The Pack

---
name: 01-ownable-idea-audit
description: Run when somebody needs to know what they are actually known for. Trigger - "run 01" plus their current one liner, headline, bio and three to five recent posts.
---

# TASK 01 — THE OWNABLE IDEA AUDIT

## INPUT
Their headline, their one liner, their about
paragraph. Three to five recent posts, pasted
whole. What they sell, to whom, at roughly what
price band. Anything they have already tried to
be known for and dropped.

## PROCESS
1. Strip to the claim. Delete every adjective,
   credential, year count and client logo. What
   is left is the actual claim. If nothing is
   left, say that first, before anything else.
2. Run the echo test. Put the stripped claim in
   the mouths of three other people in the same
   market. If it still works in all three, it is
   a category description with their face on it,
   not an ownable idea. Say which three.
3. Sort into the six shapes. Coined category,
   coined problem, method, identity label,
   X versus Y, central argument. Pick the shape
   the evidence supports, not the shape they
   want. Coined category is the vanity choice.
4. Hunt the discounted thing. The difference is
   usually something they wave away because to
   them it is obvious. It lives in the posts, in
   the asides, never in the bio. Quote it back.
5. Name the enemy. What is wrong with how this is
   done today, said plainly enough that somebody
   in the market would push back.
6. Check it against the trigger. What was the
   buyer searching for the day before they found
   this person. If the idea does not meet them at
   that moment, it is a positioning statement for
   peers, not for buyers.
7. Kill the rest. Write the graveyard out loud:
   three good ideas, and the reason each one dies.
   Killing good ideas is the hard half of this.
8. Write the one sentence. Fifteen words or
   fewer, no coined term in it yet.

## OUTPUT
WHAT THEY CLAIM NOW, stripped. ECHO TEST, with
the three names it also fits. THE SHAPE, one of
six, with why. THE DISCOUNTED THING, quoted from
their own post. THE ENEMY. THE ONE SENTENCE. THE
GRAVEYARD, three killed ideas and why. WHAT TO GO
PROVE, one piece of evidence to publish inside 30
days.

## RULES
- Hand back ONE idea. Not a shortlist. A shortlist
  is doing the easy half and leaving them the hard
  half.
- No coined category unless they already have
  distribution. Categories need an audience to
  carry them. Coined problems do not.
- Never build the idea on their expertise or their
  years. Expertise is not what makes anyone
  choose them.
- If the idea needs proof they do not have, it
  goes in WHAT TO GO PROVE. Never write it as
  though the proof already exists.

## EDGE CASES
- The claim passes the echo test only because they
  invented a word for it. The coined term is doing
  all the work and the sentence underneath is
  generic. Strip the term, rerun the test on the
  plain English version, and report that result
  instead. A name is not an idea. If the plain
  version fails, they have branding, not position.
- The strongest ownable idea is the one that
  shrinks their market or contradicts what they
  currently sell. Do not smooth it. Name the cost
  in plain numbers of who it excludes, hand back
  both the safe version and the sharp one, and say
  which one you would bet on and why. Then let
  them choose. This is a decision about their
  next two years, not a copy tweak.

## GOOD LOOKS LIKE
"WHAT YOU CLAIM NOW, stripped: 'I help B2B SaaS
companies grow through better marketing.' That is
it. Everything else in the headline was adjectives
and years.

ECHO TEST: failed. I put that sentence in three
other mouths from your own feed and it worked in
all three. When a line survives being said by your
competitors, it is a category description wearing
your face.

THE DISCOUNTED THING: post from [DATE], seventh
paragraph, you wrote 'most founders I meet have
never sat in on a single sales call at their own
company.' You buried it. That is the idea. You
discounted it because to you it is obvious. To
them it is an accusation.

THE SHAPE: coined problem. Not a category. You do
not have the distribution to drag a new category
around, and you do not need one. Febreze did not
invent a category, they named nose blindness, and
then everyone had it.

THE ONE SENTENCE: Founders write messaging for a
buyer they have never listened to.

THE GRAVEYARD: 'the fractional CMO who actually
ships' dies because it is a competence claim and
competence is table stakes. 'Revenue-first
marketing' dies the echo test in under a second.
'The anti-agency' dies because the enemy is a
business model, not a belief, and you cannot build
five years of writing on hating agencies.

WHAT TO GO PROVE: sit in on ten sales calls for
one [CLIENT] and publish what the founder had
wrong. Not a case study. A transcript with the
gap in it. Inside 30 days."

---
name: 02-why-we-buy-issue
description: Run to draft the weekly issue. Trigger - "run 02" plus the finding or the buyer observation, the example you want to use, and what you want readers to do.
---

# TASK 02 — THE WEEKLY ISSUE

## INPUT
One finding: a study, a behavioural effect, or
something you watched a real buyer do. The example
you want to hang it on. What you want the reader
to go and do afterwards. Anything you are unsure
about in the source.

## PROCESS
1. Find the turn. The issue only exists if there
   is a moment where the reader expected A and
   gets B. Write the turn as one sentence before
   writing anything else. No turn, no issue.
2. Open in the reader's situation, not in the
   study. They are at their desk with a page that
   is not converting. Start there. The research
   arrives as relief, not as a lecture.
3. Give the finding a plain English one liner a
   marketer could repeat at standup, then the
   source line underneath: who ran it, when, on
   how many people, and what it does not prove.
4. One example, public only. A brand anybody can
   go and look at. If the only good example is a
   [CLIENT], see the edge cases.
5. Write STEAL THIS. The smallest version of the
   idea somebody can run before lunch. Under
   twenty minutes. If it needs a workshop, it is
   not steal this, it is a product.
6. Kill the third point. Two examples beat three.
   Less is more is not a slogan here, it is the
   edit.
7. P.S. does exactly one job. A product, a reply
   prompt, or a referral. Never all three.
8. Pull the tease. One LinkedIn post and one X
   post that stand alone and pay off without the
   click, plus the line that makes clicking worth
   it anyway.

## OUTPUT
THREE SUBJECT LINES. PREVIEW TEXT. THE ISSUE,
full draft. STEAL THIS. SOURCE LINE. P.S. THE
TEASE, one LinkedIn and one X, native, no "check
out my newsletter".

## RULES
- No study without who ran it, when, on how many
  people, and what it does not prove. A finding
  with no source line is an opinion and gets
  labelled OPINION in the draft.
- Never open with the name of the effect. Nobody
  ever opened an email because it said anchoring
  bias. They open because it says their page is
  not converting.
- Never invent a brand result, a lift percentage
  or a customer quote. If you want a number and
  do not have one, write the mechanism instead.
- Curiosity gap in the subject line only if the
  issue pays it off in the first three lines. A
  tease with no payoff is a device, and devices
  cost more than they earn.

## EDGE CASES
- The finding is a famous behavioural science
  study that has since failed to replicate.
  Priming and ego depletion are the obvious ones,
  but check anything pre-2015 and beloved. Do not
  quietly drop it. Write the issue about the
  replication instead, because "the study you have
  been quoting for ten years fell over" is a
  better turn than the original finding was, and
  it is the whole reason people read this rather
  than a guru.
- The only example that actually proves the point
  is a private client result. You cannot name
  them, and the anonymised version is weaker than
  no example at all. Convert it into a mechanism
  written as a hypothetical the reader can test
  themselves, say in the issue that you have seen
  it work and cannot show it, and go find a public
  brand doing the same thing. Never write "a
  client of mine" as if that were evidence.

## GOOD LOOKS LIKE
"SUBJECT: Your pricing page is doing the wrong job

Here's the thing about your pricing page.

You think it answers 'how much'. It doesn't. By
the time somebody scrolls it, they already have a
number in their head. Your page is not informing
them. It is confirming or breaking a guess they
already made.

Which means the job of that page is not clarity.
It is repair.

THE FINDING: people evaluate a price against
whatever number they saw first, even when that
number was arbitrary. Tversky and Kahneman, 1974,
the anchoring work. Lab study, small samples by
today's standards, replicated many times since for
the basic effect, not a guarantee for your
checkout.

WHERE YOU CAN SEE IT: look at how Apple lays out
storage tiers. The one they want you to buy is
never first and never alone.

STEAL THIS: open your pricing page. Find the first
number a visitor sees. If it is your cheapest
plan, you anchored them low and now every other
tier reads as expensive. Move one thing above it
today. Takes ten minutes. Watch what happens to
tier two.

P.S. If you want the words buyers actually use
when they think a price is wrong, that is what
Wallet-Opening Words is for. [LINK]"

---
name: 03-argument-or-device
description: The screen. Run before anything goes out and before anything gets accepted. Trigger - "run 03" plus the hook, email, mechanic, sponsor ask or partnership on the table.
---

# TASK 03 — ARGUMENT OR DEVICE

## INPUT
The thing on the table. A hook, a subject line, a
launch email, a scarcity mechanic, a sponsor ask,
an affiliate offer, a partnership. Who it lands
on. What it is meant to make them do.

## PROCESS
1. Say the argument in one sentence. Not the
   topic, the argument: a thing you believe that
   somebody could disagree with. If you cannot
   write it, stop here and return DEVICE, NO
   ARGUMENT. Everything else is wasted work.
2. Classify the mechanic. Which one is it doing:
   pratfall, envy, fascination, curiosity gap,
   loss aversion, social proof, scarcity, or a
   plain claim. Name it. Unnamed mechanics are
   the ones that get away with things.
3. Run the reversal test. Would this still be fair
   if the reader could see exactly what it is
   doing to them, and would you be comfortable
   explaining the mechanic in the issue itself?
   You teach this stuff for a living. If the
   answer is no, you are selling snake oil to
   people who bought the anti snake oil
   newsletter.
4. Find the inferred promise. Not what the copy
   says. What a tired reader will conclude it
   said. If the result only exists in the
   inference, the copy is lying with correct
   sentences.
5. Check the scarcity is real. A cap you will
   honour, a date you will not move, a cohort you
   will actually close. If any of those would bend
   for one more sale, it is fake and it comes out.
6. For anything inbound, ask the reader's
   question, not yours. Does the thing being
   promoted work for the person on the list. Money
   is not the test. Relevance and whether it works
   are the test.
7. Verdict. SEND, FIX, or KILL. If FIX, name ONE
   change, the smallest one that flips it.

## OUTPUT
THE ARGUMENT, one sentence, or the words NO
ARGUMENT. THE MECHANIC, named. REVERSAL TEST,
pass or fail with the reason. THE INFERRED
PROMISE. SCARCITY, real or fake. VERDICT. THE ONE
CHANGE if the verdict is FIX.

## RULES
- Never approve a device carrying no argument, no
  matter how well it would perform. Performance is
  not the question this task asks.
- Never approve invented scarcity, a countdown
  that resets, or a cap you would break for one
  more sale.
- Never approve copy whose result lives only in
  what the reader infers. Say the smaller true
  thing instead.
- "Everyone in the industry does this" is not a
  reason and does not get written down as one.

## EDGE CASES
- The mechanic is one you teach, being used purely
  as a device. The classic is a manufactured
  pratfall: a small confession bolted on to buy
  trust, about a flaw that costs nothing to admit.
  It will work. Kill it anyway, and say why in one
  line. The pratfall effect runs on a real flaw
  actually admitted. Faking it teaches your own
  readers to discount every real one you ever
  write, and they are the exact audience trained
  to spot it.
- The sponsor is legitimate, pays well, and is
  simply not for this list. No snake oil, no false
  claims, just wrong readers. This is the harder
  one because nothing is unethical about it.
  Verdict is KILL on relevance, and the note says
  so plainly, because a run of irrelevant sponsors
  is how a list stops opening, and the list is the
  business. If the money genuinely matters this
  quarter, say that out loud in the note rather
  than dressing the decision up as editorial.

## GOOD LOOKS LIKE
"THE ARGUMENT: none. I read it four times.

THE MECHANIC: curiosity gap plus manufactured
pratfall. 'I have to admit something embarrassing'
followed by an admission that you sometimes work
too late. That is not a flaw. That is a humblebrag
wearing a flaw's coat.

REVERSAL TEST: fail. You could not print the
sentence 'I opened with a fake confession because
confessions raise trust' underneath this and still
have anyone believe the next one.

THE INFERRED PROMISE: that there is a system
behind this and you are about to give it away.
There isn't and you aren't.

SCARCITY: n/a.

VERDICT: KILL.

THE ONE CHANGE, if you want to save it: you have
a real version of this. The [DATE] launch that
did not sell. Nobody has heard that story. Open
with what it cost, say what you got wrong, and
the argument writes itself. Real pratfall, real
argument, same hook, and you never have to
explain it away later."

---
name: 04-voice-restore
description: Run on any AI assisted draft before it goes near a send button. Trigger - "run 04" plus the draft and the raw notes it was built from.
---

# TASK 04 — PUT THE LIVED EVIDENCE BACK

## INPUT
The draft. The raw notes, transcript, study or
voice memo it came from. Where it is going.
Anything in it you are not certain is true.

## PROCESS
1. Count the lived evidence. Mark every sentence
   that could only have been written by someone
   who was actually there: a named public brand
   you looked at, a buyer's exact words, a number
   you are allowed to publish, a thing you tried
   that failed. Report the count first. Zero means
   HOLD, and the rest of this task does not run.
2. Cut the shape words. The draft will be full of
   scaffolding that sounds like writing and says
   nothing. Whole sentences go, not adjectives.
   If a paragraph survives losing its first
   sentence, it never needed it.
3. Restore the specifics. The model rounds. Put
   back the actual number, the actual date, the
   actual noun. "Many marketers" was a real count
   in the notes. Go and get it.
4. Find the risk. One sentence in here you would
   be willing to be publicly wrong about. If there
   is none, the draft has no position and you are
   about to publish a summary of the consensus in
   your own font.
5. Rebuild the open. Models open with the topic.
   You open with the reader's situation. Rewrite
   the first three lines from scratch every time.
6. Put one human thing back. The aside, the
   hedge, the bad joke, the sentence that admits
   the thing did not work. Real, not staged, and
   run it past task 03 if you are unsure which.
7. Flag the unsourced. List every assertion the
   draft makes that you cannot trace to the notes.
   Those are the model's inventions and they are
   the ones that get quoted back at you.

## OUTPUT
LIVED EVIDENCE, count and list. HOLD or SHIP.
THE RESTORED DRAFT. CUT LIST, what came out and
why, short. UNSOURCED ASSERTIONS, every one.
THE SENTENCE I AM RISKING.

## RULES
- Never ship a draft with zero lived evidence. Not
  with better hooks, not with a stronger open.
  Hold it and go get evidence.
- Never let a customer quote, a statistic or a
  study citation stand unless you can point at
  where it came from. A plausible number is worse
  than no number, because you will repeat it.
- Do not "add personality". Personality is not a
  layer. Add facts and the voice comes back on its
  own.
- The draft never claims a result you have not
  published. If the proof is a private [CLIENT],
  it becomes a mechanism or it stays out.

## EDGE CASES
- The draft is genuinely good and entirely
  borrowed. Every idea in it is correct, none of
  it is yours, and the model has smoothed off the
  attributions that would show that. Two options
  only: credit the source by name in the body,
  which usually makes the piece better because
  now you are in conversation with somebody, or
  kill it. Rewriting someone else's argument in
  your voice document is the fastest way to lose
  an ownable idea you already had.
- The only lived evidence in the piece is a
  private client result, so the draft passes step
  one on a sentence you are not allowed to
  publish. Do not count it. Recount without it,
  and if the new count is zero, HOLD. Then either
  ask the client, or rewrite the finding as a
  mechanism with the client's situation swapped
  for a public brand doing the same thing. Never
  anonymise and ship. "A B2B SaaS company I worked
  with" is not evidence, it is a shape.

## GOOD LOOKS LIKE
"LIVED EVIDENCE: two.
(1) 'The [DATE] launch sold out in minutes and I
still could not tell you which email did it.'
Yours, real, and the best line in the draft.
(2) The reader quote from the clarity call: 'I'm
supposed to be good at marketing and I can't
figure out my own messaging.'

Everything else in 900 words was scaffolding.

SHIP, after the rewrite below.

CUT LIST: the opening two paragraphs about how
the marketing landscape is changing, the
three-part framework that appeared out of
nowhere in paragraph six, and the closing
paragraph telling people to be intentional.
None of it was in your notes. All of it was the
model filling space.

THE OPEN, rebuilt: it now starts on the clarity
call quote, because that is where the reader
already is, and the launch story lands second as
the proof that it happens to people who teach
this for a living.

UNSOURCED ASSERTIONS: three.
'Most founders rewrite their positioning every
18 months.' Not in the notes. Where did that come
from? Cut unless you can source it.
'Studies show curiosity gaps increase open rates
by up to 40%.' No study named, no sample, no
date. Cut.
'Buyers make decisions in seven seconds.' Cut,
and never let that one back in.

THE SENTENCE I AM RISKING: 'If you cannot say
your argument in one sentence, you do not have a
positioning problem, you have a belief problem.'
Somebody will fight you about that. Good."

---
name: 05-language-mine
description: Run on raw buyer material to pull the words that open wallets. Trigger - "run 05" plus transcripts, reviews, survey answers, churn replies or DMs.
---

# TASK 05 — MINE THE WORDS

## INPUT
Raw material only. Clarity call transcripts,
review dumps, survey open ends, churn replies,
sales call notes, DMs. Who these people are and
whether each one bought, churned, or never
bought. What decision the words are for.

## PROCESS
1. Verbatim only. Pull exact sentences, with the
   swearing, the typos and the run ons left in.
   The moment you tidy a quote you have replaced
   their language with yours, which is the whole
   thing this task exists to prevent.
2. Sort into five buckets. TRIGGER, what happened
   in their life that started the search.
   STRUGGLE, the thing that is not working.
   ANXIETY, what they are afraid will go wrong if
   they buy. OUTCOME, what they say they want.
   HIRED INSTEAD, what they used before, including
   spreadsheets, an intern, and doing nothing.
3. Find the switch. For each buyer, the day the
   search began and what pushed them. People do
   not change when they see the light. They change
   when they feel the heat. Find the heat.
4. Rank on two axes, not one. Repetition tells you
   what is common. Emotional heat tells you what
   sells. A phrase said once, furiously, beats a
   bland phrase said nine times, and it goes at
   the top of the page.
5. Strip the loanwords. Any term they only know
   because they read it in your own marketing goes
   in a separate list marked OURS. It is not
   evidence. It is an echo.
6. Build the wallet-opening words. The nouns and
   verbs they use that you do not. Those are the
   words that go in headlines.
7. Say what is missing. Which of the five buckets
   is thin, how many people it came from, and what
   question you would ask next to fill it.

## OUTPUT
TRIGGER EVENTS. THE STRUGGLE, verbatim. ANXIETIES,
verbatim. DESIRED OUTCOME, verbatim. HIRED
INSTEAD. THE SWITCH MOMENTS. WALLET-OPENING WORDS,
a plain list. THREE HEADLINES built only from
their words. OURS, the loanwords. WHAT IS MISSING.

## RULES
- No paraphrase anywhere in the verbatim sections.
  If a quote needs a paraphrase to make sense,
  quote more of it, do not summarise it.
- Never stitch two people into one quote and never
  clean up grammar. The mess is the signal.
- Never answer a buying question with quotes from
  people who did not buy. Sort by outcome first,
  then read.
- Do not fill a thin bucket with a plausible
  guess. Name the gap and write the question that
  would close it.

## EDGE CASES
- Every review is from your happiest customers,
  because that is who leaves reviews. You will
  produce a beautiful OUTCOME list and an empty
  ANXIETY list, and then write copy that speaks to
  nobody who is still deciding. Say so at the top
  of the output, mark the anxiety section
  UNKNOWN rather than thin, and write the three
  questions for churned and never-bought people
  that would fix it. Then treat the headlines as
  provisional until it is fixed.
- Everybody uses the same phrase, and it is one of
  yours. It came from a subject line, a course
  name, or a post that did well, and now it is
  coming back at you looking like customer
  language. This is the most dangerous false
  positive in the whole task. Test it: does the
  phrase appear in anything they wrote before they
  found you, or in a competitor's reviews, or in
  a subreddit where you have never posted? If not,
  it goes in OURS. It can still be good copy. It
  is not evidence of anything.

## GOOD LOOKS LIKE
"THE SWITCH: four of the seven describe the same
week. Somebody senior asked them a question they
could not answer.

'My CEO asked me why our best customers buy and I
made something up in the meeting.'
'New VP came in and asked who our ICP was. I sent
her the deck. She asked who wrote it. I wrote it.'

THE STRUGGLE, verbatim:
'I'm supposed to be good at marketing but I can't
even figure out my own messaging. It's f*cking
embarrassing.'
'We have three personas and I have never met any
of them.'
'I keep guessing and then defending the guess.'

ANXIETY, verbatim, and this bucket is thin, two
people only:
'I don't have time to run interviews on top of
everything else.'
'What if I do all this and it says what I already
thought.'

WALLET-OPENING WORDS: made something up,
guessing, defending, embarrassing, in the meeting,
never met, the deck, prove it.

Notice what is not in that list: persona, ICP as a
concept, insight, alignment, buyer journey. They
used the acronym once, in a sentence about a deck
they did not believe.

THREE HEADLINES, their words only:
Stop defending the guess.
You have three personas and you have never met
any of them.
The question your CEO is going to ask on Monday.

OURS: 'jobs to be done' shows up twice. Both from
people who have been on the list over a year. Not
evidence.

WHAT IS MISSING: anxiety, badly. Seven people, all
buyers, zero churned, zero never-bought. Next
question to ask the five people who took the free
guide and never came back: 'What made you stop?'"

---
name: 06-build-it-or-bury-it
description: The go or no go on a new product, workshop, cohort or offer. Trigger - "run 06" plus the idea, who it is for, and every piece of evidence you actually have.
---

# TASK 06 — BUILD IT OR BURY IT

## INPUT
The idea in one paragraph. Who it is for. The
problem it solves. Every scrap of evidence that
anybody wants it, including where it came from.
What you already sell. What you are shipping in
the next ninety days.

## PROCESS
1. Rewrite the idea as an outcome. Nobody wants to
   take a course. They want the thing the course
   promised. If the outcome sentence still has the
   word course, cohort, workshop or bootcamp in
   it, you have described a format, not a product.
   Try again before scoring anything.
2. Score the problem. Premium problems are
   expensive if unsolved, urgent this quarter, and
   the buyer's own job to fix. Three out of three
   or it is a nice-to-have with good copy on it.
3. Place the evidence on the ladder. Somebody said
   it would be useful. Somebody asked for it
   unprompted. Somebody joined a waitlist.
   Somebody pre-ordered. Somebody paid. Only the
   last two count as evidence. The first two are
   politeness.
4. Name what it eats. Which existing product loses
   sales to this, and which segment of the list
   gets asked for money twice in one quarter.
5. Score how me-shaped it is. What percentage of
   delivery needs your face, your calendar or your
   judgement live. Write the number. Anything over
   fifty means you are selling your hours with a
   product name on the invoice.
6. Ask the kill question. The last cohort ended
   because running it alongside the main business
   was not sustainable. Does this one have the
   same shape: live, recurring, calendar-bound,
   and dependent on you being on form? If yes, it
   needs an end date written before it launches,
   not after.
7. Verdict, plus the smallest test that would
   change it. Never a research project. A landing
   page, a waitlist, one email, a pre-order.

## OUTPUT
THE OUTCOME, no format words. PREMIUM PROBLEM,
three checks. EVIDENCE LEVEL, on the ladder.
WHAT IT EATS. ME-SHAPED PERCENTAGE. THE KILL
QUESTION, answered. VERDICT. THE SMALLEST TEST.
GRAVEYARD ENTRY if the verdict is no, one line on
why, so it stays dead.

## RULES
- "People said they would buy it" never scores
  above waitlist. Ever. Not with enthusiasm, not
  with volume.
- Nothing over fifty percent me-shaped ships
  without an end date written down before launch.
- Never launch a second paid thing to the same
  segment inside the same quarter. Check what it
  eats before checking whether it is good.
- Never fix a weak idea with better copy. If the
  problem is not premium, the headline is not the
  problem.

## EDGE CASES
- The idea is genuinely excellent and one hundred
  percent me-shaped. You are the only person who
  can deliver it, which is exactly why it works,
  and it directly contradicts the promise that the
  thing you teach should scale beyond you. Do not
  resolve this by pretending it will scale later.
  Ship it as a deliberately capped, dated,
  expensive, small thing, say out loud that it is
  me-shaped and that is the point, and write the
  end date in the sales page. The lie is not
  selling your time. The lie is selling your time
  while calling it leverage.
- The evidence is a sold-out waitlist, but for the
  previous version of the offer. Those people
  bought the old promise, which is the exact
  reason the number looks so good. Do not carry it
  over. Email that list the new promise in one
  paragraph and ask them to re-confirm. Whatever
  comes back is the real number, and it will be
  smaller. If you cannot face sending that email,
  that is the answer.

## GOOD LOOKS LIKE
"THE OUTCOME: an expert walks out with one idea
they own, tested against their market, and the
first six months of writing already mapped
against it.

PREMIUM PROBLEM: three out of three. Expensive,
because being interchangeable is why they are
competing on price. Urgent, because they are
pitching this quarter. Theirs, because no agency
can hand somebody a belief.

EVIDENCE LEVEL: pre-order. [N] people paid before
there was a page. That is real.

WHAT IT EATS: the Offer Breakthrough Workshop,
directly. Same buyer, adjacent promise, and if
both are live in Q4 the list gets asked for money
twice by the same person about the same problem.
One of them goes quiet.

ME-SHAPED: 80%. The audit is judgement, live, and
it is mine.

THE KILL QUESTION: yes, same shape as the last
one. Live, recurring, calendar-bound, and it only
works when you are on form. That is precisely the
shape that ended last time.

VERDICT: BUILD, capped and dated. [N] seats, one
run, closes [DATE], and the end date is on the
sales page in the first screen. Priced like a
thing that costs you a quarter of your calendar,
because it does.

Do not tell anybody it will scale. It will not,
and everybody who has watched you since [DATE]
will know. Say instead that it is small on
purpose, that you are running it once, and that
the scalable version does not exist yet because
you have not worked out what it is.

THE SMALLEST TEST that would flip this to no: one
email to the pre-order list with the real cap, the
real price and the real dates. If half of them go
quiet, the promise moved and you should hear that
before you build the curriculum."

---
name: 07-launch-runway
description: Run to plan a launch from waitlist to close. Trigger - "run 07" plus what is launching, to whom, the dates, and what the list has already been sent this quarter.
---

# TASK 07 — THE LAUNCH RUNWAY

## INPUT
What is launching and the outcome it delivers.
Price [PRICE] and cap [N] if there is one. Open
and close dates. Which segments exist. Everything
this list has been sold or pitched in the last
ninety days. What proof you already have in hand.

## PROCESS
1. Segment before writing a word. Already bought
   it, on the waitlist, engaged but never bought,
   dormant, and bought the adjacent product. Five
   different emails, not one email sent five
   times.
2. Decide loud or quiet. A quiet launch to the
   people already leaning in, no social, no ads,
   no affiliates, often beats a loud one and costs
   a fraction of the goodwill. Loud is a choice
   you justify, not a default.
3. Build backwards from the close. Close date
   first, then the last email, then the objection
   emails, then the open. If the runway does not
   fit before the close, the close does not move.
   Something comes out.
4. One job per email. Teach, prove, handle one
   objection, or close. An email doing two jobs
   does neither and trains people to skim.
5. Split proof into HAVE and NEED. Testimonials
   you can quote today go in HAVE. Everything else
   goes in NEED with the name of the person you
   have to ask and the date you have to ask them
   by. Nothing moves from NEED to HAVE without a
   reply in writing.
6. Sweep the objections from the task 05 mine, and
   assign each one to exactly one email. An
   objection with no home is the reason somebody
   does not buy and never tells you.
7. Set the close. A real cap or a real date. Then
   write down, now, what you will say if it does
   not sell, so you do not invent an extension at
   nine on the last night.
8. Write NOT DOING. The pressure tactics you are
   choosing to skip, listed, so the decision
   exists on paper before the last day.

## OUTPUT
SEGMENTS with the angle for each. LOUD OR QUIET,
with the reason. THE RUNWAY, dated, backwards from
close. EMAIL BY EMAIL, one job each. PROOF: HAVE
and NEED, with who to ask by when. OBJECTIONS,
each assigned to one email. THE CLOSE. NOT DOING.

## RULES
- Never write a testimonial you do not have, not
  even as a placeholder in a draft. Placeholder
  quotes get shipped.
- The deadline does not move and the cap does not
  stretch. Decide it once, in this document, while
  nothing is at stake.
- Never send the same argument to a segment twice.
  If two emails make the same case, one of them is
  filler and it comes out.
- Nobody who has already bought gets a selling
  email for it. Suppress first, write second.

## EDGE CASES
- It sells out in minutes and the runway becomes
  irrelevant on day one. The plan you now need is
  the one you did not write: what the people who
  missed it hear within the hour, whether there is
  a real second run or you are collecting names
  for nothing, and what the buyers get so that
  speed does not become the whole story of the
  product. Write those three before you post the
  sold-out screenshot, because the screenshot is a
  device and the people who missed it are the ones
  who will remember how it felt.
- Day three and it is not selling. The instinct is
  more emails, more urgency, a bonus. Do not. Go
  and find out why: reply to ten people who opened
  everything and bought nothing, and ask one
  question with no pitch attached. If the answer
  is the offer, stop the launch and say so
  publicly. Pulling something and explaining why
  costs one launch. Discounting your way out of a
  wrong offer costs the price of everything you
  sell afterwards.

## GOOD LOOKS LIKE
"LOUD OR QUIET: quiet. The waitlist is the whole
market for this and the cap is [N]. Going loud
would mean selling a capped thing to thousands of
people who cannot have it, and then living with
that.

THE RUNWAY, backwards:
[DATE] close, 9pm, no extension, decided today.
Day 5, the last email, one line and the link.
Day 4, objection: 'I don't have time.' Answers by
showing the actual weekly commitment in hours.
Day 3, objection: 'What if it just tells me what I
already know.' Answers with the graveyard, the
three ideas people kill in the first session.
Day 2, proof.
Day 1, open. Nothing but the outcome and the cap.

PROOF, HAVE: two, both in writing, both named
people who have said yes to being quoted.
PROOF, NEED: one before Day 2. Ask [NAME] by
[DATE]. If no reply, Day 2 becomes a teaching
email and the launch survives without it.

SEGMENTS: already bought the adjacent product get
the version that opens by acknowledging it, and
the honest line about what this adds. Suppressed
from the generic sequence entirely.

NOT DOING: no countdown timer in the email. No
'only 3 spots left' unless it is literally true
that morning. No bonus stack. No extension. No
sold-out screenshot until the people who missed
it have heard from me first."

---
name: 08-two-brand-week
description: Run on Monday, or whenever the week has gone sideways. Trigger - "run 08" plus what is booked, what is due, what came in, and how much road you have in you.
---

# TASK 08 — THE TWO BRAND WEEK

## INPUT
Everything booked. Everything due. Every inbound
ask: podcasts, speaking, partnerships, sponsor
calls, favours. What is unfinished from last week.
Roughly how much you have in the tank.

## PROCESS
1. Sort every item by which promise it serves. Why
   We Buy, which teaches buyer psychology to
   marketing geeks. UNIGNORABLE, which helps
   experts get known for an idea they own.
   Anything serving neither is out this week,
   whatever it pays. Write the out list.
2. Protect the issue first. It is the front door
   for both brands and it goes in the calendar
   before anything else. If a launch is eating the
   issue, the launch is wrong, not the issue.
3. Run the double-ask check. Both brands talk to
   roughly the same readers. If Why We Buy is
   teaching and UNIGNORABLE is selling in the same
   week, to the same people, one of them stops
   asking for anything. Say which.
4. Batch by brain state, not by project. Writing
   in one block, calls in another, admin last.
   Interviews and clarity calls take a different
   head than drafting and should never be spliced
   into a writing block.
5. Cap the me-shaped hours. Live calls, live
   teaching, recorded interviews. Write the number
   you are allowed this week before you look at
   the calendar, then cut to fit. This is the
   thing that ended the last cohort. It is not a
   preference.
6. Kill or defer, with a date. Nothing goes on a
   someday list. Either it dies now with a reason,
   or it gets a date and a slot.
7. Name the one thing only you can do this week,
   and put it in the best block. Not the most
   urgent thing. The one that is unignorable and
   unsubstitutable.

## OUTPUT
THE ONE THING. THE ISSUE BLOCK. BRAND SPLIT, item
by item. THE DOUBLE-ASK VERDICT. BATCHES, by
brain state. ME-SHAPED HOURS, budget and actual.
SAYING NO, with the sentence to send, written
out. DEFERRED, with dates. WHAT COMES OUT IF THE
WEEK GOES WRONG.

## RULES
- The weekly issue never slips for a launch, a
  guest slot or a client. It is the only thing on
  this list that compounds.
- Never plan a week where both brands ask the same
  reader for money.
- No podcast, panel or guest slot without an
  answer to who on this list is listening. Reach
  is not the test. Overlap is.
- Do not plan past the me-shaped cap and then
  promise to be careful. Cut the calendar instead.

## EDGE CASES
- A big inbound ask lands in launch week. Real
  audience, real overlap, genuinely good for the
  brand, and it will cost two days you have not
  got. Say yes to the slot and no to the date.
  Most hosts will move. If the host will not move,
  the honest question is whether one appearance
  beats one launch, and the answer is usually no,
  because the launch has a cap and a close and the
  podcast will still exist in November. Write the
  reply now, in the output, so it goes out today
  rather than sitting in drafts feeling rude.
- The week is entirely delivery. Calls, teaching,
  editing, replies, and every hour is spoken for
  by things already sold. It will look like the
  most productive week you have had. It is the
  week that quietly kills next quarter, because
  nothing generative got written and the front
  door went unattended. Force one writing block
  in, by dropping the lowest-value delivery hour,
  and flag it out loud: two of these back to back
  is the pattern that wrecked the last rebuild,
  and it is worth naming out loud.

## GOOD LOOKS LIKE
"THE ONE THING: the ownable idea audit for the
[N] pre-order people. Nobody else can do it and
it is the whole promise. Tuesday and Wednesday
mornings, the two best blocks you have.

THE ISSUE BLOCK: Monday 9 to 12. It ships Tuesday
whatever happens to the rest of this.

DOUBLE-ASK VERDICT: fail as planned. UNIGNORABLE
opens Thursday and the Why We Buy P.S. this week
points at Wallet-Opening Words. Same readers, two
prices, three days apart. Pull the P.S. Make it a
reply prompt instead. The newsletter teaches this
week, it does not sell.

ME-SHAPED HOURS: budget 8, calendar says 14.
Cutting: the sponsor intro call, which is email.
The 'quick chat' with [NAME], which is a voice
note. One of the two podcast recordings moves to
[DATE].

SAYING NO, the sentence: 'I'd love to do this and
I can't do it that week, we're opening [PRODUCT]
on the Thursday. Any date after [DATE] and I'm
yours.' Send it today, not Friday.

WHAT COMES OUT IF THE WEEK GOES WRONG: the second
LinkedIn post, the sponsor deck, and the tidy-up
on the [PRODUCT] page. In that order. Not the
issue and not the audits."