How to write a company initiation

Prompt

Add this as a memory called 'Initiation Writing Process'

Initiation note workflow

Core principle

Do not start by writing. Start by solving the investment debate.

A good initiation note is not a company description. It is a proof chain:

> What is the market missing? What has changed? What must be true? What evidence supports it? How does it flow into the model? Why does valuation follow?

The note should only be written once that chain is clear.

How to use this memory

When helping me with an initiation note, use this as the default workflow:

1. Orient first, without forcing an opinion.
2. Build the financial base early.
3. Identify the old / stale market view.
4. Reduce the note to one big debate.
5. Work out the economic mechanism.
6. Do company-specific diligence based on that debate.
7. Translate strategy into unit economics.
8. Build base, upside and downside cases.
9. Choose valuation after the thesis is clear.
10. Write only once the proof chain is solved.

If any section can be written without having worked anything out, treat it as likely filler.

Interaction protocol — gated, step-by-step

When I ask Primer to run this initiation workflow, it must be run as a gated sequence, not as an autonomous end-to-end project.

Default behaviour:

1. Do one step only. Complete the next step in the sequence and stop.
2. Return a step output. Summarise what was learned, the evidence used, the provisional implications, and any open questions.
3. Make a clear recommendation. Say whether the step is good enough to move on, needs more work, or has revealed a problem.
4. Ask for permission to proceed. End by asking whether to continue to the next numbered step.
5. Wait. Do not start the next step until I explicitly say to continue / yes / go ahead / next step.

This gating instruction overrides the normal tendency to work through all ten steps independently. If the user asks for a full initiation note, still begin with Step 1: orientation and pause after that step unless they explicitly instruct otherwise.

Use this status format at the end of each step:

- Step completed: [number and name]
- Recommendation: [move on / deepen this step / rethink]
- Next proposed step: [number and name]
- Question for user: Shall I continue to the next step?

1. Start with orientation, not opinion

First build the plain factual map:

- What the company sells
- Who the customer is
- Main divisions, geographies and channels
- Revenue streams
- Cost structure
- Key KPIs
- Recent history
- What has gone right or wrong
- What investors probably already think

At this stage, do not try to force the final stock view. Understand the company first.

2. Build the financial base early

Before deciding the angle, understand the financial engine:

- Revenue by segment
- Gross margin
- Operating costs
- EBITDA, EBIT and PBT
- Cash conversion
- Capex
- Working capital
- Leases
- Tax and interest
- Net debt or net cash
- Dividend or buyback capacity
- Forecast sensitivity

The model often tells you what the note is really about. Sometimes the story is growth. Sometimes it is margin recovery. Sometimes it is cash flow. Sometimes it is balance sheet repair.

3. Identify the old market view

Define the stale perception the note is written against:

- What do investors currently believe?
- What historical issue are they anchored on?
- Is that view still fair?
- What has changed?
- What evidence would force investors to update their view?

The initiation should not just say, “this is a good company.” It should say, “the market is using the wrong mental model.”

4. Define the big debate

Reduce the note to one central question, such as:

- Is margin improvement structural or temporary?
- Can the company grow beyond its core market?
- Has a recovery story become a cash-return story?
- Is the balance sheet now an asset rather than a constraint?
- Can the business drive frequency in a low-frequency category?
- Is the market undervaluing cash generation?
- Is the growth opportunity real or just narrative?

Do not begin the full write-up until this debate is clear.

5. Work out the economic mechanism

The debate needs a mechanism. Avoid vague phrases like “self-help”, “brand strength”, “loyalty” or “growth opportunity” unless the money-making route is explicit.

Map the machine:

> Customer behaviour → revenue driver → margin effect → cash flow → valuation impact

Example:

> Higher retention → lower acquisition cost → better contribution margin → higher cash generation → higher justified multiple

The mechanism tells you what evidence is needed.

6. Do company-specific diligence

Do not use the same checklist for every company. Let the debate decide the work.

If the thesis is growth, test:

- Market size
- Share gains
- Customer acquisition
- Repeat purchase
- Pricing power
- Channel expansion
- Product adoption
- Competitor response

If the thesis is margin recovery, test:

- Gross margin bridge
- Cost savings
- Operating leverage
- Mix shift
- Input costs
- Pricing
- Fixed versus variable cost base

If the thesis is cash return, test:

- FCF bridge
- Capex needs
- Working capital volatility
- Balance sheet headroom
- Dividend capacity
- Buyback capacity
- Pension or lease drag

If the thesis is customer behaviour, test:

- Frequency
- Retention
- Average order value
- Lifetime value
- Attach rates
- Subscription or membership economics
- Price comparison
- Customer satisfaction

7. Translate strategy into unit economics

Do not just say an initiative is attractive. Work out what it is worth.

Find the smallest useful economic unit:

- Per customer
- Per order
- Per store
- Per subscriber
- Per member
- Per product
- Per site
- Per contract
- Per pound of buyback

Then ask:

- What is the revenue?
- What is the gross profit?
- What is the incremental cost?
- What is the acquisition cost?
- What is the repeat rate?
- What is the renewal rate?
- What is the contribution?
- What is the payback?

If the unit economics do not work, the thesis probably does not work.

8. Build base, upside and downside cases

Separate what is underwritten from what is optionality.

Case	Purpose
Base case	What happens with reasonable execution and no heroic assumptions
Upside case	What happens if the mechanism works better than expected
Downside case	What breaks the thesis

The Buy case should not depend entirely on the boldest assumption. The note should make clear what is already in the forecasts and what is not yet priced in.

9. Choose valuation after the thesis is clear

Valuation should follow the debate.

Thesis type	Useful valuation lens
Long-term growth / margin optionality	Scenario DCF
Cash generation / shareholder returns	FCF yield, DCF, buyback analysis
Recovery / earnings normalisation	Normalised P/E or EV/EBITDA
Multi-division business	Sum-of-the-parts
Asset-heavy business	NAV or replacement value
Early-stage growth	Unit economics and long-term scenario analysis

Do not force a standard valuation method onto every company. Use the method that best captures the investment debate.

10. Start writing only when the proof chain is solved

Writing should begin only when these questions are clear:

- What is the call?
- Why now?
- What is the market missing?
- What is the central debate?
- What evidence changes the debate?
- What are the key model drivers?
- What is in the base case?
- What is upside optionality?
- What breaks the thesis?
- Why does valuation follow?

If those answers are not clear, keep working.

Recommended note structure

1. Front-page thesis — recommendation, target price, upside/downside, and the one-sentence investment case.
2. Why now — what has changed and why the market view may be stale.
3. The big debate — the central investor question.
4. Mechanism and evidence — how the thesis actually works.
5. Business foundations — why the base business is resilient enough.
6. Market context — only the market data needed to answer the debate.
7. Forecasts and scenarios — base, upside, downside and key sensitivities.
8. Valuation — method, assumptions and link to the proof chain.
9. Risks — the specific things that would break the thesis.

Final test

Before writing any section, ask:

> What had to be worked out before this section could be written?

If the answer is “nothing”, the section is probably filler.

The right sequence is:

1. Understand the company.
2. Build the financial base.
3. Identify the stale market view.
4. Define the big debate.
5. Work out the economic mechanism.
6. Gather company-specific evidence.
7. Convert the mechanism into unit economics.
8. Build base, upside and downside cases.
9. Choose valuation.
10. Then write.

The note should feel inevitable by the time it is written.