Training needs analysis examples: six roles, worked through

How to read these examples

Our guide to training needs analysis explains the method: three levels (organization, task and person) and six steps. This post is the other half: six worked examples, one per role, so you can see what a finished analysis looks like and copy the parts that fit your team.

The examples are illustrative. The teams and numbers are made up to show the shape of the work, not research findings. Each one follows the same pattern:

• The business goal: one result, in one line. This is the organization level.
• Four to six skills the goal depends on, each with a required and a current level on a 1–5 scale: 1 aware, 2 learning, 3 can do it, 4 does it well, 5 can teach it. This is the task and person level.
• The evidence behind each current level: observation, a test, system data or a manager’s review. Never a guess on its own.
• The top gap, its cause, and the training (or other fix) that closes it.

Example 1: a people manager

A logistics company promoted eight team leads in a year, most of them its best operators. Six months on, two of their teams have the highest staff turnover in the business. Goal: bring turnover in those two teams down to the company average within a year.

Evidence: an anonymous upward survey from each team, exit-interview notes from HR, a review of six recent performance reviews, and a conversation with each lead’s own manager. Leadership training is one place where research backs this step: in a meta-analysis, Lacerenza and colleagues found that programs built on a needs analysis worked better, as did programs with practice, feedback and sessions spread over time.

• Run a weekly one-to-one that ends with an agreed action: required 4, current 2. Evidence: upward survey.
• Give specific feedback on a mistake, the same week: required 4, current 2. Evidence: upward survey and exit interviews.
• Write a performance review with examples and goals: required 3, current 2. Evidence: review of six recent reviews.
• Plan shifts and workload fairly: required 4, current 4. Evidence: manager review and rota data. No gap.
• Handle a conflict between two team members: required 3, current 2. Evidence: HR case notes.
• Top gap and fix: feedback and one-to-ones, the two skills exit interviews name most. A short course on both, three practice conversations spread over six weeks, and each lead’s manager sitting in on one real one-to-one. The upward survey runs again after three months.

Example 2: a maintenance technician in manufacturing

A packaging plant has 18 maintenance technicians across three shifts. Unplanned stops on its two filling lines have risen for three quarters in a row, and the same faults keep coming back. Goal: cut unplanned downtime on those lines by 25% this year, with no lockout incidents.

Evidence: a year of work-order history from the maintenance system, repeat failures by asset, near-miss reports, and a supervisor watching each technician do two tasks. Levels are recorded per person and per machine, since someone can be an expert on one line and new to the other. Tasks a regulation names go to the top, whatever else the data says. In the US, for example, OSHA’s lockout/tagout standard requires retraining whenever a job assignment, the machines or the energy control procedures change.

• Apply lockout/tagout on the filling line: required 4 for everyone, current 3 for the two newest technicians. Evidence: observation. Closed first, whatever its score.
• Find the root cause of a repeat failure: required 4, current 2. Evidence: work orders show the same seal replaced 11 times on one filler.
• Read the line’s electrical drawings to trace a fault: required 3, current 2. Evidence: supervisor observation.
• Do the weekly preventive maintenance checklist: required 4, current 4. Evidence: work-order completion data. No gap.
• Close a work order with the cause and parts recorded: required 4, current 2. Evidence: half the closed orders say only “fixed”.
• Top gap and fix: root cause on repeat failures. A short course on a simple method such as “5 whys”, then practice on the plant’s own five most frequent faults with a senior technician. The work-order gap is partly a form problem: make cause and parts required fields. Repeat failures are checked again after a quarter.

Example 3: an IT team

A 10-person IT team is moving the company’s file servers and email to a cloud service. Since the first wave, tickets reopened within a week have doubled, and one change took the shared drives offline for a morning. Goal: finish the move with no unplanned outage, and bring reopened tickets back under 5%.

Evidence: the ticket system (reopened tickets and escalations by category), the change log and the outage review, and a one-hour hands-on lab in a test environment. The skill rows don’t have to start from a blank page: NIST’s NICE Framework describes technical work as task, knowledge and skill statements you can borrow. Keep the tool apart from the fundamentals. A course on the new admin console won’t fix weak troubleshooting.

• Set up user access and permissions in the new service: required 4, current 2. Evidence: hands-on lab.
• Troubleshoot a sync or sign-in fault step by step: required 4, current 3. Evidence: reopened tickets.
• Plan a change with a tested rollback: required 4, current 2. Evidence: change log and outage review.
• Write a knowledge-base article users can follow: required 3, current 2. Evidence: review of ten recent articles.
• Spot and report a phishing email: required 4, current 4. Evidence: the last simulated phishing test. No gap.
• Top gap and fix: access setup and change planning, both tied to the outage. A hands-on course in the test environment, built from the vendor’s admin guide and the team’s own runbook, and a rule that every change needs a written rollback before it is approved. Reopened tickets are checked after the next wave.

Example 4: a sales team

A software company’s 14 account executives book plenty of first meetings, but few turn into proposals. The sales director asks for “negotiation training”. Goal: raise the share of first meetings that reach a proposal from 20% to 30% in two quarters.

Evidence: CRM data on where deals stall, 30 recorded first calls scored against a short checklist, and two days of a manager joining calls. The data points somewhere other than the request: deals stall right after the first meeting, long before anyone negotiates.

• Ask discovery questions that find the buyer’s real problem: required 4, current 2. Evidence: call recordings.
• Tailor the demo to that problem: required 4, current 3. Evidence: call recordings and the manager’s notes.
• Agree a dated next step before the call ends: required 4, current 2. Evidence: the CRM, where most deals have no next step.
• Handle a price objection: required 3, current 3. Evidence: call recordings. No gap yet.
• Keep the pipeline up to date: required 4, current 3. Evidence: CRM audit.
• Top gap and fix: discovery and next steps, not negotiation. A short course with the team’s own best call as the model, role-play practice with realistic buyers, and managers scoring one call a week against the same checklist. Conversion is checked every month.

Example 5: a customer support team

A software company’s 20-person support team gets a new product release every month. Customer satisfaction on tickets about the latest features is well below the rest. Goal: bring it level with other tickets within two releases.

Evidence: a quality review of 50 tickets against the team’s scorecard, satisfaction comments tagged by theme, and the reopen rate by feature. Our training needs analysis guide walks through a different support example, where only one of two gaps needed a course. Here, too, one gap is not about skill.

• Explain a new feature correctly the first time: required 4, current 2. Evidence: quality review and reopen rate.
• Troubleshoot a fault by reproducing it: required 4, current 3. Evidence: quality review.
• Calm an angry customer in writing: required 4, current 4. Evidence: satisfaction comments. No gap.
• Find the answer in the knowledge base: required 4, current 3. Evidence: the quality review, where agents cite old articles.
• Escalate a bug with the details engineering needs: required 3, current 2. Evidence: engineering sends a third of bug reports back.
• Top gap and fix: new-feature knowledge. A 15-minute course built from each month’s release notes, ready before the release, with a quiz. The old knowledge-base articles are a content problem: give each one an owner and a review date. Satisfaction on new-feature tickets is checked after each release.

Example 6: a clinical team on a hospital ward

A medical ward with 30 nurses is getting new infusion pumps, and its incident reports show medication errors clustered on night shifts. Goal: no pump-related incidents after go-live, and fewer medication errors on nights over the next six months. This example is generic: clinical competencies and sign-off rules come from your own hospital and regulator.

Evidence: incident reports, observed practice audits (a senior nurse watches a medication round and a handover), the device supplier’s competency checklist, and a short skills check on the new pump. Self-ratings are only a starting point. In clinical work, levels come from observation.

• Program and check the new infusion pump: required 4, current 1. Evidence: skills check. Everyone starts here, since nobody has used it yet.
• Give medication with the full set of checks: required 4, current 3 on nights and 4 on days. Evidence: observed rounds and incident reports.
• Hand over a patient using a structured format such as SBAR: required 4, current 3. Evidence: observed handovers.
• Spot a deteriorating patient and escalate early: required 4, current 4. Evidence: chart audit. No gap.
• Top gap and fix: the new pump, a gap everyone shares, and the night medication checks. Hands-on pump practice with sign-off by an assessor before go-live, a short refresher on the checks for night staff, and a look at night staffing and interruptions, which training alone won’t fix. Incident reports are reviewed monthly.

Turning your examples into a plan

Look across the six examples and the same habits repeat. Each starts from a business result, not a course title. Each uses at least two kinds of evidence. Each has a row with no gap, which is training you don’t need to buy. And in most of them, one gap turned out to be a form, a rule or a staffing problem.

To build your own, list the roles and skills in our free training needs analysis tool. Rate the required and current level of each skill on the same 1–5 scale and add how many people it covers. The tool ranks the gaps by priority and gives you a CSV. It runs in your browser, and nothing you type is sent to us.

If you prefer a spreadsheet, the Excel template below has sheets for goals, roles and tasks, the skills gap and the plan, on the same 1–5 scale. Then turn each top gap into an objective you can check, and decide now how you will measure it. Our post on how to measure training effectiveness shows how.

How Classbase is designed to help

Classbase doesn’t write the analysis for you. Deciding what good looks like for a nurse or a sales rep is your team’s call. It is designed to make the steps after it faster.

Attach the documents you already have, such as release notes, a runbook or a pump’s competency checklist, and the AI drafts a course as short, interactive cards. Nothing reaches a learner until an author approves it. For skills a quiz can’t show, add a role-play with an AI character, graded against your own criteria. Assign the course with a due date, and reminders go only to people who are behind. Reports show who finished and which skills each group scores lowest on, so your next analysis starts from evidence.

If you want to see one of these examples built from your own documents, book a demo.