The 7-step model — and a healthy irreverence
The module anchors on the seven-step model from Beth McGoldrick and Deborah Tobey's Needs Assessment Basics (ATD Press, 2016), tagged "from problem to solution to improvement": assess needs, define objectives, design learning, develop materials, deliver training, evaluate training, revise programme. What I loved was the trainer's refusal to be precious about it. She walked the mapping out loud: "Assess ialah analyze lah. Lepas itu design kan. Lepas design, develop kan. Lepas develop, implement — implement dia guna apa? Deliver." Then the punchline, quoting Shakespeare by way of Syekh Saeed: "A rose by any other name smells the same." New frameworks are mostly ADDIE rebranded — "cuma Debra dia pandai, dia buat dalam framework" — Tobey's real contribution being the explicit problem-to-solution-to-improvement arc and the seventh step, revise, which ADDIE leaves implicit. Models are tools, not religions; know the ancestry and you stop being impressed by renaming.
The working steps — outcomes first, people later
The operational flow the workbook details runs: determine the desired business outcomes → link those outcomes with employee behaviour → identify trainable competencies → assess competencies → determine performance gaps → prioritise training needs → determine how to train → conduct a cost-benefit analysis → plan for training evaluation. Notice where it starts: not with people, with business outcomes. The trainer's example was digitalisation again: if employee behaviour cannot support the digitalisation goal, measure — cannot support at all? 50%? 25%? "I tak boleh agak-agak, I kena assess." And if the staff are fine but the infrastructure is the blocker, no amount of training closes that gap — you are back to non-training solutions. Alongside the steps sit the six questions relevant to making a TNA: what training is needed and why, where, who needs it, how it will be provided, how much it will cost — and the sixth, which she called the question nobody ever asks: "Apa impak kalau tak ada training?" What is the impact of NOT training? Answer that and the budget conversation inverts — her example was a company needing 400 certified staff for next year's project with almost none in hand. Suddenly training is not a cost centre; it is the project's critical path. Data collection for all of this splits into quantitative methods (surveys, tests) and qualitative (observation, interviews, documentation review of existing data) — and the strongest analyses combine both.
Prioritising — gap size is not the only factor
The segment that most changed my thinking was prioritisation, taught through a two-person puzzle. The company benchmark is the goalpost. Person A assesses at 18. Person B at 40. Who goes for training first? The reflex says A — biggest gap, most need. But the trainer complicated it immediately: criticality matters. If A's role is critical — "dia sahaja yang dalam department ni," the only one who can unlock next year's project — A goes first however expensive the journey. But there is a second legitimate strategy: send the smallest gaps first, because they convert to fully competent people fastest — "so that kita ada banyak yang competent person dalam organization." Quick wins build the bench while the long projects run. Either way, the decision needs infrastructure: "you must have a record, and you must have a matrix." A competency matrix — who is at what level against which standard — is what turns prioritisation from politics into arithmetic.
Case Study
a competency matrix for the AI rollout
I applied the full flow to the decision I actually face: who across Mana Mana's teams gets AI-tooling training first, with limited seats per cohort. Business outcome first: reduce time spent on repetitive guest-communication and reporting tasks by half within two quarters. Linked behaviour: staff drafting with AI assistance and verifying output, instead of typing everything from scratch. Trainable competencies: prompt drafting, output verification, escalation judgement — all three genuinely K, S and A, so this passes the is-it-even-a-training-problem filter (the tools are already deployed; infrastructure is not the blocker). Then the matrix. Ten staff, scored 0–5 against the standard on the three competencies, from a short practical test — not self-rating, because everyone self-rates a three. The spread: two people at 4 (near-competent), five clustered at 2, three at 1. Prioritisation, using both lenses from class. Criticality: the reservations lead scored 2, but she is the only person handling OTA disputes — her gap gates the whole outcome, so she is in cohort one regardless. Quick wins: the two 4s need half a day, not a course — coach them this week, and they become peer support (the 20% stream from 70:20:10) for everyone else. The remaining seats go to the 2s; the 1s follow next cohort with the 4s as buddies. Cost-benefit on one line: a cohort costs roughly two working days of coverage; the impact of not training — the question nobody asks — is the outcome slipping a quarter, which costs more than every cohort combined. Evaluation planned before delivery: repeat the same practical test at day 30 and compare against today's matrix, which now doubles as the baseline data the rationale slide promised.
Reflections
Day 1 ends with the course eating its own cooking: CT01 taught how people learn; CT02 begins the discipline of deciding whether, what and whom to train — and today's session closed with an assessment exercise, because "tak sihlah kita cakap-cakap tapi kita tak buat." The through-line I am keeping: outcomes before behaviours, behaviours before competencies, competencies before gaps, gaps before priorities — and always, before any of it, the question that prices the whole exercise: what happens if we don't?