Where we are
You have a sequence with worked examples in it. Tonight is what the learners actually do, which is the part that determines whether anything survives the week.
★ What you walk out with tonight. One activity that is genuinely your objective rather than adjacent to it, a retrieval moment placed on purpose, a spacing plan even if you only have one session, and a sentence for when somebody asks you to make it flow better.
LiveThe illusion, stated plainly2 min▶
People judge how much they have learned by how easy it felt. That judgement is close to backwards. A fluent explanation feels like understanding, and re-reading feels like consolidation, and neither produces much that survives a fortnight.
The consequences for you are specific and awkward:
- Your best-designed sessions may get slightly worse feedback scores than your worst ones.
- Sponsors will ask you to remove the parts that work, because those are the parts people found hard.
- Track B session 5 exists partly to give you a number to answer that with.
None of this means make it unpleasant. Difficulty for its own sake is just bad design. It means the effort should be on the thing they came to learn, and the smoothness should be everywhere else.
Being made to recall it beats being shown it again
6 min liveLiveWhere to put retrieval, concretely4 min▶
- Before you teach it. A guess costs thirty seconds and makes the answer land. "Which of these six do you think break a consumer?" They will be wrong and that is what makes the correction stick.
- Three-quarters of the way in. Closed book, two minutes. This is the highest-value two minutes in most sessions, and it is also your only honest read on how it is going while you can still change something.
- At the very end, as the exit. Not a feedback form. One question they answer from memory, handed to you on the way out - it doubles as your data for track B session 5.
- A week later, by email. Two questions. Costs you five minutes to write and does more for retention than extending the session by an hour.
Say why you are doing it. "This will feel harder than watching me do it, and that is why it works" takes eight seconds and converts most of the resistance, because people will accept difficulty they understand the purpose of.
Same minutes, spread out, and the activity that is the objective
6 min liveLiveIs the activity the objective, or next to it?4 min▶
Take your objective's verb and check the activity uses the same one. It sounds too obvious to be a real failure mode and it is the most common one there is.
| Objective verb | Activity that is adjacent | Activity that is it |
|---|---|---|
| Classify a schema change | Discuss why contracts matter | Classify ten schema changes |
| Draft a deprecation notice | Review a good example notice | Draft one, swap, review a peer's |
| Diagnose a failing load | Watch a demo of the diagnosis | Diagnose one, with the logs |
| Choose between two designs | Learn the criteria | Choose, and defend it to somebody |
Everything in the middle column is useful and none of it is the objective. The middle column is what fills sessions when nobody checked, because it is easier to build and much more comfortable to run.
Read your objective's verb. Read your activity. Are they the same verb? If your objective says classify and your activity says discuss, you have built a different session from the one you promised.
Design one activity, properly
★ 22 min · everyone buildsSame verb. Your objective's verb and your activity's verb are identical, or you are building something else. Four-second check, most commonly failed.
Real material. Ten real schema changes from your own repository, not invented ones. Invented material teaches a method that works on invented material.
Retrieval. Closed book somewhere. Two minutes, three-quarters through, and it is also your read on how it is going.
Space it. Either three sittings, or one session plus two emails. Write the emails now while you still care - nobody writes them later.
Say why. Eight seconds explaining that the hard part is the point. It converts most of the resistance and it costs nothing.
Write your activity in one line, starting with the objective's verb. If you cannot start with that verb, you have found the problem.
Go and get real material. Ten actual examples from your own systems, tickets, documents. This is the most time-consuming part of course design and it is what separates a session that transfers from one that does not.
Include two ambiguous cases on purpose. Cases where competent people disagree. The argument is where the real understanding shows up, and it is the part people remember.
Place your closed-book two minutes. Three-quarters through. Write the actual question now.
Write the two follow-up emails now. Day 7 and day 30, two questions each, scheduled. Five minutes today; nobody ever writes them in three weeks.
Write your eight-second sentence. "This next bit will feel harder than watching me, and that is why it is in here."
Now audit an existing session against the middle column. How much of it is discuss, review, watch, learn-about? That fraction is how much of it is adjacent to its own objective.
If you take one thing from tonight. Check that your activity's verb is your objective's verb. It takes four seconds, it is the most commonly failed check in this discipline, and failing it means the session teaches something other than what it promised.
Before session a5
30 min plus material gathering- Collect ten real cases. From your own systems. This is the homework that actually takes time and it is the one that matters.
- Find your two ambiguous ones. Where competent colleagues would disagree. Check by asking one.
- Verb-check three existing activities against their objectives. Count how many are adjacent.
- Write and schedule the two follow-up emails. Now, not later. Later does not happen.
- Bring your ten cases to a5, because a5 turns them into assessment questions with distractors that teach.
Three questions
Clear all three and this session gets stamped in your passport.
1 · Your best-designed session gets slightly worse feedback than a smoother one. What is the most likely explanation?
Fluency feels like understanding. A smooth session where you explained everything beautifully produces confident people who cannot do the thing. This is why track B session 5 exists - you need a number that is not the feedback form.
2 · Your objective says "classify a proposed schema change". Your main activity is a group discussion about why data contracts matter. What is wrong?
Discussion is useful and it is not classification. The four-second check - is the activity's verb the objective's verb - is the most commonly failed one in this discipline, because adjacent activities are easier to build and much more comfortable to run.
3 · Spacing the same 90 minutes over three sittings makes day-one performance slightly worse. Why do it?
The day-one dip is the honest cost and it is why this is hard to sell - designers get judged on the peak, which is the number that does not matter. The transfer board in track B session 4 puts figures on both ends of that trade.
What this session teaches, and where it came from
The retention figures referenced here are produced by the transfer board in b4-transfer.html and are a teaching model, listed in full in the course's source map.
Designer a4 cheat sheet · pin this
You have an activity that is the objective. Session a5 turns your ten real cases into assessment: questions that test the ability rather than recall of your slide, wrong answers built out of the misconceptions people actually have, and feedback that changes the next attempt instead of announcing a verdict.