Training Needs Analysis (TNA): A Step-by-Step Guide for L&D Teams

What is a training needs analysis?

A training needs analysis is a systematic way of finding out whether training is needed, for whom, and what it should achieve. It sits at the very start of the learning cycle, before any content is designed or bought. Its job is to answer one practical question: where is the gap between the performance we need and the performance we are getting, and is a lack of knowledge or skill the cause?

The classic framework comes from William McGehee and Paul Thayer, who in 1961 described training needs as something to be examined at three levels: the organisation, the operation (the tasks a job involves) and the individual. More than sixty years later the structure still holds, because each level answers a question the others cannot. Organisation analysis tells you where the business is going and which capabilities matter most. Task or role analysis tells you what good performance in a job actually looks like. Person analysis tells you who is falling short and by how much.

In many companies the word “analysis” is generous. Training requests arrive as solutions (“we need a negotiation course”), the learning team books the course, and nobody checks whether negotiation skill was the real constraint. A TNA reverses that order. It treats every request as a symptom to be diagnosed, which is why it is the first step in the learning and development cycle and the foundation for measuring whether training worked at all.

Why does a training needs analysis matter so much?

The most expensive training is the programme that was never needed. It consumes budget, it takes people away from their work, and when performance does not improve it damages the credibility of the learning function. A TNA protects against that in four ways.

First, it focuses spend on the gaps that matter. Every organisation has more possible training topics than it has money or time. A TNA forces a ranking based on business impact, so the sales team’s pipeline problem gets attention before a nice-to-have presentation skills workshop.

Second, it separates skill problems from everything else. A large share of performance problems have nothing to do with what people know. Targets are unclear, tools are slow, processes contradict each other, or the incentive system rewards the wrong behaviour. Training cannot fix any of these, and sending people on a course while the real cause stays in place guarantees disappointment.

Third, it creates the baseline for evaluation. You cannot show improvement without knowing the starting point. The metrics you gather during the TNA (error rates, time to competence, customer scores, sales conversion) become the before picture that the Kirkpatrick and Phillips evaluation models compare against later.

Fourth, it builds sponsorship. When line managers help define the gap and agree how success will be measured, they are far more likely to release people for training and reinforce the new behaviour afterwards. That manager support is one of the strongest predictors of whether learning transfers to the job.

What happens at each of the three levels?

Organisation analysis

Organisation analysis links learning to strategy. Start with the business plan, not the training catalogue. What is the company trying to achieve in the next one to three years? Entering a new market, launching a product line, automating a process, improving margins, meeting a new regulation? Each goal implies capabilities the organisation will need more of. This is also where you look at strategic workforce planning data: which roles will grow, which will shrink, and where you plan to build skills internally rather than hire them.

Useful inputs at this level include the strategic plan, departmental objectives, KPI dashboards, customer complaint trends, audit and compliance findings, exit interview themes and engagement survey results. The output is a short list of priority capability areas, each tied to a business outcome a senior leader cares about.

Role or task analysis

Role analysis describes what good performance looks like in a specific job. It breaks the role into its key tasks and, for each one, the knowledge, skills and behaviours required to do it well. Job descriptions are a starting point but are usually too generic. Better sources are observation of top performers, interviews with managers, process documentation, quality standards and competency frameworks.

A practical technique is to ask managers to describe the difference between their best performer and an average one in concrete terms. “She qualifies leads before booking demos” is useful. “He has a great attitude” is not. Where the role is changing, for example because of new software or AI tools, the analysis should describe the future version of the task, not only the current one.

Person analysis

Person analysis compares each individual, or each group, against the standard defined in the role analysis. Sources include performance reviews, skills assessments, manager ratings, work samples, quality data, certification records and self-assessments. Self-assessment alone is unreliable, because people tend to overrate skills they rarely use and underrate skills they use daily, so combine it with at least one objective measure.

The output is a picture of where the gaps sit: across a whole team, in one location, among new hires, or in a few individuals. That distinction matters because a gap shared by everyone points to a design or process issue, while a gap in a handful of people is a coaching or development issue.

How do you run a training needs analysis step by step?

The steps below work for a single team or a whole organisation. Scale the effort to the size of the decision: a two-week check is enough for a £10,000 workshop; a multi-year leadership programme deserves more.

1. Define the business problem. Write one sentence describing the result that is not happening. “Customer onboarding takes 34 days against a target of 20” is a problem. “Staff need project management training” is a solution in disguise.
2. Agree the performance standard. With the line manager, describe what good looks like and how it will be measured. This becomes your evaluation target.
3. Gather evidence. Use a mix of data (KPIs, quality metrics, system reports) and conversation (manager interviews, focus groups, observation). Two or three sources that agree are worth more than one large survey.
4. Identify the gaps. Compare current performance with the standard at role and person level, and record the size of each gap.
5. Diagnose the cause. For each gap, ask whether it is caused by missing knowledge or skill, or by something else. The test in the next section helps.
6. Prioritise. Rank the learning gaps by business impact, number of people affected, urgency and cost of closing them. A simple 1–5 score on each factor is usually enough.
7. Recommend solutions. For each priority gap, propose the lightest intervention that will work: a job aid, coaching, a stretch assignment, microlearning, or a formal programme. Remember the 70-20-10 principle: most development happens on the job.
8. Set success measures. For each recommendation, record the behaviour you expect to change and the business metric it should move, with a baseline and a date.

Is it really a training problem?

This is the single most valuable question in the whole analysis. Performance consultants often use a blunt test: if this person’s life depended on doing the task correctly, could they do it? If the answer is yes, the problem is not a lack of skill, and training will not solve it.

Thomas Gilbert’s Behavior Engineering Model, published in the 1970s, is a more structured version of the same idea. It separates environmental factors (information, resources, incentives) from individual factors (knowledge, capacity, motives), and argues that most performance problems sit in the environment, not the person. In practice, check these causes before you approve any training request:

Which data collection methods work best?

No single method gives a complete picture, so most TNAs combine two or three. Choose based on how many people are involved, how much time you have, and how sensitive the topic is.

• Performance data and KPIs. The most objective source. Error rates, sales conversion, handle times, audit findings and customer scores show where results fall short. They tell you where the problem is, but rarely why.
• Manager interviews. Fast and rich in context. Managers know who struggles and with what. Use a consistent set of questions so the answers can be compared across teams.
• Focus groups with employees. Good for understanding causes. People will often tell you the system is the problem before management will.
• Surveys and skills assessments. Useful at scale. Keep them short, ask about specific tasks rather than broad competencies, and pair self-ratings with manager ratings.
• Observation and work samples. The best way to see the gap between how work is supposed to be done and how it is actually done. Time-consuming but highly reliable.
• Skills data in your HR systems. Many HRIS and learning platforms now hold skills profiles. Treat them as a starting point; they are only as good as the data people entered. If you are choosing a platform, the LMS comparison covers which tools support skills mapping.

What should a training needs analysis report include?

Keep the report short enough that a busy executive will read it. Two to five pages, plus appendices for the detail, is usually right. A practical structure:

1. The business problem and its cost. One paragraph, with numbers where possible.
2. The performance standard. What good looks like and how it is measured.
3. Findings by level. Organisation, role and person, with the evidence behind each.
4. Gap diagnosis. Which gaps are learning gaps, and which need a non-training fix and an owner.
5. Prioritised recommendations. A ranked table: gap, audience, proposed solution, estimated cost, expected outcome.
6. Evaluation plan. The baseline metrics, the target and when you will measure again.

The evaluation plan is the part most often missing, and it is the part that makes the difference later. Agreeing the success measure before the programme starts is what turns training from a cost into an investment you can defend.

How often should you run a training needs analysis?

Think of TNA as two activities. The first is an annual or strategic review that runs alongside business planning, looks across the whole organisation and shapes the learning budget for the year. The second is a targeted analysis triggered by an event: a new system, a regulation, a merger, a spike in errors or a training request from a manager. Targeted analyses can be done in days rather than months.

Organisations with a strong learning culture increasingly run a lighter, continuous version: managers discuss development needs in regular one-to-ones, skills data is updated as people complete projects, and the learning team reviews trends each quarter. That rhythm catches gaps earlier and avoids the large, slow surveys that are out of date by the time they are published.

How does TNA connect to the rest of learning and development?

A needs analysis is only the first step. Its output feeds programme design, where you decide the blend of on-the-job experience, coaching and formal learning. It feeds the choice of delivery, from microlearning for quick, frequent reinforcement to longer leadership development programmes for complex capabilities. And it feeds evaluation, because the baseline and success measures you set now are what you will compare against later.

If you are building an L&D function from scratch, start with a small, well-run TNA on one business problem that leaders care about. A visible win there earns the credibility to apply the same discipline across the organisation. You can find the rest of the cycle, from design to evaluation, in the Learning & Development hub.

Frequently Asked Questions