Training Needs Analysis (TNA): What, Why, and How — A Complete Operational Guide for HR & L&D

Short answer: Training Needs Analysis (TNA) is the systematic process of identifying the gap between expected and actual performance/competence, confirming that gap can actually be closed by training (rather than a motivation, tools, or process issue), and translating it into measurable learning objectives. The classic McGehee & Thayer framework dissects it at three levels — organizational, task, individual — and in the ADDIE model, TNA is the Analysis phase: the foundation that lets the whole program be designed correctly and its impact be measured.

Most TNA articles stop at the definition, the three levels, and a list of methods (surveys, observation). All true, yet not enough for someone who must actually run a TNA and answer for the result. This guide adds the parts usually missing: the root-cause gate (is this really a training problem?), number-based prioritization (DIF), competency mapping to SKKNI, and the bridge from gap to measurable objective and ROI baseline — with a worked example you can copy.

Intended readers: HR / HC / L&D / SDM teams and unit leaders in private companies, state-owned enterprises (BUMN/BUMD), government agencies, institutions, associations, and non-profits.

Quick navigation

1. What TNA is (definition, 3 levels, place in ADDIE)
2. Why TNA matters (the cost of skipping it)
3. The root-cause gate: is this really a training problem?
4. How to run a TNA: seven steps
5. Data-collection method matrix
6. Prioritizing with DIF Analysis (worked numbers)
7. Competency mapping to SKKNI
8. From gap to measurable objective & ROI baseline
9. TNA report contents & role split (RACI)
10. Worked TNA example (illustrative)
11. Common mistakes & how to avoid them
12. Glossary
13. FAQ
14. Next step

What TNA is (definition, 3 levels, place in ADDIE)

Training Needs Analysis is the systematic process to: (1) find the gap between expected and actual performance/competence, (2) test whether that gap is genuinely caused by a knowledge/skill deficit, and (3) translate the valid gap into measurable learning objectives. The output becomes the basis for program design — and the zero point for measuring impact.

Allison Rossett (1987) treats needs assessment as an umbrella that must uncover five things: optimal performance, actual performance, stakeholder feelings, the causes of the gap, and solutions. "Assessment" and "analysis" are used interchangeably in the field; what matters is whether the process answers all five, especially the causes.

The classic three levels (McGehee & Thayer, 1961)

The most influential framework comes from Training in Business and Industry (McGehee & Thayer, 1961), which dissects need at three levels:

• Organizational analysis — business goals, strategy, climate, and resources. The question: does training align with the organization's direction, and will the work environment support its application?
• Task/operational analysis — the critical tasks and competencies the role demands, compared against a standard (e.g. SKKNI or an internal competency dictionary). The question: which competencies, at what level, does this role genuinely need?
• Person analysis — who needs training, their starting level, and application barriers on the job. The question: for these specific people, which gaps are real?

Ostroff & Ford (1989) extended this framework by adding a unit/sub-unit dimension and a data-interpretation stage — a reminder that TNA data must be read in context, not swallowed raw.

TNA's place in ADDIE

In the ADDIE instructional-design model (Analysis, Design, Development, Implementation, Evaluation), TNA is the core of the Analysis phase. It is not an opening formality; it decides whether the next four phases produce impact or just activity. A mistake in Analysis means the whole program points the wrong way, however good the facilitator.

Working definition: TNA is not "asking employees what training they want". TNA is proving the gap, filtering whether it deserves to be closed by training, and turning it into a measurable target.

Why TNA matters (the cost of skipping it)

Skipping TNA does not save time; it moves the cost somewhere invisible. Four concrete consequences:

• Activity spending, not capability. Training runs, participants give high satisfaction scores, then nothing changes on the job — because the root cause was not a skill deficit.
• "Sheep-dip" training. Everyone is trained on the same thing with no specific reason. Big budget, low relevance, thin impact.
• Unmeasurable impact. Without a TNA baseline, evaluation stops at Kirkpatrick Level 1 (satisfaction). Level 3 (behavior), Level 4 (business results), and Phillips ROI (Level 5) are impossible to calculate with no zero point.
• Indefensible decisions. When management, audit, or procurement asks "what is this program based on?", "the team requested it" is weak; "competency gap X against standard Y, highest DIF priority" is strong.

TNA is the difference between a training budget that becomes an investment and one that becomes a cost. This is also the core of how to choose a corporate training vendor: a healthy vendor declines to quote before a TNA.

The root-cause gate: is this really a training problem?

This is the step TNA articles almost always skip, yet it is the most decisive. Before any gap is recommended for training, pass it through this gate.

The Mager & Pipe (1984/1997) test: ask the screening question — "If their life depended on it, could this employee do it?" If the answer is yes, but they don't do it, this is not a skill deficit. Training will not solve it; the cause lies in motivation, consequences, tools, process, or expectations.

Thomas Gilbert's Behavior Engineering Model (1978): most performance barriers are environmental, not individual. Check these six cells before concluding "needs training":

Domain	Cell	Diagnostic question
Environment	Information & feedback	Are expectations clear and feedback timely?
Environment	Resources & tools	Are tools, systems, and processes adequate?
Environment	Incentives & consequences	Is good performance rewarded and poor performance consequential?
Individual	Knowledge & skills	Is there genuinely a knowledge/skill deficit? (← training's domain)
Individual	Capacity	Is the right person in the right role?
Individual	Motives	Are personal motivation and expectations aligned?

Only the knowledge & skills cell is truly solved by training. If TNA finds the root cause in another cell, the recommendation is not training: fix the process, tools, feedback, or incentives. Honesty at this gate saves the largest budget.

Core: Training for a non-skill problem is cost with no result. The root-cause gate turns TNA from "designing training" into "deciding whether training is the answer".

How to run a TNA: seven steps

1. Set the business goal & scope — start from the result the unit/organization wants (organizational level). Without a business goal, TNA loses its benchmark.
2. Define expected performance/competence — the role standard: critical tasks, the competency levels demanded. Make it explicit and measurable (use SKKNI or a competency dictionary).
3. Measure actual performance/competence — gather evidence with several methods (see the matrix below). Triangulate.
4. Compute the gap — gap = expected − actual, per task/competency, per participant segment.
5. Pass the root-cause gate — for each gap, run the Mager & Pipe + Gilbert test. Filter out what is not a training problem; record non-training recommendations separately.
6. Prioritize — valid training gaps are ordered with DIF Analysis (below), aligned to business goals and budget.
7. Translate into measurable objectives + baseline — each priority becomes a specific learning objective, with a starting condition as a baseline and behavior/result-level success indicators.

The output of steps 1–7 is a TNA report that becomes the basis for program design (ADDIE Design phase) and, at once, the evaluation framework.

Data-collection method matrix

No single method is enough; each carries bias. Pick a combination to fit the question, then triangulate at least two sources.

Method	Strongest for	Strengths	Limitations
Survey/questionnaire	Broad reach, perceived needs	Fast, cheap, large scale	Self-report bias; shallow
Interview	Depth, context, causes	Nuance, flexible	Time-costly; interviewer bias
Focus group (FGD)	Consensus, team dynamics	Multiple viewpoints	Dominant-voice skew
Direct observation	Task reality on the ground	Objective on real behavior	Time-costly; Hawthorne effect
Work samples & KPI/HRIS data	Factual performance gap	Objective, already exists	Doesn't explain the cause
Assessment/competency test	Actual competency level	Measured against a standard	Needs a valid instrument
360 feedback	Behavioral gap	Many perspectives	Needs cultural maturity

Rule of thumb: use objective data (KPI/assessment) to prove the gap, and qualitative methods (interview/FGD) to explain its cause — both are needed to pass the root-cause gate.

Prioritizing with DIF Analysis (worked numbers)

Once the list of valid training gaps is gathered, it is usually too long to budget at once. DIF Analysis turns the wish list into a decision order. Rate each task/competency on three dimensions:

• Difficulty (D) — how hard it is to master (1–4 scale).
• Importance (I) — how critical to role/business success (1–4 scale).
• Frequency (F) — how often it is performed (1–4 scale; or relative frequency).

Compute a priority score (a common approach: D × I × F) then categorize:

• Over-train — important & difficult yet infrequent: train formally plus reinforce with job aids/checklists and periodic practice (easily forgotten).
• Train — high on two dimensions: a prime formal-training candidate.
• No-train — low score: a quick guide, demo, or peer coaching — not a class.

Illustrative example (numbers to demonstrate the method, not client data):

Task/competency	D	I	F	Score (D×I×F)	Category	Decision
Vendor contract negotiation	4	4	2	32	Over-train	Formal training + checklist + periodic simulation
Sales-data analysis	3	4	4	48	Train	High-priority formal training
Filling an internal claim form	1	2	4	8	No-train	Job aid + short demo

DIF makes the decision transparent: what gets budgeted first is the highest score aligned to the business goal, and every "not now" has a written reason.

Competency mapping to SKKNI

At the task/role level, "expected performance" should be tied to a standard so it is not subjective. In Indonesia, SKKNI (the Indonesian National Work Competency Standard) set by Kemnaker provides a set of competency units usable as a benchmark.

The mapping flow:

1. Identify the role and its critical tasks (output of task analysis).
2. Map each task to the relevant SKKNI competency unit (or an internal competency dictionary where a sector SKKNI is not yet available).
3. Set the required level per unit, then measure the actual level (assessment/test).
4. The gap per competency unit becomes the basis for module design — and, if the organization needs formal recognition, the basis for a certification scheme via a BNSP-licensed Professional Certification Body (LSP).

Tying TNA to SKKNI makes the gap auditable ("against national standard unit X") and connects training to a certification path where relevant. For deeper LPK/BNSP/LSP legality, see the vendor-selection guide.

From gap to measurable objective & ROI baseline

A TNA that stops at "there is a gap in X" is not finished. The closing step turns findings into something measurable:

• Specific learning objectives — stated as an observable competency/behavior change ("able to draft a negotiation proposal that includes a BATNA and a zone of agreement"), not a topic title ("negotiation training").
• Baseline — record the starting condition (assessment score, KPI, behavior) before training. This is the zero point.
• Tiered success indicators — map to Kirkpatrick: Level 2 (competency gain), Level 3 (on-the-job application), Level 4 (business indicators). For large programs this baseline enables a Phillips Level 5 ROI calculation = (net benefits ÷ program cost) × 100.
• 70-20-10 check — if TNA shows the need is better met through experience/assignment or social interaction than a class, design it that way; do not force a classroom format just because that is what is usually bought.

Without a TNA baseline, impact can only be guessed. With a baseline, impact can be calculated — and that is what separates L&D the budget trusts from L&D that gets cut first when efficiency hits.

TNA report contents & role split (RACI)

A good TNA report contains, at minimum:

• Business goal & scope; methods and data sources (with their limitations).
• Expected vs actual performance/competence; a gap table per segment.
• Root-cause gate results: training gaps vs non-training recommendations (process/tools/incentives).
• DIF priorities with scores and reasons.
• Measurable learning objectives, baseline, and evaluation indicators.
• Format recommendation (class/blended/on-the-job) and a scope estimate, not invented prices.

A healthy role split:

Activity	HR/L&D	Unit leader	Line manager	Employee	Vendor
Business goal & scope	C	A/R	C	I	I
Define standard competencies	R	A	C	I	C
Collect actual data	A/R	I	C	R	C
Root-cause gate	A/R	C	C	I	C
Priorities & measurable objectives	A/R	C	I	I	C
Validation before proposal	C	A	I	I	R

(R = Responsible, A = Accountable, C = Consulted, I = Informed.) A healthy training vendor is at minimum Consulted and helps validate the TNA before drafting a proposal.

Worked TNA example (illustrative)

Demonstration scenario (not real client data; numbers illustrate the method):

A sales unit targets a higher conversion rate. Organizational level: goal = raise conversion; context = a new product, a complex sales cycle. Task level: critical competencies = customer-needs diagnosis, objection handling, value-proposition drafting. Individual level: assessment shows the team is strong at presentation, weak at needs diagnosis.

Root-cause gate: interviews reveal two things — some of the team do not know how to run a diagnostic discussion (skill deficit → training's domain), some know but do not do it because incentives only chase demo volume (environmental root → recommendation: fix the incentive scheme, not training).

DIF: "customer-needs diagnosis" scores D3·I4·F4 = 48 → Train priority. Measurable objective: "able to run a diagnostic conversation that surfaces three of the customer's hidden needs". Baseline: initial diagnostic role-play score + current conversion rate. Evaluation: Level 3 (frequency of real diagnostic conversations), Level 4 (conversion movement), noting that the incentive fix runs in parallel so attribution stays fair.

Lesson: without the root-cause gate, the whole team would be retrained — half the budget wasted on people whose problem was incentives, not skill.

Common mistakes & how to avoid them

Key takeaways:

• Asking "what training do you want" → start from the business goal and a proven gap, not a wish list.
• Skipping the root-cause gate → run Mager & Pipe + Gilbert before recommending a class.
• A single method → triangulate objective data (KPI/assessment) with qualitative (interview).
• Treating all needs as equal → order with DIF; every "not now" has a written reason.
• Stopping at "there is a gap" → turn it into a measurable objective + baseline so impact can be calculated.
• Subjective standards → tie to SKKNI / a competency dictionary so it is auditable.

Glossary

• TNA (Training Needs Analysis) — the systematic process of identifying, validating, and translating competency gaps into measurable training objectives.
• Performance gap — the difference between expected and actual performance/competence.
• Behavior Engineering Model (BEM) — Thomas Gilbert's (1978) framework mapping performance causes to six environmental & individual cells.
• DIF Analysis — a prioritization method based on Difficulty, Importance, Frequency.
• SKKNI — the Indonesian National Work Competency Standard; a competency-unit benchmark set by Kemnaker.
• ADDIE — an instructional-design model: Analysis, Design, Development, Implementation, Evaluation.
• Baseline — the measured starting condition before an intervention; the zero point for impact evaluation.

FAQ

What is Training Needs Analysis (TNA)?

TNA is a systematic process to identify the gap between expected and actual performance/competence, confirm the gap can genuinely be closed by training, and translate it into measurable learning objectives. The classic McGehee & Thayer (1961) framework dissects need at three levels: organizational, task/role, and individual. In the ADDIE model, TNA is the Analysis phase — the foundation before design, development, implementation, and evaluation.

What is the difference between TNA, training needs assessment, and needs analysis?

In practice the terms are used interchangeably. Per Allison Rossett (1987), assessment is the umbrella activity that uncovers five things: optimal performance, actual performance, stakeholder feelings, the causes of the gap, and solutions; analysis is the activity of examining the data to decide whether and what training is needed. What matters is not the label but whether the process answers all five — especially the causes, because not every gap is a training problem.

Why is TNA important and what is the risk of skipping it?

Without TNA, the training budget becomes activity spending: participants are satisfied, but performance does not change because the real root cause was not a skill deficit. TNA prevents 'sheep-dip' training (everyone trained on the same thing for no specific reason), sets the baseline that makes impact measurable (Kirkpatrick Level 3–4, Phillips ROI Level 5), and makes budget decisions defensible to management, audit, and procurement.

What are the three TNA levels per McGehee & Thayer?

Organizational level: business goals, climate, and resources — whether training aligns with strategy and will be supported by the work environment. Task/role level: the critical tasks and competencies the role demands, compared against a standard (e.g. SKKNI or an internal competency dictionary). Individual level: who needs training, their starting level, and application barriers. Ostroff & Ford later extended this with a unit dimension and a data-interpretation stage.

How do I confirm a gap is really a training problem?

Pass it through the 'root-cause gate' before recommending training. The Mager & Pipe question: 'If their life depended on it, could the employee do it?' If they could but do not, this is not a skill deficit — the cause is motivation, consequences, tools, process, or expectations (Thomas Gilbert's Behavior Engineering Model: most performance barriers are environmental, not knowledge). Training for a non-skill problem wastes budget with no result.

What data-collection methods are used in TNA?

Combine several: surveys/questionnaires (broad reach, fast, cheap, but prone to self-report bias), interviews (depth, context), focus groups/FGD (consensus, dynamics), direct observation (task reality, time-costly), work samples & performance/KPI/HRIS data (objective), assessments/competency tests (measured actual level), and 360 feedback (behavior). Triangulate at least two sources so findings are not biased by a single method.

How do I prioritize many training needs?

Use DIF Analysis: rate each task/competency on Difficulty, Importance, and Frequency. High scores on Importance and Difficulty with a given frequency = a formal-training priority ('Over-train' when critical yet infrequent, needing reinforcement/job aids). Low-scoring needs are handled with a quick guide or coaching, not a class. DIF turns a wish list into an evidence-based decision order.

How does TNA relate to SKKNI and competency mapping?

At the task/role level, the competency gap is measured against a standard. In Indonesia, SKKNI (the Indonesian National Work Competency Standard, set by Kemnaker) provides competency units usable as a benchmark: map the competencies the role demands to relevant SKKNI units, measure the actual level, and the gap becomes the basis for training design and, if needed, a certification scheme through a BNSP-licensed LSP.

How are TNA findings turned into something measurable?

Translate each gap into a specific, measurable learning objective (an observable competency/behavior change, not a topic title), set the starting condition as a baseline, and define success indicators at the behavior and business-result level. That baseline is what makes Kirkpatrick Level 3–4 evaluation and a Phillips Level 5 ROI calculation possible — without a baseline, impact can only be guessed, not calculated.

How long does TNA take and who runs it?

A focused TNA for a single program can finish in 1–2 weeks; a full organizational or multi-unit TNA takes several weeks. The process owner is usually HR/L&D, with the unit leader as sponsor (owner of the business goal), line managers contributing task context, and employees providing actual data. A healthy training vendor helps run or validate the TNA before drafting a proposal.

Next step

You now have the full operational flow: set the goal, prove the gap, filter through the root-cause gate, prioritize with DIF, tie it to SKKNI, then turn it into a measurable objective with a baseline. The sensible next step is to run a focused TNA on your single most urgent need — before budgeting any program.

Neksus starts every program from a TNA: diagnosing the gap, filtering whether training is the answer, and designing measurable objectives (Kirkpatrick/Phillips) — so the budget targets real change. Discuss your team's need and request an initial TNA session via the Neksus contact page — no obligation, as the right starting point.

Also read:

• How to Choose a Corporate Training Vendor / Provider
• Organizational Change Management
• Leadership for First-Line Managers
• Digital Transformation for Executives
• See the full training catalog →

---

Last updated: 17 May 2026. The frameworks cited (McGehee & Thayer 1961; Rossett 1987; Gilbert 1978; Mager & Pipe; Ostroff & Ford 1989; ADDIE; Kirkpatrick; Phillips; DIF; SKKNI) are attributed to their original sources. The worked example is illustrative to demonstrate the method; Neksus does not display client names or success statistics.