Three, nine, three
Rather than lecture the template, Martin built a session plan live, taking group one's morning topic — backpacks, under the programme title "Lighten the Load" — and filling in the form on screen while we watched the decisions happen. Overview first: session title "Types of Backpacks", trainer — your own name, "on Friday, you are the trainer-facilitator" — duration fifteen minutes, twelve participants. Work, health and safety: office health and safety as the sensible default, with a nod to the MCO era when every plan carried social-distancing guidelines. Then the arithmetic that structures everything downstream: fifteen minutes splits into roughly three minutes of opening, nine of body, three of closing. Three plus nine plus three. The opening rows are simply GLOSS written vertically; the closing rows are OFF; and the nine minutes in the middle are where the actual design decisions live. Before any content existed, the time was already budgeted — which is exactly backwards from how most people plan, and exactly why most people run over.
The scope decision
The learning outcome came next, and with it the afternoon's sharpest teaching. He wrote: identify the three types of backpacks for various outdoor activities. Three. "I talk not about five, I do not talk about ten. I talk only about three. You don't put in your slides ten bags — you just put three bags." And then the warning, delivered flat: "If I see already you write ten or fifteen, I know your session will be not good. It is impossible in fifteen minutes. You stand here and you speak like a machine gun." The general principle: every session, of any length, is the same decision about how much content fits the container. His example from manufacturing — the seven wastes of lean. In fifteen minutes you can either survey all seven shallowly, or pick one waste and go deep. Both are legitimate sessions. What is not legitimate is all seven, deeply, in fifteen minutes: "that one will not work." Broad or deep — decide, then write the outcome to match the decision, because the outcome is a promise about scope. Write three, teach three, assess three. It is the rock-pebbles-sand jar from Day 1, but inverted: the classic version says fit the big rocks in first. The session-planning version says: first admit how big the jar actually is.
A memo for your head
Watching him fill the GLOSS rows settled the script question for good. Under grab attention he wrote: trainer asks probing question — example: how many of you have been on a hike recently? — trainees respond by show of hands. Not "Good morning everyone, welcome to my session" — "you don't write your opening welcome. We are writing some memo for our head: what actually I want to do, and what are the key reactions I am expecting." The plan records intentions and expected responses, and reading it back, the actual sentences reconstruct themselves: that is what people mean when they say it's like a script, and why it isn't one. The link row: ask the hands that went up what backpack they used — prior knowledge, surfaced from the people who have it, and a quiet tactic inside it: don't cold-call the people who didn't raise their hands; they have never hiked and have nothing to share yet. Outcome row: copy the learning outcome across — literally copy it. Structure: today, lecture and discussion. Stimulate: show a real backpack, or a picture of a huge mountain — "okay, now we are ready, we want to climb Mount Kinabalu, let's go!" Then the timing column, with the granularity rule attached: half-minute blocks. Half a minute for grab, half for link, half for outcome — the three minutes assemble themselves. Never plan in ten-second slices: "The moment someone asks you a question, your ten seconds are already gone." Ten-second plans are written by engineers, he added, looking at no one in particular, and broken by humans every single time.
The arithmetic of nine minutes
The body rows made the template earn its keep. Two minutes: lecture on the three backpack types — expedition pack, eighty litres, waist harness; day pack, twenty litres; sling bag, single strap. Not phrases, pointers — "for everything you want to talk about, you have some pointers there: what is important to talk." The test of sufficiency is portability: "I can take this session plan and give it to someone else, and they can run the same session without being me there." If the pointers don't carry the session, they are not the right pointers. Next row, the activity: a discussion, with its instructions written out — analyze the five given scenarios, discuss with your team which backpack is most suitable, write down findings — because instructions belong in the plan and on the slide, both. Two minutes to discuss, ninety seconds for groups to share findings. Then he ran the total like an accountant: two plus two plus one-and-a-half — five and a half minutes against a nine-minute body. "So now you need to think: do I extend my lecture? Do I extend the discussion — that means more complex scenarios maybe? Or do I do something additional?" The something-additional became a third block: three minutes on wearing an expedition pack properly, one volunteer, step by step. Nine minutes, accounted for. And the workflow rule that closes the loop: write the session plan first to clear your mind, then build the slides, then come back and update the plan's resources column — slide two, slide four, handouts — so plan and deck match. The plan drafts the slides; the slides correct the plan.
The quiet one who delivered
Then he showed us real plans from past cohorts — photographs only for the participant ones; "you can take pictures, but you cannot have the copy." The one he lingered on came with a story he clearly re-tells every cohort. A young financial accountant — very silent for the whole training, the kind of participant you would bet against. Her colleague was the opposite: outspoken, confident, and already the company's long-term trainer. On Friday she delivered one of the best sessions he had assessed, and the colleague — who had followed none of the recommendations — "messed up," in Martin's words, and the company was shocked when the results went back. "This is assessment." The quiet ones prepare. Her plan: "The Story Behind the Numbers — How Every Action Shapes Profit." Fifteen minutes, nine pax. One outcome: identify the five main components of a Profit and Loss statement — "very clear: not ten, not a hundred components. Five." Methods: lecture, an interactive game, a case scenario; assessment: quiz. "So I can see already — identify, lecture, training game, quiz. Alignment is done." Her opening rows were written fuller than the template needs — the engaging question spelled out word for word: do you think your daily work, approving forms, checking quality, managing staff, has anything to do with company profit? — and Martin's verdict on the excess was gentle: she was a first-timer and nervous, and "if you are nervous, at least you have some anchor point." Bullet points would do; her sentences were scaffolding, not sin. Her stimulation line he read out with real admiration: by the end, you'll see that profit isn't just about numbers — it's about you and what you do every day. "It's not about my theory. It's about you. This can be very stimulating, because it's very personal."
Jump for Profit, PLC 101, and the nine-page warning
The middle of her session was a game called Jump for Profit. A line of tape on the floor; cost items appear on screen — food cost, packaging, labour, utilities, marketing, repair and maintenance; participants physically step left if it belongs to Cost of Sales, right for Operating Expenses, with three to five seconds to think. Her plan specified all of it: the tape, the seven items, the thinking time, the debrief. For the close she ran a scenario — government introduces a new foreign-worker law, HQ launches a promotion, the outlet still has to maintain the air-con — and a quiz through a QR code, then a Q&A she had prepared questions for. Six slides in total. "It's not about how many slides — but you can see it's very detailed." His one improvement: the reflection step should have carried three prepared questions to steer the debrief toward the outcome. A second good plan — "PLC 101", an automation session with a catchy title, the physical PLC hardware in the room to touch, a role play with its briefing written out, a Google Form quiz with the link pasted into the plan, and a future row promising that next session, we learn to program like a pro. A bit long in places, "but more is better than lesser." Then the counter-example, anonymised and devastating: a nine-page plan whose rows read "introduction, ice breaking, three minutes", "overview of title and learning outcome", "EPF roles, functions and benefits — slides one to four", "verification review — seventeen slides." Seventeen slides, no trainee action in any row, no interaction anywhere. "When I get your session plan, I don't want to listen to you already." This person will talk for fifteen minutes, and the plan says so before they open their mouth. The scoring bands are correspondingly blunt: plans like the first two put you in the four-to-five range; the nine-pager is one-to-two. "The session plan is very easy to score."
Send it early
The logistics advice at the end was really a story about the same accountant. She sent her draft plan on Wednesday evening — the same evening she learned the material. That bought her a written reply with corrections, a Thursday review, and five to ten minutes of direct guidance; by Friday she was polished. The alternative path is the cohort tradition Martin has given up fighting: "I always say this, and it never fully materializes — at 2 AM on Friday morning, we are sleeping. We are not reading your session plan." And still, every cohort, the 2 AM submissions arrive from people who then look a particular kind of tired on Friday afternoon. The afternoon closed with a half-hour team exercise — a session plan for our own morning topics, template in the WhatsApp group — and the reminder of what is actually collected: the final session plan goes to him for the report; slides are ours to keep, though he will give feedback on them if we send early. Feedback, but not corrections: "I will not correct your slides." The boundary between coaching and doing-it-for-you, drawn exactly where a good trainer draws it.
Case Study
planning my own fifteen minutes
Tonight I did the thing Martin says nobody does: I started my Friday session plan on Wednesday evening. The session comes from our group's AI course — which even acquired a proper marketing poster tonight, "AI Odyssey", because apparently I procrastinate on scoping by designing posters — and scoping was exactly where the machine-gun warning hit me. My instinctive topic was "how to prompt AI effectively", which is the ten-backpacks version: models, techniques, examples, caveats. Fifteen minutes. Machine gun. The seven-wastes decision cut it down honestly: one waste, deep. Outcome: identify the three components of an effective AI prompt — context, instruction, output format. Not five components, not a framework tour. Three. The plan wrote itself onto the three-nine-three skeleton with today's examples as donors. Opening, three minutes of GLOSS in half-minute blocks — grab is a live before/after on screen: the same request to AI, lazy prompt versus structured prompt, visibly different results; link is right-hand-left-hand-joke on who has used ChatGPT this week; outcome copied onto its slide; structure — lecture and one activity; stimulate — this is the difference between AI writing for you and AI guessing for you. Body, nine minutes: two-minute lecture, three pointers, one per component. Then Jump for Profit, stolen with pride as "Jump for Prompt": tape line on the floor, eight prompt fragments on screen, step left if it's context, right if it's instruction — three to five seconds of thinking time, exactly as specified by a financial accountant I will never meet. Then pairs rewrite one genuinely bad prompt against a scenario card, ninety seconds of sharing. Closing: QR-code quiz — the link already pasted into the plan — Q&A with two questions pre-seeded to two classmates, and a future row: in the next session, we would put these prompts to work on your actual inbox. The plan is one page. The trainer column has a verb in every row; the trainee column has no empty cells — the nine-page EPF ghost checked both columns over my shoulder. It goes to Martin tomorrow, not at 2 AM Friday, because the lesson of the quiet accountant is not really about sleep: the draft you send early gets a second author. And her other lesson is already in the plan — the stimulation line is personal, not technical. Not "prompting is an important skill." You already know how to brief a colleague; tonight you learn to brief the machine.
Reflections
Watching a session plan get built live was the most useful ninety minutes of the course so far, because every cell of the template turned out to be a decision, and every decision had a failure mode he had personally watched someone fail: scope too wide, machine gun; timing too fine, broken by the first question; rows without trainee actions, fifteen minutes of monologue foretold; a plan faked for the assessor, worse than no plan. The template is a form; the teaching was the jurisprudence. And the person the afternoon kept returning to was not a framework but the quiet accountant. Silent all course, precise on paper, devastating on delivery — while the confident in-house trainer improvised and sank. Every framework this week has been pointing at the same unglamorous conclusion: delivery is the visible tip of preparation. On Friday, fifteen minutes will be watched, and the four days underneath them — the outcome verb, the alignment row, the half-minute blocks, the seeded questions, the strip of tape on the floor — will be doing the actual performing. The session plan is where a session goes to become inevitable.