Training Needs Analysis When The Solution Is Already Chosen

A training needs analysis finds the gap between current and required results, then tests whether training should close it.

With a request that already names a course, work in four steps.

State the result gap with the sponsor, rule out four causes training cannot reach, gather operational evidence, and recommend a means with a measure.

Most requests reach the learning function in that form, such as a leadership programme for the top fifty or a communication workshop before a restructure.

By then the sponsor has chosen the intervention and assumed the budget.

The analysis converts the request back into a result the business owns.

What A Need Is

Kaufman and Guerra-Lopez define a need as a gap between current results and desired results, together with the consequences of that gap.

Used as a verb, as in "we need more training", the word selects the solution before anyone has stated the problem.

They call an assessment that starts from a chosen method a quasi needs assessment, because training is a means and the result it should change is still unnamed.

The familiar training needs assessment belongs to that group, since it begins after training has been picked.

The analysis a sponsor should pay for begins one step earlier, with the result that is missing.

The labels learning needs analysis and training needs assessment overlap heavily in practice, and the starting point matters more than the label.

Three Levels Of Analysis

Most guides to training needs analysis name three levels, which are organisational, task and individual.

At the organisational level the question is which results the business is trying to produce over the next year or two.

At the task level it is what the work requires of the person doing it.

At the individual level it is who falls short of that requirement, and by how much.

Rummler and Brache, as Gupta, Sleezer and Russ-Eft describe them, diagnose performance at the organisation, the process and the individual job instead.

A failure at any one of those levels can limit what the organisation achieves.

A request nearly always arrives at the individual level, as in "these people need training", while the cause often sits in a process.

The Sponsor Conversation

The first question is who is asking and whether that person owns the result.

Barbazette defines a client as someone with both a need to be involved and the authority to make the critical decisions.

She notes that the person who made contact may be only a messenger.

A request passed on by an HR colleague for an operations director means the director has to attend the second meeting.

Her next step is to agree the result wanted with no solution agreed, plus the boundaries of the assessment and the date the plan will be presented.

A sentence that opens that conversation is "Before we design anything, I would like two weeks to confirm what is driving the result, so that what we build closes it."

Six questions carry the conversation.

• What result is not being achieved

• How do we know, and from which source

• Which population, and where is the gap widest

• What would look different if the gap closed

• What has already been tried here

• What else could be causing this

The sixth question keeps the analysis honest, because most performance gaps have causes that training cannot reach.

People Management Insight surveyed more than 250 HR and L&D professionals for MHR in 2026.

Of those, 22 per cent described senior leaders as highly involved and visible in L&D.

Another 52 per cent described the leaders as supportive but not closely involved.

MHR sells a learning platform.

The conversation therefore has to be short, structured and booked in advance.

Causes Training Cannot Reach

Mager and Pipe begin by asking whether the gap is important enough to pursue, and they drop it if the evidence says no.

Then they ask whether the person could do the task if really required to do it.

Where the answer is yes, the skill exists, and training cannot supply what is already there.

Gilbert reaches the same place from the other side, with six conditions for performance in two groups.

Data, instruments and incentives sit in the environment, and knowledge, capacity and motives sit in the person.

He considered the first three the most likely to improve performance, because the environment is easier to change than the person.

Kaufman and Guerra-Lopez cite an estimate that instruction suits roughly 20 per cent of performance problems, all things being equal.

Four causes need an explicit answer before design starts.

• Clarity. People do not know what good looks like, or the standard changed without notice

• Conditions. The process, tools or caseload make the wanted behaviour slower than the unwanted one

• Consequences. The wanted behaviour costs the person time or standing, while the unwanted one produces results

• Time. The skill exists, and competing demands leave no hours to use it

Mager and Pipe's questions on obstacles add a lack of authority or tools, and policies that follow a way of working that ought to change.

Barbazette builds a fictional scenario from a real client exchange, and it shows these causes at work.

An information systems manager asked for time management training for about sixteen operators, programmers and analysts, because co-workers on the other twelve-hour shift complained.

Diagnosis found that the department clerk had not been replaced during a hiring freeze.

The case log was largely ignored, and newer analysts had never been trained in handing off cases.

The standard was to clear user requests within 48 hours, and it was met 70 per cent of the time.

The recommendation was a revised orientation, case management training from a master performer and a job aid, and the time management class was dropped.

Evidence Without A New Survey

A questionnaire is quick, and it records wants, ordered by whichever topic was discussed most recently.

Stronger evidence usually exists in systems the business already runs.

• Error rates, rework and cycle times

• Complaint categories and audit findings

• Win and loss reasons from the pipeline

• Two hours of observation beside the person doing the work

• Interviews with people who meet the standard and people who do not

• Interviews with the managers who receive the output

• Exit and onboarding records

Observation shows the workarounds, the screens that slow the task and the point where people stop following the process.

A survey is useful later, for reach and for testing a hypothesis that already exists.

In the MHR survey, 67 per cent of respondents reported L&D impact through employee surveys and 55 per cent through completion rates.

Only 20 per cent used business outcomes such as sales, customer satisfaction or quality.

The figures sum well above 100 per cent, so respondents named several methods.

An analysis that starts from business data gives the later evaluation something other than satisfaction scores to measure.