Template pack

§The body of this whitepaper described what each step is for and deliberately left out how long it takes and who sits where. That is what this pack is for. Each template is a skeleton: the headings and the rules, so that you make each decision once and not every week.

§ A.11. Research plan#

§Half a page, written before any research. Assumptions: three to six, each as a claim that could be false. Target: one description tight enough to say who is out. Sources: one line each for desk, experts, communities, customers, saying what question each is expected to answer. Would change our mind: one sentence naming the finding that would make you stop. Revisited after every source.

§ A.22. Outreach message#

§Three sentences, never four. Who you are, in plain words. What you are trying to understand, framed as the problem and not the product. What you are asking for, specifically: a conversation of a stated length, at their convenience, about how they handle the problem today. No pitch, no link, no attachment. Sent to three times as many people as you need, expecting a third to reply.

§ A.33. Conversation guide#

§One page. Opening, in exact words: who you are, that you want to learn and not sell, that there are no wrong answers. Their world: two questions about their role and a normal day. The last time: "walk me through the last time this happened," followed by the six follow-ups: tell me more, what did you do next, how did that affect you, why was that important, what else did you try, what happened then. What they have tried: every workaround and why they stopped. Only if they raise it: the solution, one sentence, then back to them. Close: the commitment asks, in order. Introduction. Permission to return. A place in a pilot. Never say: a list of the phrases from chapter 15 that contain the product.

§ A.44. Same-day conversation page#

§Filled before you leave. Code, profile, date. Three verbatim quotes. The specific event they described and its cost in their units. Every workaround mentioned. What they agreed to at the end and how readily. One line: did this change what I believe, and how.

§ A.55. Problem analysis matrix#

§One table. Rows are problem clusters from the wall. Columns: frequency, pain, current solution, will pay. Frequency and pain from the stories. Current solution from the workarounds; "none" is a warning. Will pay only from evidence: said unprompted, or already pays for a worse version. Ranked, then the top two or three go to definition.

§ A.66. Problem definition document#

§Statement, in the five-line form: for, who, is a problem because, currently they, which. In the customer's words. Evidence: the quotes behind each line and the numbers you have. Current solutions: what people do today, why it is not enough, the gap. Impact: business, emotional, financial, time, in the customer's units. Context: market size as far as you know it, the competitors and adjacent tools, anything regulatory. Validation: what the loop in chapter 7 returned, from whom. Open: assumptions still open and the gaps you are carrying into Part II by choice.

§ A.77. Solution canvas#

§One page per candidate. Core concept: one sentence, key differences, main assumptions as claims that could be false. Journey: discover, start, get value, share, one line each. Requirements: must-haves, nice-to-haves deferred, hard parts, resources. Fit: who exactly, which line of the problem statement it changes, why now, who else is close. Risks: technical, market, business model. The one assumption that kills it: one sentence.

§ A.88. Brief for a machine#

§Cover line: "This is the file you upload." Scope: one paragraph. Personas: two, with names, ages, habits and a quote each. Rules already decided: numbered statements. Dataset: paste-ready, in local format: names, items, prices, places. Screens allowed: five at most, one line each, each with the number of the question it answers. Out: explicitly, the features that will not appear.

§ A.99. Prompt rules#

§Kept beside the brief, never uploaded with it. Flow before screen. One screen per message. Real data, never placeholder text. Describe the moment, not the interface. Ugly version first. One change at a time. When it fails, describe what you see. Ask for the failure states: empty, error, no signal. Do not ask for a design system. Attach images with exact instructions. Save the prompts that worked. And above all of them: upload only the brief.

§ A.1010. Field guide#

§One page for the whole team. What we are validating, and which open decision each item unblocks. The method in one paragraph: life first, then situations, then think-aloud, then commitment. The three mistakes: instructing, helping, explaining. Question bank: seven questions about their life, choose four or five per session. Opening, in exact words. Task table: each task as a situation, with what you are watching for. Closing asks: commitments, in order of cost. Roles: one who speaks, one who writes. Setting: their device, their hands, recording with permission.

§ A.1111. Same-day synthesis sheet#

§Code, profile, date. Three verbatim quotes, one marked critical. Where they got stuck, for how long, what they said while stuck. What they thought the unit of value was. What they agreed to at the end, and with what face. One thing you learned that you did not want to hear. Confidence: high, medium or low, and why. What this setting could not test.

§ A.1212. Tension map#

§Three columns: validated, open, discarded. One row per assumption carried since the research plan. Every row: the verbatim quote and the interviewee code that moved it. Every open row: an owner, a date, and how it will be closed. Read at the top of every strategy conversation.

§ A.1313. Two-act memo#

§A few pages, prose. The proposal in one line. Act one: what is in, what is out, why first; usually the piece that depends on no one and closes the biggest open question. Act two: why second, how it starts, in how small a footprint. Deferred: what is not in version one and why it is deferred rather than discarded. This week: the open rows whose date is now, each with a number to hit. Proposes; does not decide.