Training needs analysis: how to find the gaps worth training

What a training needs analysis is, and when to run one

A training needs analysis (TNA), also called a training needs assessment, is the work you do before you build or buy training. It answers three questions: what needs to be trained, who needs it, and whether training will fix the problem at all.

In a major review of the science of training, Salas and colleagues describe it as a diagnosis of what needs to be trained, to whom, and in what organizational system. You don’t need one every week. Run one when something changes:

• New hires are joining a role, and you need to know what they must be able to do in their first months.
• A new system or process is coming, such as a new CRM, an ERP or a new safety rule.
• A team’s results have dropped, and someone has asked for “a course” to fix it.
• It’s time for the annual training plan and budget.
• You run a training center and a corporate client sends a brief: the analysis turns their request into a program you can plan and deliver.

Why it gets skipped, and what that costs

In practice, the analysis is the step teams skip. Someone asks for a course, the deadline is close, and the course gets built. The research shows how common this is. When Arthur and colleagues pooled 397 results from 162 sources on training in organizations, only 6% of them (22 of 397) came from studies that reported any needs assessment. The authors note that many more may have done one without writing it down, but the pattern is hard to miss.

The same meta-analysis found that training, on average, works: effect sizes of 0.60 to 0.63 depending on what was measured, which the authors call a medium to large effect. So the question is rarely whether training can work. It’s whether this training targets the right gap. Salas and colleagues make the same point: what happens before and after training matters just as much for program success as the training itself.

It isn’t a niche idea, either. ISO 10015:2019 gives organizations of any type or size guidelines for competence management and people development. ISO 30422:2022 gives guidance on organizing workplace learning for both short-term operational needs and long-term skills needs. In Saudi Arabia, one of the three pillars of the Human Capability Development Program under Vision 2030 is lifelong learning, focused on upskilling and reskilling citizens who have finished their education. A good TNA is how one team’s plan connects to goals like these.

The three levels of training needs analysis: organization, task and person

The classic model, the one Arthur and colleagues used, looks at three levels: organizational analysis, task analysis and person analysis. The US Office of Personnel Management uses the same three, calling them organizational, occupational and individual assessment. Each level has its own questions and its own evidence.

Start at the top. If you start with the person, you tend to train people in what they ask for, not in what the business needs.

• Organization: Which business goals can training help reach? Where is it needed most? What at work will help, or block, people using what they learn? Evidence: strategy and team goals, audit findings, new systems on the way.
• Task or role: What must someone in this role do well? Which tasks are frequent, critical or risky? What does “good” look like, in terms you can observe? Evidence: job descriptions, procedures, interviews with top performers.
• Person: Who needs training, and in what? Who already meets the standard and can skip it? Evidence: test results, work samples, manager observation, performance data.

Is it a training problem at all?

Arthur and colleagues define needs assessment as asking whether the organization’s needs, objectives and problems “can be met or addressed by training.” Sometimes the honest answer is no.

A quick check: could the person do it if their job depended on it? If yes, it isn’t a skill gap. Look at the process, the tools, the incentives or the workload instead. A course won’t fix a broken form, a slow system or a target that rewards the wrong thing. If no, and the task matters, training can help.

Often it’s both. A new CRM needs training, and it also needs fields that make sense and a manager who checks the pipeline in it. Write the non-training fixes next to the training ones, and give each one an owner.

How to conduct a training needs analysis, step by step

Keep it small enough to finish. For one role or one change, a few days of work is usually enough. A short analysis you act on beats a long one that sits in a folder. These six steps move from the top level down:

Where the evidence comes from

Use more than one source. Each one sees part of the picture, and where two sources disagree is often where the real problem is.

• Manager interviews: what good work looks like, where people struggle, and what has changed.
• Surveys and self-assessments: fast and wide, and good for what people want to learn, but weaker as proof of what they can do.
• Performance data: sales, errors, support tickets, audit findings, and time to reach full speed in a role.
• Customer reviews and complaints: tagged by theme, they show which gaps customers feel. Our post on training needs assessment shows how to read them.
• Observation and work samples: watch the task being done, or review real output.
• Existing assessment results: if your learning platform records test answers by skill, you already have a map of where each group is weak.

A worked example: one support team

Here is a short analysis for one team, from start to finish. The company and the numbers are made up, but the shape is real.

Harbour Logistics moves its 12-person customer support team to a new ticket system. Emma Clarke, who leads the team, is asked for “training on the new system”. Before booking anything, she spends two days on an analysis.

• Goal: 22% of customers write again about the same order. The target is 15% within three months.
• Tasks: log a ticket, find an order’s status, explain a delivery delay, and escalate a damaged-goods claim. Escalation is rare but costly, so it is marked critical.
• Required level: every agent at 3 (on their own) for all four tasks.
• Current level: a 10-question quiz by task, and a review of 40 closed tickets. Everyone logs tickets well. Five of the 12 agents escalate damaged goods wrongly, and most repeat contacts come from delay questions.
• Gap and cause: escalation is a real skill gap. The delay answers are a different problem: the saved reply for delays still describes the old process.
• Plan: a 20-minute course and a practice conversation on escalation for the five agents, a one-page job aid, and a rewritten saved reply this week. Repeat contacts are checked again after six weeks. Only one of the two gaps needed a course.

From gaps to a plan: objectives and formats

Each gap becomes a learning objective you can measure: not “understand the new CRM” but “log a new lead with every required field in under two minutes”. Decide now how you will check it, so the evaluation is planned before the training is built. Our post on how to measure training effectiveness shows how.

Then match the format to the gap. Knowledge gaps suit a short course and a quiz. Skill gaps need practice and feedback. Tasks people do once a month may need a job aid more than a lesson. Group the courses for each role into a program, and add live cohort sessions where people need to practice together.

Training needs analysis template and format

A training needs analysis format is a table you can sort and share. Our free template is an Excel file with four sheets that follow the steps above: Organization goals, Roles & tasks, Skills gap, and Plan. It has drop-down lists and works out the gap for you. Two grey sample rows show how to fill it in.

The Skills gap sheet is the heart of it: one row per role and skill. Here are its ten columns, filled in with the support team above:

• Role: who it’s for. “Support agent”.
• Skill: one thing you can observe. “Escalate a damaged-goods claim”.
• Required level: on the 1 to 4 scale. “3, on their own”.
• Current level: what the evidence shows. “2, with help”.
• Gap: required minus current, worked out by the sheet. “1”.
• Evidence source: where the current level came from. “Quiz and ticket review”.
• Priority: high, medium or low, from impact and urgency. “High”.
• Training type: the format that fits the gap, or “not training”. “Short course and practice conversation”.
• Owner: a role, not a name, so the sheet survives a reshuffle. “Team lead”.
• Due date: when the gap should be closed and checked. “Six weeks from today”.

Technical and maintenance training needs analysis

For technical and maintenance roles, start from the equipment and the tasks, not from a course catalogue. List the assets each team looks after, the tasks on each one, and what happens if a task is done badly: an injury, downtime or a failed inspection.

Tasks a regulation names go to the top, whatever a survey says. In the US, for example, OSHA’s lockout/tagout standard requires training for employees who service equipment, and retraining whenever their job assignment, the machines or the energy control procedures change. Check the rules that apply where you work.

The best evidence is already in your systems: work-order history, repeat failures on the same asset, near-miss reports and audit findings. Watch the task where you can, and record the level per person and per task, because a technician can be an expert on one machine and new to the next.

IT training needs analysis

IT skills shift with every new platform, so separate the tool from the fundamentals. A move to a new cloud service is a short, practical gap. Weak troubleshooting or security habits are a deeper one, and a tool course won’t close them.

You don’t have to write the task list from a blank page. The NICE Framework from NIST (SP 800-181r1) describes cybersecurity work as Task, Knowledge and Skill statements. Pick the ones that fit each role and use them as the rows of your skills gap sheet.

For evidence, look at reopened tickets, escalations, failed changes and incident reviews, next to a short hands-on test. A certificate shows what someone studied, not always what they can do on your systems.

Manager training needs analysis

Leadership training is where the analysis shows up in the research. In a meta-analysis of leadership training, Lacerenza and colleagues found that programs built on a needs analysis led to more learning and more transfer to the job.

Start from what managers in your company must actually do: run one-to-ones, give feedback, hold performance reviews, hire and handle conflict. Then gather evidence from more than the manager: their own manager, upward or 360-degree feedback from the team, and team data such as turnover or engagement survey results.

For team training needs, look for the gaps a whole team shares. If most of a team scores low on the same skill, the cause is often how the team is set up or led, not each person.

How Classbase is designed to help

Classbase doesn’t do the analysis for you, and it shouldn’t: the judgment about goals and gaps belongs to your team and your managers. Where it helps is the step after.

Once the gaps are clear, describe the course and attach your manuals or decks, and the AI drafts the outline and then each module as short, interactive cards. Nothing reaches a learner until an author keeps it and publishes. Programs group courses for a role into one path that opens section by section, assigned once. Every test answer counts toward a skill, and reports answer plain questions from your own data, so your next analysis starts from evidence.

If you have a list of gaps and want to see it turned into a program, book a demo and we’ll show you.