INTD 317W04BRING YOUR W03 MATERIAL

Synthesis &
ecology.

THE PAYOFF SESSIONTUE 29 SEP 2026

This session carries the course. If the payoff exercise lands, the notebook justifies itself for the rest of term. If it does not, the notebook becomes documentation homework by week six.

Prep: affinity mapping materials, stakeholder ring template printed, notebook rubric visible on the wall, and three fallback examples from your own real projects in case the room's material is too thin to show a difference. Rehearse the payoff.

Tables seeded from the W03 domain sign-up — at least one person per table who has made a journey map or a stakeholder map before.

Synthesis, and the ethics of encoding.

CLUSTERING IS A CLAIM
When you group two quotes together you are asserting they are the same kind of thing. That is an argument, and it can be wrong.
NAME THE CLAIM
A theme called “trust” says nothing. A theme called “they trust the nurse and not the system that pays her” can be argued with, which is what makes it useful.
ENCODING IS POWER
Whatever you encode becomes the thing the rest of the project optimises for. Everything you left out stops existing for every downstream decision.
THE OUTLIER, AGAIN
Same failure as last week's transcript. A theme that fits everyone usually fits nobody in particular.

Five minutes. Ask it for a persona.

Give me a persona for a health app user.

WHAT COMES BACK, RELIABLY
Sarah, 34. Busy marketing manager.
Values efficiency. Tech-savvy but time-poor.
Frustrated by cluttered interfaces.

What's wrong with it? Do not let anyone say “it's generic” and stop there.

Needs no prior training to appreciate, which is exactly why it works in a mixed room — the student who has never made a persona can see this is bad.

Steer toward: it describes a user, and it describes an individual, and it is a composite of nobody. Then the turn: personas are out as a built artifact in this course. They are the most rigorously taught thing in the core, and emerging technology's defining problem is its effect on people who never opted in — who by definition are not users.

Affinity mapping. Your own material.

ONE QUOTE PER STICKY
Verbatim. Not your summary of what they meant.
CLUSTER IN SILENCE
Move things without explaining. Explaining early locks the first arrangement in.
THEN NAME EACH CLUSTER AS A SENTENCE
Not a noun. A claim you could be wrong about.

Twenty minutes. Five-minute micro-teach first if anyone at your table hasn't done this.

Break.
Ten minutes.

BACK AT 9:40LEAVE YOUR CLUSTERS UP

Stakeholder ecology. Five rings.

DIRECT USERSINDIRECT USERSNON-USERS AFFECTEDDOWNSTREAM IN TIMENON-HUMAN SYSTEMS

Then the harder pass, and the one that is actually graded: which of these did your research reach, and which are you guessing about?

Micro-teach opens this too. Then build for real, on their own domain.

The second pass is where the learning is. Most rings will be guesses in week four, and that is fine — the skill is knowing which is which and saying so.

Now bring the model in.

THE LIFT

Genuinely strong at who am I forgetting. It will push the ring diagram past the obvious transaction faster than a table of four people will.

THE LIE

Generic categories with no local knowledge. “Regulators.”“The community.” It does not know which regulator or whose community, and it will not tell you that it doesn't.

THE CATCH

Your research tells you which of these are real here. The model's list is a hypothesis. Yours is evidence. Knowing the difference is the whole job.

This is the same question week twelve asks with nine more weeks of evidence attached: who bears the cost.

THE PAYOFF · 70 MIN

Does any of
this actually
work?

Seventy minutes, bought by moving the Research Pack to W05. Use all of it — do not let the ecology run long.

Five steps. Your own research.

05 MIN · BASELINE
Ask a model, cold, for three design directions in your domain. Save it.
30 MIN · BUILD
Assemble into your W02 context: the problem as you now understand it, three verbatim quotes, your stakeholder ecology, two hard constraints, one stated non-goal.
05 MIN · SECOND PASS
Same request. Context supplied.
20 MIN · COMPARE
At your table. What changed? Was the second actually better, or just longer? What does it still get wrong?
10 MIN · WRITE IT
The comparison goes into decisions.md. This is your first entry that will have a parent.

If the tool, the account or the device doesn't work for you, or you'd rather not use one: take my two prepared outputs, run steps four and five on those, and write the comparison. The comparison is the assessed half either way.

The comparison is evaluative, not constructive. They are not proving they can make an artifact — they are proving they can judge one. That is the scarce skill and it is only runnable with a cohort that has prior training. Say this out loud; it reframes the whole exercise for the strong students.

For some of you
not much will change.
That's a finding. →

SAY THIS BEFORE THEY STARTTHIN CONTEXT, THIN GAIN

Say it in advance, not afterwards. Announced first it is a hypothesis being tested; said afterwards it sounds like an excuse for an exercise that failed.

Thin context produces thin gains — which is the thesis stated in reverse, and it is a finding about their research rather than a failure of the method. The students who get this result have learned the most useful thing available today.

If the room's material is genuinely too thin to show anything, this is where the three prepared examples from your own projects come out.

Before you go.

RESEARCH PACK
Due next week. If today showed you your research is thin, this week is when to fix it.
NEXT WEEK
Writing the spec. Problem statements that name a deficiency instead of an aspiration.
DOMAIN SWAP
Week five is the one no-questions swap window. After that you are committed.
NOTEBOOK
Commit today's comparison tonight while you still remember what surprised you.
AUTOPSY
Two names for next week.

The model's list is
a hypothesis.
Yours is evidence. →

W05 · TUE 6 OCTWRITING THE SPEC · RESEARCH PACK DUE

Speaker notes

next
back
scroll next & back
swipe next & back
s speaker notes
o overview & jump
t start the 8:30 clock
f fullscreen
esc close

Turn your phone sideways

These slides are a fixed 16:9 canvas. In portrait they shrink to an unreadable strip; landscape gives you the whole slide.