The Formation of the Master Mind

The partnership lives in the setup rather than in each prompt: your standing context - who you are, what you are building, how you want to be argued with - belongs in Custom Instructions and saved memory, so every conversation starts with the junior partner already briefed. For a collaboration you return to, a Custom GPT turns the whole arrangement into something you can name and reopen. In the prompt itself, state the aim first (Chapter 1), name both roles (Chapter 4), and say explicitly what you are contributing so your hypothesis is treated as evidence to test rather than a conclusion to flatter.

• Standing context in Custom Instructions; the task in the prompt.
• Name both roles and say what you are contributing.
• Ask for disagreement in its own section, or you get agreement.
• A recurring partnership belongs in a Custom GPT.
• "Great hypothesis" means the framing failed. Push back.

**This is the capstone, so it is worth naming what is being combined.** The Definite Chief Aim from Chapter 1 gives it a single goal; the Master Persona from Chapter 4 names both roles; the Treasure of Specificity from Chapter 24 replaces "a fashion brand" with a real one; and the Master's Review from Chapter 21 is why the last section exists.

---

Let us set up a working partnership for this project, and keep it for the whole conversation.

## THE ROLES

**Mine - senior partner.** I own the strategy, the risk and the final call. I know this brand, this market and what I can afford to be wrong about.

**Yours - junior partner with range.** You have read more marketing campaigns than I ever will. Your job is to widen my options and test my reasoning, not to agree with it.

## MY AIM

A marketing plan for a small sustainable fashion brand in India. Specifically: [brand name], selling [product] at [price range] to [who], currently reaching customers through [channel]. The constraint that shapes everything is [budget, team size or inventory reality].

One outcome matters: [the single thing - first thousand customers, breaking even on paid acquisition, a wholesale relationship]. Anything that does not move that is interesting rather than useful.

## TO START

Five of the most innovative marketing campaigns of the last two years. For each: what they did, and the core principle that made it work - stated as a principle I could apply, not as a description of what happened.

Then, because a list of five admired campaigns is where most strategy conversations go wrong:

1. **Which of the five could my brand actually run** at my budget and scale, and which needed money, reach or luck I do not have? Be blunt about this. A campaign that worked because a national brand spent heavily is not a template for me.
2. **Which transfer to Indian sustainable fashion specifically**, and which do not travel - because of the market, the price sensitivity, or because sustainability claims are received differently here?
3. **What am I likely to pick for the wrong reasons?** Name the one I will find most appealing and tell me why it is the wrong choice for me.

## HOW I WANT YOU TO WORK WITH ME

Disagree with me in writing, not in hedges. When I state something as fact that is actually my assumption, say "that is an assumption and here is what would make it false". If you find yourself calling my thinking strong, check whether you have examined it - a partnership where the junior partner agrees is just a more expensive search engine.

**Set this up properly once:** put the brand, the market, the constraint and that last paragraph into Custom Instructions, so you do not rebuild the briefing every session - and if you come back to this often, make it a Custom GPT called "[Brand] strategy partner".

**What is combined here:** the Blueprint of Creation from Chapter 2 structures the output; the Search for Absolute Truth from Chapter 13 governs how the data is handled; the Reversal of Perspective from Chapter 23 is why the last question exists - and Controlled Imagination from Chapter 18 is the reason the analysis is fenced off from the interpretation.

---

I need help with a problem that genuinely needs both halves - my knowledge of this company and your patience with fifty documents.

## THE DIVISION OF LABOUR

**What I bring - context and a hypothesis.** My instinct is that our employee turnover is high because of poor middle management, not because of salary. I have been here long enough to have a view, and long enough to be attached to it.

**What you bring - analysis without my investment in the answer.** You have no stake in my being right, and you can read fifty exit interviews without getting tired on the thirtieth.

## THE DATA

Anonymised exit interview text from our last 50 departures:

```
[Paste data here]
```

## TASKS - IN THIS ORDER, AND THE ORDER MATTERS

1. **Analyse first, blind to my hypothesis.** Identify the top three most frequently mentioned reasons for leaving, with how many interviews mention each and representative quotes. Do this as if I had told you nothing. I have deliberately put my hypothesis above the data and I want you to work from the data anyway - that sequence is the whole point of this exercise.
2. **Then test my hypothesis against what you found.** Does the data support it, contradict it, or fail to address it? Those are three different answers and the third is the most likely. Quote the lines that bear on it either way.
3. **Then the uncomfortable part.** What is in this data that I did not ask about and would not have looked for? A hypothesis makes you blind in a specific direction, and you are better placed to see what I have excluded than I am.

## HOW TO HANDLE THE LIMITS OF THIS DATA

Be straight about these rather than burying them:

- Exit interviews are a filtered source. People leaving rarely criticise a manager by name, and "better opportunity" often means something else. Where you suspect the stated reason is a polite substitute, say so and show the line that made you think it - but mark it as inference, clearly, not as a finding.
- Fifty is enough to see a pattern and not enough to settle an argument. If a difference is small, say it is small.
- Do not infer anything about individual people's circumstances, health or motives beyond what the text says.

## AND THEN

Tell me what I should actually do to find out - what question to ask, of whom, to test this properly. The exit interviews are evidence I already have; a real answer probably needs the people who have not left.

If the data supports me, say so plainly. But check whether you are agreeing because the evidence points that way or because I told you my view before you read it.

**The fullest combination in the book, and worth naming.** Definite Chief Aim (Ch 1) sets the single goal. Master Persona (Ch 4) names the roles. Organized Knowledge (Ch 3) supplies the story only I have. Specificity (Ch 24) fixes the audience and the room. Commence with the End in Mind (Ch 10) decides the closing line before the opening. The Chain of Thought (Ch 22) is why this happens in stages. The Master's Review (Ch 21) is the final section. And the Negative Command (Ch 15) is the list of what must not appear.

---

We are writing a keynote together. I will be the one standing up, so the vision is mine and the craft is yours.

## MY ROLE - THE VISION

**Audience:** a room of roughly [number] young entrepreneurs in Hyderabad, mostly [stage - pre-revenue, first customers, first hires]. Many of them are in the hardest part and are not saying so.

**Core message:** persistence matters more than genius.

**What I want them to do differently on Monday:** [the specific change - make the call they have been avoiding, stop waiting for the idea to be perfect, go back to the thing that failed]. This is the test for every line in the speech.

**The story only I have:** [your actual failure - what you tried, what it cost, the specific moment you nearly stopped, what you did instead]. Do not smooth this. The useful part of a failure story is the part that is still embarrassing, and a room full of people currently failing can tell the difference immediately.

**Length:** [minutes]. Roughly [minutes x 130] words.

## YOUR ROLE - THE CRAFT

Structure, rhythm, the shape of sentences meant to be heard rather than read. Where a paragraph needs to become three short lines. Where a pause does more than a word.

## HOW WE DO THIS - IN STAGES, NOT ONE DRAFT

**Stage 1.** Before writing anything, ask me the five questions you need answered to write this well - about my story, the room, or what I actually believe. Five that change the speech, not a form to fill in. Then give me the closing line first: the last sentence they walk out carrying. We decide the ending before the beginning, because everything else is the road to it.

**Stage 2.** The structure - opening, three movements supporting the core message, conclusion. One line per section on what it must land. I approve this before any prose exists.

**Stage 3.** The full draft, in Canvas so I can work on it line by line rather than regenerating the whole thing to fix one paragraph.

## WHAT MUST NOT BE IN IT

- "Failure is a stepping stone to success", "follow your passion", "the only way to do great work is to love what you do", and anything else this audience has heard at five previous events.
- Any statistic I have not given you. If a number would strengthen a point, tell me to find one rather than supplying it - an invented figure from a stage is the one mistake I cannot take back.
- Quotations you are not certain of. One or two real quotes on persistence are welcome, and a misattributed quote in a speech about honesty is worse than none. Mark every quote as "verify this" and give me the source you believe it comes from, so I can check it before Thursday.

## THEN REVIEW YOUR OWN DRAFT

A short section headed "Where this is weak":

- Which of the three movements is thinnest.
- Any line that sounds good on the page and will land flat spoken aloud. Those are different things and this is the failure mode of written-out speeches.
- Where you have used my story as decoration rather than as argument.
- Every claim or quote I must verify before I stand up.

Be honest in that section. You can tell me the structure sags in the middle; you cannot tell me whether a quote is real, so put every one of those in the verify list rather than reporting the draft as clean.