Applied library · Young Founders
The Horizontal Strategy - Satoshi Light horizontal logo A chamfered H symbol with a right-facing arrow beside the words The Horizontal Strategy. TM
Applied library · The Validation Canvas · Young Founders edition

Writing it down doesn’t make it true.

The Validation Canvas for young founders

You have an idea. Maybe it is a good one. This course will not tell you that, because nobody can. It will show you how to find out for yourself, one honest step at a time, before you spend your money or your Saturdays on a guess. There is a one-page idea card to start with, seven short modules with a box to fill in for your own idea, a table of the grown-up words, and four young founders whose decisions are followed all the way to an answer. Everything you write is saved on this device.

Begin
Young Founders edition

Before you start

Most people with an idea ask one question: is it good? That question cannot be answered, by you or by anyone else, and it gets answered anyway, by whoever is most confident, by how you feel today, or by a friend who would never say no to you. This course asks a different question: what are you about to spend on this idea, and has the idea earned it yet? That question has an answer, because you can check it.

Here is how the course is arranged. First, an idea card: one page where you write down what you are counting on. Then seven modules. Each one teaches one part of the method and ends with a box you fill in for your own idea. The ten questions before you start. The eight things you have to be honest about, and a ninth the canvas works out by itself. The seven questions every claim has to answer, and the five-step scale that says how much your evidence is worth. The bar your idea has to clear, and the gate that checks it. Then the answer, with its conditions. At the end there is a table of the grown-up words, because the boxes you fill in here are the same boxes real founders use. Only the labels are different.

The founders we follow The same four young founders appear in this course and in the Wisdom course, so you only have to meet them once. Jay, 15, makes hats for Roblox avatars. Priya, 16, resells hoodies and trainers on Depop. Theo, 18, builds a Fortnite island. Dev, 16, washes cars on his street. They are made up. The rules, fees and sums they run into are real, and were checked on 21 October 2026. The decision we follow through every module is Priya’s: she has £20 of birthday money and four free Saturdays, and she wants to use them to buy hoodies at charity shops and sell them on Depop. A parent has agreed to set up the payment side, because Depop needs an adult’s bank account and ID for anyone under 18. If she goes ahead, the £20 and the four Saturdays go on this. If she does nothing, the money stays in the drawer and the Saturdays go on something else. Not deciding is also a decision.
If you are the mentorRead the same page the young founder reads; there is no separate version. Your job is not to judge the idea. It is to ask the question at the end of each module, out loud, and then keep quiet while they answer. The course is deliberately demanding of enthusiasm, including yours: every hopeful claim is asked to show its evidence. When they write something vague, do not fix it for them. Ask what would prove it wrong. The boxes they fill in use the same structure as the real Validation Canvas, so their work exports straight into the adult tool when they are ready. The last module is a table that bridges the two sets of words. Wherever an account, a payment or an age rule comes up, you are the adult the course means when it says “ask a parent or a teacher”.
The idea card

What am I counting on?

Watch the film for the idea card, then read on. Film, 5 min 01, with captions.

Before the seven modules, fill in one card. It takes five short answers, and it is the shortest honest description of a business idea there is. You will come back to it at the end, and you will probably change it. That is what it is for.

The five questions are: What am I counting on? Who pays, and with what? What would count as proof? What does it cost me? What is the first test? Each one has a box below. Write a guess if a guess is all you have. A guess written down can be checked. A guess kept in your head cannot.

Priya’s idea card Counting on: that strangers on Depop will pay about £18 for a hoodie I bought for £4. Who pays: the buyer, with their own money, into an account a parent runs for me. Proof: a stranger, not a friend, pays for one. Cost to me: £20 of birthday money, four Saturdays, and a parent’s time to set up the account. First test: list three hoodies at £18 and see whether two sell to strangers within four weeks.
Mentor: the question to ask here “Which of these five could be wrong?” All five can. If they say none, start again.

Your idea card

Five answers. Short ones. You can change them later, and you should.
Module 01

The ten questions before you start

Watch the film for this module, then read on. Film, 5 min 48, with captions.

Most people think a business plan is for working out whether an idea is good. Nobody can work that out. So this canvas asks a different question first: what exactly am I about to spend on this, and has the idea earned it yet? “Good” is a feeling. “Earned” is something you can check against evidence. The whole course is built on that swap.

Before you fill in anything else, you answer ten short questions. grown-up name: the decision frame They are: the idea in one sentence; who the customer is, as a real person; what is going on for them; what you want out of it; what you are deciding; what it will cost you; when you have to decide by; what going ahead actually means; what success looks like, as a number; and what would make you stop. The last two are written now, before you know anything, because a target you set after seeing the results is not a target. In the tool, the answer stays locked until all ten are filled in. That lock is the first rule of the method.

THE TEN QUESTIONSThe canvas refuses the question everyone asks, and answers a better one."Is this a good idea?"no one can answer it, andeveryone does anywayTHE TENThe idea, one sentenceThe customer, a real personWhat is going on for themWhat you want out of itWhat you are decidingWhat it will cost youDecide by whenWhat going ahead meansWhat success looks likeWhat would make you stopinsteadTHE ANSWERlocked until allten are filled inWithout the ten questions it is a description of an idea. With them, it is a decision.
The vague question replaced by ten short answers, feeding an answer that stays locked until all ten are done

What to examine

How it connects to the eight thingsThe ten questions connect to Evidence first: a target written in advance is what lets the evidence disagree with you later. They connect to Saying yes, because the commitment you are weighing is the same kind you will ask customers to make. And they connect to What they want in your own head: a locked answer protects you from the person most likely to skip the ten questions, which is the person who already believes.
Priya’s ten answers Priya arrived with the usual question: is reselling hoodies a good idea? The ten questions replaced it. Idea: buy hoodies at charity shops for about £4 and sell them on Depop for about £18. Customer: a sixth-former who wants a particular brand without paying the shop price, and searches Depop by brand and size. Not “people who like hoodies”. Decision: spend £20 and the next four Saturdays. Deadline: a parent has agreed to run the payment account for one term; if nothing has sold by half term, in four weeks, the account closes. Going ahead means: buy up to five hoodies, photograph them with every mark showing, list them, and post each one within two days of a sale. Success, written before anything is listed: three sold to strangers within four weeks. Stop: nothing sold in four weeks, or a hoodie costing more than £8 to buy. The hardest line to write was the stop line. Ask a parent or a teacher to read your ten with you, because one of the ten usually involves them.
Mentor: the question to ask here “What number would make you stop, and would you actually stop?” Wait for the second half.

Your ten questions

Answer them for a real idea of yours. The check below works like the real tool: it stays quiet until all ten are filled in.

Failure patterns

Module 02

The eight things you have to be honest about

Watch the film for this module, then read on. Film, 5 min 42, with captions.

With the ten questions done, the canvas asks eight more, one at a time. Each one pins down something specific. grown-up name: the eight areas They are: who actually has the problem, and who pays; what they really want and what they are worried about; how big the problem is next to what they do now; what you are promising, and where the promise stops; how they will find out about you; what they have to do, risk or pay to say yes; whether the sums work; and whether you can make and deliver it every time, not just once.

How you answer matters more than how much you write. Each of the eight wants a claim, not an essay: one or two sentences that could later be shown to be wrong. “Teenagers like hoodies” cannot be wrong, so it is worth nothing. “Sixth-formers who search Depop by brand will pay £18 for a clean hoodie in their size” can be wrong, so it is a claim. “I don’t know yet” is a good answer too. Written down, it becomes a question with your name on it. Left blank, it becomes a gap that quietly votes yes.

THE EIGHT THINGSEight you answer. A ninth the canvas works out from your answers.01Who really has it02What they want03The problem now04The promise05How they hear06Saying yes07The sums08Making it happenTHE SEVEN QUESTIONSrun underneath all eight, asking the same seven things about everything you claim09Does it allfit togetherworked out,not writtenNo answer stands on its own, and the ninth only speaks once the other eight have.
The eight things you answer, sitting over the seven questions, with the ninth outlined because the canvas works it out rather than you writing it

What to examine

How it connects to the eight thingsThis module is the eight things, so it connects to all of them, but three matter most. Who really has it turns your customer line into a real person with a real problem and a real wallet. The promise is where most people go vague, because a promise with an edge can let someone specific down. And The sums is where excitement meets arithmetic, which is why it is the one most often left “for later”.
Priya’s eight answers Priya’s eight answers took an afternoon, and two of them changed the idea. Who really has it: the buyer is a sixth-former spending their own money; the money arrives in a parent’s bank account, because Depop needs an adult’s details for a seller under 18. What they want: the brand, in their size, without the shop price, and not a fake. The promise got its edge: hoodies and trainers in the brands people search for, every mark photographed, posted within two days, and nothing else. The sums wrote their claim cold: a £18 sale leaves about £13 after Depop’s fee and the £4 she paid, and the buyer pays the postage. Saying yes recorded the objection a friend raised at lunch, word for word: “How do I know it’s not a fake?” That question is now answered in every listing, with the label in the photo. Making it happen found the weakest answer: the parcelshop is only on her way home on Saturdays.
Mentor: the question to ask here “Which of these eight answers would you least like to be tested?” That one is the weakest, whatever they say next.

Your eight answers

Answer each of the eight for your idea in one or two sentences that could be wrong. The check below looks for the words that mean you stopped checking.

Failure patterns

Module 03

The clash check

Watch the film for this module, then read on. Film, 5 min 55, with captions.

This is the part no other canvas does. You do not fill in the ninth thing. The canvas works it out by holding your eight answers up against each other and asking one question: can all of these be true at the same time? grown-up name: coherence You wrote each answer on its own, maybe on its own day. When two of them clash, it is not because you were careless. It is what perfectly reasonable answers do when they have never been read together.

A clash always has the same shape: two answers, each fine on its own, impossible together. A promise to post within two days, and a parcelshop you can only reach on Saturdays. A refund promise that makes saying yes easy, and a sum that only works if no parcel ever comes back. Nobody wrote a contradiction on purpose. The clash check exists because the most expensive mistakes in a business are not wrong answers. They are two answers that cannot both be true, and you cannot see that from inside either one.

THE CLASH CHECKThe ninth thing is a cross-examination. Your answers give evidence against each other.THE PROMISE, 04Posted within two days.Every mark photographed,not other schools.THE SUMS, 07The parcelshop is only onthe way home on Saturdays.clashSAYING YES, 06Refund if not as described,so nobody has to commit.THE SUMS, 07Each hoodie leaves about £13.One return takes most of a sale, by design.clashNeither answer is wrong on its own. Together they cannot both be true, and only the clash check sees it.
Two clashes at Priya’s: the promise against the parcelshop, and the refund against the sums

What to examine

How it connects to the eight thingsThe clash check connects to The sums most often, because the numbers are where two answers finally have to share a page. It connects to The promise by enforcing its edge: an edge the plan ignores is not an edge. And it connects to Evidence by giving you the best list of tests you will ever get, because every clash is a question with exactly two possible answers, and one cheap test usually settles which.
Priya’s clashes Priya’s canvas produced two flags, and both survived being argued with. The first: her promise says posted within two days, and her Making it happen answer says the parcelshop is only on her way home on Saturdays. Both cannot stand. Either the promise softens, or the route changes. Priya changed the route: there is a parcelshop two streets from school, so a parcel now goes the next school day, and Saturday morning covers anything sold on a Friday. Every parcel leaves within two days, and the listing says so. The second flag was the refund. “Refund if not as described” against about £13 left per hoodie meant that one returned parcel, with postage both ways, would wipe out most of a sale by design. So the photographs now show every mark, the description gives the chest measurement, and the refund promise stays, because the honest listing is what makes it rare. Neither fix was clever. Both were invisible until the answers were made to argue with each other.
Mentor: the question to ask here “Read me answer four and answer eight, one after the other.” Then say nothing. The pause usually does the work.

Your clash check

Write down the clashes between your own answers. A canvas with no clashes recorded has usually not been read against itself.
Answer AAnswer BThe clash, in one sentence

Failure patterns

Module 04

The seven questions every claim has to answer

Watch the film for this module, then read on. Film, 5 min 58, with captions.

Underneath every one of the eight things, the canvas runs the same seven questions, and they never change. grown-up name: the evidence layer What are you claiming? Why do you believe it? What have you actually seen? What does not fit? What breaks if you are wrong? What would you test next? And what result would count as a pass? The first three show how little most claims rest on. The last two turn that into something to do.

Two of the seven do most of the work. What does not fit makes you write down the evidence against you, and an empty line there almost never means there is none. It means nobody looked. What breaks if you are wrong asks how much the whole idea leans on this one claim: does it fall over, does it wobble, or does it barely notice? grown-up name: criticality: load-bearing, material, supporting That label is what the gate reads later, so it is worth an argument now. Most claims feel as if the whole idea leans on them. Most do not. The one that does deserves your next test.

THE SEVEN QUESTIONSOne claim, walked through the seven questions that put a price on it.CLAIM: "Strangers who search for the brand will pay £18 for a hoodie."WHY I BELIEVE ITSix friends said "yes, I’d buy one" at lunch.WHAT I HAVE SEENPeople saying yes. No money, no order, nothing done.WHAT DOES NOT FITTwo friends’ parents already bought a £6 revision guide and said no more.IF I AM WRONGThe whole idea falls over. Nothing else matters if this is wrong.NEXT TESTAsk the 25 people in my science class to pay up front, the week the timetable goes up.PASS MARKEight paid orders, written down before I ask anyone.The seven questions do not ask whether you believe the claim. They ask what the claim would survive.
Priya’s will-they-pay claim run through all seven questions

What to examine

How it connects to the eight thingsThe seven questions are evidence made into a habit, so they connect to Evidence by definition. In practice they hit Who really has it hardest, because that is where the most confident unchecked claims live. And they connect to What they want in both directions: the customer’s wants, which you are claiming to know, and yours, because the “what does not fit” line catches you looking only for good news.
Priya’s claim, tested Priya ran her central claim through the seven questions: strangers will pay £18 for a hoodie she bought for £4. Why she believed it: eight friends said they would buy one. What she had seen: nothing. A yes at lunch is a sentence, not a purchase. What did not fit, once someone looked: Depop’s suggested price for that brand, built from items that actually sold, sat between £12 and £16. If she is wrong: the whole idea falls over. Next test: list three hoodies at £18 for two weeks. Pass mark, written before the first listing: two sold to strangers. The test cost three Saturdays and £12 of stock. It is the difference between asking your customers and asking your mirror.
Mentor: the question to ask here “Who have you talked to who has no reason to be nice to you?” Friends say yes. The evidence is in the people who have no reason to.

One claim, seven questions

Take the claim your whole idea leans on most and make it answer all seven.

Failure patterns

Module 05

How much is your evidence worth?

Watch the film for this module, then read on. Film, 6 min 07, with captions.

Once a claim has some evidence behind it, the canvas asks how strong that evidence is, and it grades on a five-step scale. grown-up name: the evidence scale Step one: I think so. Step two: they said so. Step three: I saw it myself. Step four: they did something. Step five: they paid. The steps translate what people say into what it is actually worth, and the exchange rate is strict. “I would definitely buy one”, said warmly by a real person, is still step two.

The scale keeps itself honest with ceilings. Evidence that is people saying things stops at step two, however many people and however excited they are. Evidence of people doing things (replying, saving a listing, turning up, giving you their time) stops at step four. Only paying reaches step five. This is not unkind about your friends. It respects the difference between a kind sentence and money changing hands, and every founder who counted kind sentences as customers has paid the difference themselves.

THE FIVE STEPSFive steps, and a ceiling set by what kind of evidence it is.1I think sosaid by you2They said sosaid by them3I saw it myselfseen by you4They did somethingdone by them5They paidpaid by them"I would definitely buy one" can still be step 2."Here is the £18" is a different kind of sentence.Saying stops at 2. Doing stops at 4. Paying is the only way to 5. Excitement does not climb the steps.
The five steps rising, with the said, did and paid ceilings

What to examine

How it connects to the eight thingsThe scale connects most directly to Saying yes, because commitment is exactly what the steps measure: what the customer was willing to do or pay. It connects to How they hear as a filter on flattery, since a reply to your post is only evidence at the level of what the person then did. And it connects to The problem now, because what people already pay for their current fix is step-five evidence about the problem, sitting in plain sight.
Priya’s test result Priya’s listing test ran for two weeks, and the steps did their job. Of the three hoodies listed at £18, two sold to strangers in another city, one at £18 and one at £16 after a price drop. Two paying customers is step five, for one claim about one kind of buyer: people who search for that brand will pay about £16 to £18 for a clean hoodie in their size. It says nothing yet about other brands, and the canvas keeps the claim exactly that small. Meanwhile the eight friends who “would definitely buy one” stayed where they belonged, at step two. Three of them were sent the listing. None bought. Forty people liked the listing, which is step four at best, and the canvas does not let a like count as a sale.
Mentor: the question to ask here “Which step is that, really?” Ask it every time they say “people loved it”. Do not accept a step above two for anything that was only said.

The grading bench

Grade three real claims of yours. The bench applies the ceilings for you, the way the real tool does. The gate in module 06 takes each claim’s step from here, so grade a claim here before you lean on it there.
ClaimWhat kind of evidenceWhich step you would give it

Failure patterns

Module 06

How big a promise are you making?

Watch the film for this module, then read on. Film, 5 min 50, with captions.

Now the two halves meet. The ten questions told the canvas what you are about to spend. The steps graded what you actually know. This module is the sum between them. Every commitment sits on a ladder: Look, Try, Small run, Go, Grow. grown-up name: the commitment ladder: explore, experiment, pilot, build, scale Each rung sets a bar, from step one up to step five, because the same idea can deserve a yes to a small try and a no to a full launch. “Has this idea earned it?” means nothing until you say what “it” is.

Then the gate reads your claims against the bar, one at a time, never as an average. grown-up name: the gate The rules are short. No claim the idea leans on can sit below the bar. No claim the idea leans on can have evidence against it. The ten questions must be complete, with the targets written in advance. That is why an idea can look strong overall and still fail the gate: seven good answers and one wrong claim that everything leans on is a good-looking idea that does not work. The gate exists so that one broken leg cannot hide behind three good ones. One note about the ladder below: Looking needs no proof at all, and Trying needs a named test rather than a step, because you try something exactly when you do not know yet. From a Small run upward, the steps shown are the usual bar.

THE BAR AND THE GATEYour commitment sets the bar. The gate reads claims one at a time, never an average.Lookbar: step 1Trybar: step 2Small runbar: step 3Gobar: step 4Growbar: step 5Priya's askTHE GATENo leaned-on claim below the barpassesNo leaned-on claim with evidence against itpassesAll ten questions done, targets written firstpassesA good average with one broken legnever passesNo averages. An idea can look great overall and still rest on one wrong claim it cannot do without.
The commitment ladder with its bars, and the gate rules that read claims one at a time

What to examine

How it connects to the eight thingsThe gate connects to The sums as its enforcer, because what you are committing is money and time, and the bar is what stops step-two evidence spending step-five money. It connects to Making it happen through your “can I actually deliver this” claims, which are read by the same gate. And it connects to What they want at its sharpest: the gate cannot feel how excited the room is, and that is the feature the most excited person in the room will like least.
Priya at the gate Priya’s ask was Go: £20 and four Saturdays. Bar: step four. The gate read three leaned-on claims against it. Strangers will pay: step five, for one brand, after the listing test. Hoodies can be bought for about £4: step three, because one Saturday at two charity shops found three at that price, which is one trip, not a supply. A parent will run the payment account for the term: step two, because it had been said at the kitchen table and nothing had been set up. Verdict, as asked: blocked, and the gate named the rule: two leaned-on claims below a Go bar. Verdict for the rung below, a Small run of five hoodies: passes, on one condition. The account, the ID check and the Direct Debit are set up with the parent before any stock is bought, which turns the step-two claim into a step-four one, because somebody did something. The idea did not fail. The size of the yes changed.
Mentor: the question to ask here “Which rung is this really?” Most young founders name the rung below the one they mean, because the bar is lower. Make them say the real one.

The gate, run on your idea

Set your commitment and list the claims your idea leans on, written as you wrote them on the bench in module 05. The gate takes each claim’s step from the bench, and gives no result until your ten questions in module 01 are answered. To keep things simple this course uses one bar for every claim; the real tool sets a different bar for each of the eight areas, because a delivery claim and a demand claim need different proof.
Claim the idea leans onCurrent step

Failure patterns

Module 07

The answer, and the five honest routes

Watch the film for this module, then read on. Film, 6 min 04, with captions.

The answer is not a score and not a pat on the back. grown-up name: the verdict It is a sentence with parts. A yes comes with conditions, limits and a date to check again, because a yes with no conditions is a mood with a signature. A not-yet names its route, and there are five. Test more, when one named test would settle the open question. Go smaller, when the evidence is real for one group and missing for everyone else. Ask a different question, when the ten questions were pointed at the wrong decision. Wait, with a written trigger for when to come back. And Stop, when the idea has had its fair hearing. Every answer also names the rule that decided it, so you are never arguing with a feeling.

Then the tool does one small thing that matters more than any feature: it writes down its answer and yours, separately. You can overrule it. Sometimes you should, because the canvas reads your claims and you read the world. But the difference is written down, with your reasons, and it goes into a one-page record with the ten questions, the graded claims and what the gate said. grown-up name: the Validation Decision Report Three months later, that page is the difference between learning from the decision and having the same argument again.

THE ANSWER AND THE ROUTESA yes carries conditions. A not-yet names its route. Both name the rule that decided.THE ANSWERrule namedGOwith conditions, limitsand a date to check againTest moreone open question; one named test settles itGo smallertrue for one group; commit there onlyDifferent questionthe ten questions pointed the wrong wayWaitcome back on a written triggerStopthe idea has had its hearingThe tool writes down its answer and yours, separately. The gap is a record, because the final call stays with you.
The answer branching into go-with-conditions and the five named routes

What to examine

How it connects to the eight thingsThe answer connects to Saying yes one last time, because it is a commitment decision and behaves like one: conditional, dated, revisitable. It connects to How they hear inside your own house, since an answer with a named rule can be explained to a parent in two minutes. And it connects to Evidence going forward: the record’s targets and check-again date are what reopen the decision on schedule, not after something has gone wrong.
Priya’s answer Priya’s one-page record went to her mentor, and then to the parent whose account it would be. Tool’s answer: Go, smaller. Group: buyers who search for the one brand that sold. Conditions: the Depop account, the ID check and the Direct Debit set up with the parent before any stock is bought; the price paid for each hoodie tracked on one sheet; nothing bought above £8. Check again: four weeks, against three sold to strangers. The original Go, £20 and four Saturdays on any brand, was recorded with its route, Go smaller, and its rule, two leaned-on claims below the bar. Priya’s own decision matched the tool’s on everything except one thing: she kept a second brand on her shopping list even though the evidence for it was step one. Her reason was written down: it costs £4 to find out, and she would count it separately. The parent read one page, asked two questions, and set up the account that evening. Nobody discussed whether reselling hoodies was a good idea. There was nothing left for that question to do.
Mentor: the question to ask here “If the tool says not yet and you go anyway, what will you write down as your reason?” If they cannot write it, that is the answer.

Your answer, written down

Close the loop on the decision you framed in module one. The record is the product.

One page with your decision, your graded claims, what the gate said and your answer, ready to print or save as a PDF.

Failure patterns

Module 08

The real words

Watch the film for this module, then read on. Film, 6 min 24, with captions.

Every box you filled in on this course is a box on the real Validation Canvas. The same ten questions. The same eight things. The same seven questions under each claim, the same five steps, the same bar, the same gate, the same routes. Real founders, with real money on the table, use exactly this. The only thing that was different here was the labels.

The table below bridges the two sets of words. Read it once, and the adult tool will not feel like a different instrument. Your mentor can point at this page whenever a grown-up word turns up. When you are ready, press Export my work at the foot of this page, open the real tool, go to Canvases and choose Import JSON. Your ten answers arrive in its frame and your eight answers as claims in its eight areas, marked untested until you add evidence there; the claims you graded here travel in its notes, for you to place in the right area. The real tool keeps its work in the browser you use, so export a backup from it as well. The idea card is this edition’s own; its five answers are a short form of the ten questions and the central claim.

What we called it hereWhat the real tool and grown-ups call it
Writing it down doesn’t make it trueFilled in is not the same as proven
The ten questions before you startThe decision frame
The eight things you have to be honest aboutThe eight areas
Who really has the problem, and who paysCustomer reality
What they want, and what they are worried aboutPsychology and motivation
How big the problem is, and what they do nowProblem and alternatives
The promise, and where it stopsOffer
How they will hear about youCommunication and access
What they have to do, risk or pay to say yesConversion and commitment
The sumsEconomics
Making it happenOperations
The clash checkStrategic coherence, the ninth area
The seven questions every claim has to answerThe evidence layer
Why I believe it / What I have seen / What does not fitBasis / Evidence / Counter-evidence
The whole idea falls over / wobbles / barely noticesLoad-bearing / Material / Supporting (criticality)
The five steps: I think so, They said so, I saw it, They did something, They paidThe evidence scale: untested belief, anecdotal indication, observed reality, behavioural evidence, commercial evidence
Said / Did / Paid ceilingsThe said, done and paid caps
The ladder: Look, Try, Small run, Go, GrowThe commitment ladder: Explore, Experiment, Pilot, Build, Scale
A claim the idea leans onA load-bearing assumption
The bar and the gateThe bar and the gate (unchanged)
The answerThe verdict
Go / Test more / Go smaller / Ask a different question / Wait / StopProceed / Test further / Narrow / Reframe / Pause / Stop
The one-page recordThe Validation Decision Report
Mentor: the question to ask here “Open the real tool. Which box do you not recognise?” There should not be one.
The map

One instrument, one loop

Here is the whole method, start to finish. The idea card says what you are counting on. The ten questions turn an idea into a decision with a deadline and a target. The eight things turn the decision into claims. The clash check reads the claims against each other. The seven questions put a price on each claim, and the five steps grade what it rests on, with a ceiling set by what kind of evidence it is. The bar rises to meet the commitment, the gate reads the claims one at a time, and the answer comes back with conditions or a route, the rule named, your call written down next to the tool’s. Nothing in the loop is clever. Its whole power is that no step can be skipped quietly.

Now use it. Open the interactive tool with your mentor, write down the thing you are actually about to spend money or Saturdays on, and run it through, badly, tonight. Your first canvas will be thin. That thinness is useful: it is a map of what you do not know yet, in order. You have finished this course when a real decision of yours has been through the loop, the one-page record exists, and something changed because of what it said. If the decision involves an account, a payment or an age rule, that is the moment to ask a parent or a teacher.

THE WHOLE INSTRUMENTOne loop, from a live decision back to a live decision.Ten questionsthe decision, pinned downEight thingseight answered, one readSeven questionsasked of every claimThe barset by what you will spendThe gateclaims, no averagesThe answerconditions or a routethe nextdecisionRun it in the interactive tool, on something real, this week. The canvas is not reading material. It is a machine you feed.
The six stations of the loop, returning to the next live decision

This week

One line closes the course.

Open the interactive tool

Saved on this device