Free resources

Tools you can use this week

Five working documents for people who build learning for other people. All of them live on this page. Nothing to download, no email required, no PDF hidden behind a form.

Each tool has three ways in. Read it here. Print it from your browser, which strips the menu and the buttons and gives you a clean sheet. Or copy it as plain text and paste it into whatever you actually work in, a doc, a ticket, a wiki page, the back of a meeting agenda.

They are all built on the same four pieces: path, practice, feedback, transfer. If a course is failing, it is almost always failing at one of those four, and almost never at the one people blame first.

Tool one

The design check

Twenty-four questions. Answer them honestly rather than generously. Any “no” is a piece of design work you have not done yet, which is useful information, not a failure.

Before you build

I can name what the learner will be able to do afterwards, in a verb somebody could watch.
I know who this is for, specifically enough to picture one of them at their desk.
I know what they already bring, and I am not about to teach them something they use daily.
I know the real time budget, not the one in the project plan.
I know what happens on the other side: what they will be asked to do, by whom, and when.
I know what success looks like for the person paying for this, and it is not just completion.

Path

At every point, a learner can answer “what do I do next” without asking anyone.
The order is a reason, not a table of contents inherited from a document.
Each step is small enough to finish in one honest sitting.
Someone who misses a week can find their way back in without starting over.
Nothing important is hidden behind a click that looks optional.
I have removed at least one thing I liked because it did not serve the outcome.

Practice

Learners do the actual thing, not a description of the thing.
There is more than one rep, and the reps vary rather than repeat.
At least one task is close enough to real work that it could be used at work.
Getting it wrong is survivable and, ideally, informative.
The difficulty is in the thinking, not in decoding the instructions.
If AI could produce a passing answer in ten seconds, I have decided whether that is fine or fatal, and designed accordingly.

Feedback

Feedback arrives while the learner still cares about that piece of work.
The learner has somewhere to use the feedback: a revision, a next attempt, a second version.
Someone learning alone still gets a loop, through models, worked examples, criteria they can apply, or peer review.
The criteria were visible before the work started, not revealed with the grade.

Transfer, before you ship

There is at least one thing a learner is asked to do outside the course, in their own context.
Somebody other than me could run this without a phone call, and I have checked by asking one.

Tool two

Path, practice, feedback, transfer

One page for one course, module, workshop, or lesson. Fill the middle column first. The right-hand column is the one that saves you, because it forces you to name the failure before a learner finds it.

What am I designing:

Who is it for, and how much time do they really have:

Fill one row per piece. Print this page to write on it by hand.
Piece	What the learner actually does	How I will know it worked	Where this will break first
Path what do I do next
Practice reps that build capability
Feedback a loop, not a comment
Transfer the Monday after

The one question to finish on. If a learner did every single thing in this design and then went back to work, what would be different on the following Monday? Write it in one sentence. If you cannot, the transfer row is not done.

Tool three

Assessment in the age of AI, a planning sheet

Built on one move: stop assessing the artefact alone and start assessing the process that produced it. Work down the sheet. It takes about twenty minutes per task and it settles most of the arguments before they start.

Step one, decide the stance

For this task, which is true?

The tool is part of the capability. Using it well is what I am assessing. Let them use it, and assess the use: the prompt, the selection, the edit, the judgement about what to keep.
The tool would short-circuit the capability. The thinking the tool replaces is exactly the thinking I need to see. Design a task where the tool cannot stand in, usually by anchoring it to something only this learner has: their own data, their own context, their own prior work, a live conversation.

Whichever it is, write it into the task instructions in one plain sentence. Most disputes come from an expectation nobody wrote down.

Step two, plan the evidence

Four kinds of evidence that stay meaningful when fluent text is free.
Evidence	What it shows	How I will collect it
Drafts	That the work moved, and in what direction.
Decisions	What was chosen, what was rejected, and why.
Revision	That feedback was understood and acted on.
Live account	That the learner can stand behind the work unaided.

Step three, questions that survive

Add one or two of these to any submission. They cost the honest learner two minutes and cost everybody else the whole game.

What did you try first that did not work, and how did you know?
Name one thing you deliberately left out. Why?
Which part of this are you least confident about, and what would settle it?
If a colleague pushed back on your main claim, what would they say?
Point to the moment this piece changed direction. What caused it?
What in here comes from your own context and could not have been written about anyone else?

Step four, the fairness check

My criteria reward reasoning and fit, not polish and fluency.
I am not relying on a detector, and I have said so.
A second-language writer with excellent thinking is not penalised by this design.
A learner with no quiet hours and a shared laptop can still produce this evidence.
If I suspect something, I have a process that starts with a conversation rather than an accusation.

No paywall, ever

If something here saved you an evening

These sheets stay free and open, with no email gate and no upsell. If they earned their keep, a coffee is a kind way to say so. If not, take them anyway and use them well.

Opens PayPal in a new tab. Entirely optional.

Buy me a coffee

Tool four

The course post-mortem

Run this within a week of the thing finishing, while you still remember. Thirty minutes alone, or forty-five with the team. The rule that makes it work: describe what happened before deciding what it means.

What ran:

When, and for how many:

1. What happened

Where did people slow down or stop? Name the exact point, not the general area.
What question got asked more than twice?
What did people skip?
What took longer than I said it would?

2. What it means

Was the sticking point a path problem, a practice problem, a feedback problem, or a transfer problem? Pick one, even if it is close.
What did I assume about these learners that turned out not to hold?
What worked better than expected, and can I use more of it?

3. What changes

The one change with the largest effect for the least work:
The change I want to make but cannot yet, and what it is waiting on:
The thing I will stop doing:

4. What I will not blame

Cross these out as they come up: motivation, attention spans, “they did not read it”, the platform, the weather. Each one is usually a design symptom wearing a costume. If learners did not read it, the question is what made reading it feel optional.

Tool five

The twenty-minute learner test

A five-minute sanity check before anything goes live. The real learner is not the one with a free afternoon. They have twenty minutes, on a phone, after a shift, with somebody talking to them.

In the first sixty seconds, this person can tell what they are meant to do.
There is something finishable inside twenty minutes, and it feels finished, not abandoned.
They can stop and come back tomorrow without losing their place.
Nothing essential requires a desktop, a printer, a quiet room, or a live meeting at a fixed hour.
The instructions read cleanly on a narrow screen.
If they only ever complete this one piece and nothing else, they still got something worth their twenty minutes.

Six ticks and it will survive a real week. Fewer than four and the design is written for a learner who does not exist.

Using these with your team

Print them, paste them into your own working documents, run a workshop off them, translate them, cut the questions that do not apply. Keep a line of attribution and a link back so people can find the source. If you want to publish an adapted version or include it in something you sell, ask first and the answer is usually yes.

There are no PDFs on this page on purpose. A PDF goes stale, sits in a downloads folder, and needs an email address to get out of. Everything here is the live page, and your browser’s print dialogue will make a clean sheet out of any of it.

If you want the thinking behind the tools, the twelve-part series is free and covers the same ground at length, and the two books go further still.

THE DESIGN CHECK
Good Learning by Design - goodlearningbydesign.com

BEFORE YOU BUILD
[ ] I can name what the learner will be able to DO afterwards, in a verb somebody could watch.
[ ] I know who this is for, specifically enough to picture one of them at their desk.
[ ] I know what they already bring, and I am not about to teach them something they use daily.
[ ] I know the real time budget, not the one in the project plan.
[ ] I know what happens on the other side: what they will be asked to do, by whom, and when.
[ ] I know what success looks like for the person paying for this, and it is not just completion.

PATH
[ ] At every point, a learner can answer "what do I do next" without asking anyone.
[ ] The order is a reason, not a table of contents inherited from a document.
[ ] Each step is small enough to finish in one honest sitting.
[ ] Someone who misses a week can find their way back in without starting over.
[ ] Nothing important is hidden behind a click that looks optional.
[ ] I have removed at least one thing I liked because it did not serve the outcome.

PRACTICE
[ ] Learners do the actual thing, not a description of the thing.
[ ] There is more than one rep, and the reps vary rather than repeat.
[ ] At least one task is close enough to real work that it could be used at work.
[ ] Getting it wrong is survivable and, ideally, informative.
[ ] The difficulty is in the thinking, not in decoding the instructions.
[ ] If AI could produce a passing answer in ten seconds, I have decided whether that is fine or fatal, and designed accordingly.

FEEDBACK
[ ] Feedback arrives while the learner still cares about that piece of work.
[ ] The learner has somewhere to USE the feedback: a revision, a next attempt, a second version.
[ ] Someone learning alone still gets a loop, through models, worked examples, criteria they can apply, or peer review.
[ ] The criteria were visible before the work started, not revealed with the grade.

TRANSFER, BEFORE YOU SHIP
[ ] There is at least one thing a learner is asked to do outside the course, in their own context.
[ ] Somebody other than me could run this without a phone call, and I have checked by asking one.

Any "no" is design work not yet done, which is useful information, not a failure.

PATH, PRACTICE, FEEDBACK, TRANSFER: A WORKSHEET
Good Learning by Design - goodlearningbydesign.com

What am I designing: ______________________________________________
Who is it for, and how much time do they really have: _____________

For each piece, answer three columns.

1. PATH (what do I do next)
   What the learner actually does: ________________________________
   How I will know it worked: _____________________________________
   Where this will break first: ___________________________________

2. PRACTICE (reps that build capability)
   What the learner actually does: ________________________________
   How I will know it worked: _____________________________________
   Where this will break first: ___________________________________

3. FEEDBACK (a loop, not a comment)
   What the learner actually does: ________________________________
   How I will know it worked: _____________________________________
   Where this will break first: ___________________________________

4. TRANSFER (the Monday after)
   What the learner actually does: ________________________________
   How I will know it worked: _____________________________________
   Where this will break first: ___________________________________

THE ONE QUESTION TO FINISH ON
If a learner did every single thing in this design and then went back
to work, what would be different on the following Monday? One sentence.
If you cannot write it, the transfer row is not done.
__________________________________________________________________

ASSESSMENT IN THE AGE OF AI: A PLANNING SHEET
Good Learning by Design - goodlearningbydesign.com

One move: stop assessing the artefact alone, start assessing the process
that produced it.

STEP ONE - DECIDE THE STANCE. For this task, which is true?
  1. The tool is part of the capability. Using it well is what I am
     assessing. Let them use it, and assess the use: the prompt, the
     selection, the edit, the judgement about what to keep.
  2. The tool would short-circuit the capability. The thinking the tool
     replaces is exactly the thinking I need to see. Design a task the
     tool cannot stand in for, usually by anchoring it to something only
     this learner has: their own data, context, prior work, or a live
     conversation.
Write the answer into the task instructions in one plain sentence.

STEP TWO - PLAN THE EVIDENCE
  Drafts       - shows the work moved, and in what direction.  How: ____
  Decisions    - shows what was chosen, rejected, and why.     How: ____
  Revision     - shows feedback was understood and acted on.   How: ____
  Live account - shows the learner can stand behind the work.  How: ____

STEP THREE - QUESTIONS THAT SURVIVE (add one or two to any submission)
  1. What did you try first that did not work, and how did you know?
  2. Name one thing you deliberately left out. Why?
  3. Which part are you least confident about, and what would settle it?
  4. If a colleague pushed back on your main claim, what would they say?
  5. Point to the moment this piece changed direction. What caused it?
  6. What in here comes from your own context and could not have been
     written about anyone else?

STEP FOUR - THE FAIRNESS CHECK
  [ ] My criteria reward reasoning and fit, not polish and fluency.
  [ ] I am not relying on a detector, and I have said so.
  [ ] A second-language writer with excellent thinking is not penalised.
  [ ] A learner with no quiet hours and a shared laptop can still produce
      this evidence.
  [ ] If I suspect something, my process starts with a conversation
      rather than an accusation.

THE COURSE POST-MORTEM
Good Learning by Design - goodlearningbydesign.com

Run within a week of the thing finishing. Thirty minutes alone, forty-five
with the team. Describe what happened before deciding what it means.

What ran: _________________________________________________________
When, and for how many: ___________________________________________

1. WHAT HAPPENED
   Where did people slow down or stop? The exact point, not the area.
   __________________________________________________________________
   What question got asked more than twice?
   __________________________________________________________________
   What did people skip?
   __________________________________________________________________
   What took longer than I said it would?
   __________________________________________________________________

2. WHAT IT MEANS
   Was the sticking point a PATH, PRACTICE, FEEDBACK or TRANSFER problem?
   Pick one, even if it is close.
   __________________________________________________________________
   What did I assume about these learners that turned out not to hold?
   __________________________________________________________________
   What worked better than expected, and can I use more of it?
   __________________________________________________________________

3. WHAT CHANGES
   Largest effect for the least work: _______________________________
   Want to make but cannot yet, waiting on: ________________________
   The thing I will stop doing: ____________________________________

4. WHAT I WILL NOT BLAME
   Motivation. Attention spans. "They did not read it." The platform.
   The weather. Each is usually a design symptom wearing a costume.

THE TWENTY-MINUTE LEARNER TEST
Good Learning by Design - goodlearningbydesign.com

The real learner has twenty minutes, on a phone, after a shift, with
somebody talking to them.

[ ] In the first sixty seconds, this person can tell what they are meant to do.
[ ] There is something finishable inside twenty minutes, and it feels
    finished, not abandoned.
[ ] They can stop and come back tomorrow without losing their place.
[ ] Nothing essential requires a desktop, a printer, a quiet room, or a
    live meeting at a fixed hour.
[ ] The instructions read cleanly on a narrow screen.
[ ] If they only ever complete this one piece and nothing else, they still
    got something worth their twenty minutes.

Six ticks and it will survive a real week. Fewer than four and the design
is written for a learner who does not exist.