> ## Documentation Index
> Fetch the complete documentation index at: https://docs.snapmarket.co/llms.txt
> Use this file to discover all available pages before exploring further.
# The thesis
> A one-page argument for why the partnership should exist, written as a claim that can be wrong.
A thesis is a one-page argument for why this partnership should exist, and it is written as a claim that can be wrong. That last property is the entire point. A claim that can be wrong has to name what would prove it wrong, and naming that is what turns a document into an anchor.
What it gets you is a marketing budget you can defend in a board meeting, and a conversation in month nine that is specific rather than vague.
Everything downstream inherits its argument from this page. The proposal prices work. The messaging document defines language. The go-to-market plan sequences activity. The monthly audit reports movement. None of them say why any of it should be happening at all. The thesis does, and it becomes the standard against which the rest gets measured.
## Why one page
The limit is a quality mechanism rather than a formatting preference. At that length there is no room to waffle on the buyer or hedge the math, so it either lands on a finding or it says nothing. No room for a long list of unknowns, so only the ones that actually block the claim survive. The constraint does the editorial work, and a thesis that needs two pages is a thesis that has not been decided yet.
## What is on the page
Six sections, in this order, because each depends on the one before it.
Where your revenue comes from today, with numbers. The machine as it currently runs, not the company history.
One kind of person, described by what they believe, and what they are shopping for regardless of who sells it to them.
Why that buyer does not currently arrive. This is the problem being purchased, stated as one mechanism rather than a list of weaknesses.
Where we reach them, in what form, at what cadence, expressed in § coordinates. This is what makes the scope conversation legible a year later.
Derived, not asserted. We start at the number you care about and work backwards to the number marketing has to produce.
Your revenue goal divided by what a customer is worth gives the customers you need. That number divided by your close rate gives the conversations. That number divided by the share of people who raise a hand gives the prospects you have to reach.
Every step divides, so every number grows as we move back up the funnel. The last one is where the marketing starts.
What must be true, who owns which part, and one sentence beginning “This thesis is wrong if.” We do not control your price, your product, your close rate, or who answers the phone when a lead comes in, and naming that before anyone is disappointed is what makes month nine a specific conversation—the reach happened, the conversations happened, the close rate did not—rather than a vague argument about whether marketing is working.
## The math, worked
Abstractly it is four nouns in a row. With your numbers in it, it is a plan.
Say you want \$1 million of new revenue this year. A customer is worth \$50,000. You close one first conversation in four, and two people in every hundred who encounter your work raise a hand.
| What we know | The division | What it gives |
| ---------------------------------- | ------------------ | ---------------- |
| $1M goal, customers worth $50,000 | 1,000,000 ÷ 50,000 | 20 customers |
| You close one conversation in four | 20 ÷ 25% | 80 conversations |
| Two in a hundred raise a hand | 80 ÷ 2% | 4,000 prospects |
Four thousand is the number that sizes everything else—how much gets published, on what cadence, for how long. It is not a target we invented. It fell out of your own revenue goal, your own customer value, and your own close rate.
This is also where partnerships get decided, and not because of anything we say. Watching your own goal divide down into a reach number you have never hit is a different experience than being handed a proposal. Sometimes it means the plan is right. Sometimes it means the goal is wrong, or the price is, or the close rate is.
## What it is not
It has no deliverable table, no calendar, no price, and no tactics list. The moment it has a deliverable table it is a proposal, and the moment it has a calendar it is a go-to-market plan. Both exist downstream and both are better at their jobs.
It is also not a diagnosis with a prescription attached. The thesis names what is broken. It does not explain how to fix it. That is a deliberate line, and it is what keeps the second conversation an argument about your business rather than a free consulting session.
## The three states
The thesis is a manuscript in miniature. It has a draft, a revision, and a final form.
| State | When | What changes |
| -------- | ------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------- |
| Draft | Written from the discovery conversation, delivered in person | Underwritten by judgment, and it says so on the page |
| Revised | At the Storybook artifact | The buyer and gap sections get rewritten from the messaging interviews, because that is where we learn the first draft was wrong. |
| Standing | When Scale’s baseline period closes | The math finally carries your real numbers, and every audit measures against it. |
At the Epilogue the thesis is rendered as confirmed, disconfirmed, or amended, with the full stack of revisions behind it. That stack is the intellectual history of the partnership, and it is what makes a handoff feel like a finished book rather than a folder of files.
We expect the first draft to be wrong about the buyer. The revision at the Storybook artifact is the plan, not a concession. Being corrected out loud is the mechanism.
## The monthly thesis review
Once a month we put the thesis on the table and ask whether it’s still true.
The data is reviewed every month and we act on it every month. The strategy itself is not on a schedule. It can hold for years, and often should, because a plan revised every thirty days never runs long enough to tell you whether it worked. What changes a strategy is evidence, not the calendar.
This is the recurring session of a partnership and it is not a status meeting. The numbers arrive beforehand in your [performance report](/system/analytics/performance-reporting). The review is what we make of them, measured against the claim we opened with.
Every chapter of your manuscript gets one of four words, used consistently so that “steady” means the same thing in September as it did in March.
| Status | What it means |
| --------------- | ---------------------------------------------------------------------- |
| Strong | Performing at or above benchmark. Protect it and do more of it |
| Steady | Working, but not optimized. Refine when there is capacity |
| Weak | Underperforming or causing concern. Intervention required |
| Not yet running | Has not started. Either still in baseline or awaiting a scope decision |
Alongside the statuses: what last month produced and what it didn’t, what we propose for next month with each recommendation tagged to the piece of work that would deliver it, and the standing question—is the thesis holding.
A review that never names something that isn’t working has stopped being useful. We would rather write an uncomfortable sentence in month four than have an uncomfortable meeting in month eleven.
During the first sixty to ninety days the review covers instrumentation and observation only, and makes no claims about improvement. [The baseline period](/scale/baseline) explains why.
### What we need from you
Read it and push back. A review received without response has become a formality, and formalities are how partnerships drift.
Context we don’t have. A pricing change, a new salesperson, a large customer who signed through a personal relationship. Missing context is the most common cause of a misread month.
And a decision on the recommendations. Each is tagged to a piece of work, so agreeing means it gets scheduled, and declining means we won’t raise it again without new evidence.
## What we will ask you
Six facts, asked in the same order every time. They are worth knowing before the conversation, because most companies discover they cannot answer all of them, and that discovery is frequently the finding the thesis is built on.
* The source of each of your last ideal customers
* Your close rate from first conversation to signed
* Average customer value and average tenure
* Sales cycle length
* What historical marketing has been completed and what it produced
* Whether your website can identify who arrived and where they came from
Most companies cannot answer all six. That is not a problem with you, it is the finding. If there is no CRM, no lead scoring, and no way to tell which conversations came from where, then no amount of content will produce an answer you can trust, and the first bet has to be instrumentation. We would rather have that conversation in month one than in month twelve, when there is a year of work to account for and no way to account for it.
So if most of that comes back empty, we do not write you a thesis with invented numbers. We write one whose bet is the plumbing itself.
## How it gets written
We have a discovery conversation and record it. We draft from that conversation over the following few days. Then we deliver the draft live and open with some version of “Here is what I think is true about your business, and here is what I would bet on. Tell me where I am wrong.”
How the thesis fits into everything else that happens before Chapter 1.
Writing one is where every partnership begins. If the argument above sounds like the thing your marketing is missing, let’s talk.