations specific. For a feedback programme, the manager might review the participant’s plan before a development conversation and discuss what happened afterwards.

The employee should know where and when the skill will be used, who can help, and how progress will be reviewed.

Support the task after formal learning

Job aids, coaching, peer discussion, and follow-up assignments can help employees apply what they learned. Choose support that addresses the likely difficulty.

A checklist may help with a complex procedure. A case discussion may help employees compare decisions. An observed task may provide more useful feedback than another general lesson.

Check whether the workplace supports the intended behavior. Processes, targets, and access permissions may still need adjustment.

Recognition should reflect useful application. Completing content is worth tracking, but improvement at work needs its own attention.

Build a review into the program

Application can reveal needs that were not visible during design. Employees may understand the concept but struggle with an exception or a tool.

Collect those findings and use them to revise the support. A clearer example, updated reference, or additional practice task may be more useful than repeating the whole course.

Agree on review points suitable for the skill. The timing should allow employees a realistic opportunity to apply it.

Test the design and prepare for delivery

A program needs to work for its users as well as on paper. Test a small part first, then examine the full experience before expanding delivery.

Prototype one meaningful part

A prototype could be a lesson, scenario, assessment, or short version of the learner journey. Choose something that tests an important design assumption.

Ask employees, managers, and subject experts to review it. Check accuracy, instructions, difficulty, relevance, and whether it can be completed in the available time.

Look closely at the assessment. If it does not reveal the capability you intended, revise it before developing more material around it.

The process can be iterative. New evidence may require a change to the objective, activity, or delivery method.

Pilot the program with the intended audience

A pilot should test learning, delivery, and workplace application. Include participants whose experience and working conditions reflect the intended group.

Gather evidence about assessment performance, technical difficulties, timing, support needs, and manager involvement. Ask what participants used at work and what prevented application.

Satisfaction is useful feedback, but it should not be the only question. A well-liked program can still fail to address the original problem.

Use the findings to decide what needs revision before rollout.

Agree on resources, responsibilities, and communication

Confirm participant selection, facilitator preparation, access to systems, support routes, and ownership of assessment and evaluation.

Communication should explain why the program exists, what employees will be expected to do, and how it relates to their work.

Decide which material to build and which to buy. Internal content may be necessary for company procedures and systems; reliable external material may support broadly applicable skills.

A combined approach can work well when purchased content is connected to internal examples, assignments, and manager discussions. Compare quality, relevance, maintenance, and full cost rather than the initial price alone.

AI can help prepare drafts, organize source material, or suggest scenarios. Review the output for accuracy, confidentiality, accessibility, and alignment with the required task before using it.

Evaluate performance and the full cost of the program

Evaluation should answer whether the program achieved its intended outcome and what needs to change next. Participation, learning, workplace application, and business results provide different parts of that answer.

Measure beyond attendance and completion

Participation records establish whether the intended employees took part. Assessment examines whether they developed the required capability.

Work samples, observations, quality reviews, or suitable system data can show application. Operational measures help examine whether the original performance problem improved.

CIPD’s evaluation guidance links learning evaluation to the performance gaps that informed the intervention.[4]

This is why measures and baseline evidence should be considered during design.

If learning improves but application does not, investigate the working conditions and support. If application improves but the business measure remains unchanged, check the assumed connection and other influences.

Include employee time in the budget

The cost extends beyond a trainer or platform. Analysis, design, expert review, delivery, administration, assessment, and updates all require resources.

Include learner time and manager support. An illustrative two-hour program for 1,000 employees uses 2,000 working hours before preparation and follow-up.

That calculation does not establish whether the program is worthwhile. It makes the resource commitment visible and encourages careful choices about scope.

Development time also depends on complexity, available material, approvals, translation, and testing. Estimate the complete work rather than only the minutes of content required.

Calculate ROI when the benefits can be estimated credibly

Financial ROI compares attributable program benefits with the total program cost:

ROI (%) = [(Financial benefits − Total program cost) ÷ Total program cost] × 100

In an illustrative example, a program costs $60,000 and produces estimated attributable benefits of $105,000 in additional profit. The net benefit is $45,000, giving an estimated ROI of 75%.

The calculation depends on the quality of the benefit estimate. Revenue, productivity, and retention can change because of pricing, demand, staffing, systems, and other factors.

Use comparisons where practical and make assumptions explicit. Manager or participant estimates can inform the analysis, but should not be treated as proof of causation.

Some programs are better assessed through readiness, quality, risk, or required standards. A financial figure is useful only when the underlying estimate is defensible.

Keep the program connected to the work

Program development continues after launch. Review the design when roles, systems, policies, customer needs, or business priorities change.

Use evidence to identify what needs revision

Learner feedback may reveal unclear instructions. Assessments may reveal a weak activity. Workplace observations may show that employees need a better reference or more practice.

A course can remain factually correct while becoming less useful to the role. Review relevance as well as accuracy.

Avoid rebuilding everything by default. Change the part that the evidence shows is failing.

Return to the original performance question

Before expanding or renewing the program, check whether its purpose remains clear. The organisation should be able to explain the problem, target behavior, support provided, and evidence of change.

If those elements are missing, more content is unlikely to resolve the uncertainty.

The strongest test comes at work: can employees handle the task more effectively, and does the organization give them the conditions to keep doing so?

Sources

[1] World Economic Forum: The Future of Jobs Report 2025 Digest. Employer expectations across industries and economies.

[2] CIPD: Learning needs analysis. Guidance on identifying learning and development needs.

[3] CIPD: Methods of delivering learning interventions. Guidance on selecting learning methods.

[4] CIPD: Learning evaluation, impact and transfer. Guidance on evaluation and workplace transfer.