Session 1: How People Learn

The most common assumption about teaching is wrong: that knowing a subject deeply makes you able to teach it. In practice the opposite often happens, because expertise makes things invisible. Once you can do something fluently, you no longer remember the steps you took to get there, or which parts were hard. An expert asks 'what do you not understand?' and is genuinely puzzled by the answer, because the thing that is obvious to them now was once not obvious at all.

What actually makes a good teacher is different. Remembering what it was like not to know. Being able to see the specific place a learner is stuck rather than assuming it is the obvious place. Explaining the same thing three different ways until one lands. And noticing when someone has gone quiet, because silence online usually means confusion rather than agreement.

Then the part nobody mentions: teaching is mostly listening and adjusting. A lesson plan is a hypothesis about what will help, and the class is the test. The teacher who follows the plan exactly regardless of the confused faces is not being disciplined — they are not teaching. The plan exists so you know where you are going; the students determine how fast you get there and which route you take.

You cannot teach someone you have not located. Before planning anything, establish three things: what they already know, what they need, and what they will do with it. A class of beginners, a class of people who half-know the subject, and a class of professionals refreshing a skill need completely different lessons, and teaching the wrong one wastes everyone's time.

The curse of the middle is worth naming. Mixed-ability classes are the hardest to teach, because the beginners are lost and the advanced are bored, and the temptation is to aim at the middle and satisfy nobody. The practical answer is to teach to the beginners in the explanation and to the advanced in the exercises — explain assuming no prior knowledge, then set tasks with an obvious extension for those who finish early. Nobody is insulted by a clear explanation; people are insulted by being held back.

Then find out what they are actually struggling with, because it is rarely what you expect. Ask them directly, in their words: 'what have you tried, and where exactly did it stop working?' Their answer tells you the real obstacle, which is often a missing earlier step rather than the step you were about to teach. Teaching the wrong step well is the most common failure in instruction, and it happens because nobody asked.

A learning objective is a statement of what the learner will be able to do at the end that they could not do at the start. Not what you will cover — what they will be able to do. 'We will cover Excel formulas' describes the teacher. 'You will write a formula that totals a column and handles blank cells' describes the learner, and it can be tested.

The test of a good objective is the verb. Observable verbs — write, build, calculate, identify, correct, produce — describe something you can see someone do. Unobservable verbs — understand, appreciate, know about, be familiar with — describe a state of mind you cannot check, which means you will never know whether the lesson worked. Write objectives with observable verbs and the assessment designs itself.

Then the discipline of few objectives. Three is the maximum for a ninety-minute session, and two is usually better. A lesson with eight objectives teaches none of them properly, because each gets a few minutes and none gets practised. The temptation to include everything you know is strong, and it is the main reason lessons run over and learners leave with nothing solid.

A lesson plan has six parts, and each answers a question. Objective — what will they be able to do? Hook — why should they care in the first five minutes? Demonstration — how is it done, shown rather than described? Guided practice — what do they do with you watching? Independent practice — what do they do alone? Check — how will you know it landed?

The order matters more than people expect. Demonstrate before you explain where you can, because seeing the finished thing gives the learner a destination, and then the explanation makes sense as the route to it. Explaining first and demonstrating later means the learner holds abstract information with nowhere to put it.

Then timing, written down and honest. Ninety minutes is not ninety minutes of content — it is roughly fifteen minutes of explanation, twenty of demonstration, thirty of practice, fifteen of questions, and the rest lost to technical problems, late arrivals and repetition. Plan for less content than feels right, because a lesson that finishes with the objective achieved beats one that runs over with the objective half-covered. And always mark what you will cut if you run out of time, decided in advance rather than in the moment.

Every complex topic is a sequence of simple ones, and finding that sequence is the core craft of teaching. The method is to work backwards from the finished thing. Take the task you want them to do, and ask 'what did I have to be able to do just before that?' Then ask it again about the answer. Five or six rounds of that question produces a step sequence, and it usually reveals two or three prerequisite steps you had forgotten you knew.

Then order by dependency, not by how the software is organised. Most subjects are taught in the order their menus are laid out, which is an accident of design rather than a logic of learning. A learner needs the concept before the tool that uses it, and needs the tool before the settings that modify it. Teaching in menu order produces people who can follow instructions and cannot think.

Finally, each step must be small enough to fail safely. If a step has seven things in it, the learner who fails will not know which of the seven went wrong, and you will spend the lesson diagnosing rather than teaching. One idea per step, one action to check it, then the next. It feels slow and it is much faster, because nobody gets lost and you do not have to go back.