New: The Offer Method, version 1 · One payment of US$149 · Every future update free

Offer Method v1 · Lifetime updates free

How much research is enough before I start?

Research done to reach a decision has a finish line: the moment you can write that decision in one sentence, with your strongest evidence beside it.

By Luyanda

Enough for what? Asked on its own, the question has no answer, and that turns out to be the most useful thing about it.

If someone already pays for your skill, an employer or a client who came back for a second piece of work, and nothing is on sale under your own name yet, finding out more before you start is the sensible move. You read about pricing. You find the people who already sell what you do, and study how they describe it. You note the questions people in your line of work ask in public. Each of those turns up something useful, and none of them arrives with a point where it’s done.

The answer here has two parts: which kind of research you are doing, and a rule for when each question in it is finished. Both come from the book this site sells, and it’s at the end of this page, once the idea is complete.

Two kinds of market research

Early in its research chapter, the book splits market research in two, and the split does most of the work of answering the question.

Academic research, in the book’s words, asks what is true about this population? It “describes people as completely as it can, and it produces a report.” Commercial research asks will someone pay to have this problem solved, and will they choose me to solve it? It “counts behaviour as evidence: money spent, work posted, workarounds built, complaints written”, and it “stops the moment a decision can be made, and it produces a decision.”

Then the line the rest of the chapter stands on: “Academic research describes a market. Commercial research decides what to do in one.”

Put the question of enough to each kind and it answers itself. A description can always be more complete. There is another group of buyers to profile, another survey to run, another report on the same industry from a different angle, and each one adds real information. Completeness has no last page, so research aimed at it has no finish line. A decision is a single event. Once you can make it, the research has done its job.

Advice on market research is often advice for the first kind: how many people to survey, how large a sample has to be before a result can be trusted, how to build a persona, how to size a market. Those are good questions when the job is to describe a population. Brought to the question in front of you, whether a stranger will pay for a skill you already have, they produce a report, and the decision is still waiting on the other side of it.

Say your skill is month-end reporting for a finance team, or writing the safety procedures a warehouse actually runs on. The academic question asks what is true of finance leaders, or of warehouse managers, as a group. The commercial question asks who is already paying someone outside their own company to do that work, and whether they would choose you.

Research that ends in a decision

Everything that follows in the book’s research chapter is the second kind, and it is laid out so that it finishes. The research runs as five questions, the five Ws: WHAT, WHO, WHERE, WHEN and WHY, in that order. Each one closes with a decision you write down before you move on. Why they run in that order is an argument of its own, and it’s in the book.

A decision gives research a shape it otherwise lacks. With a sentence to write, you know what you’re looking for, and you know to stop reading once it’s written. At the end you hold five decisions, each one something you can act on.

The book ties the research to one more thing, and it’s set before you reach the research chapter: a date. In its words, “the day your first messages go out.” Then, to be clear about which date it means: “Not the day you’ll finish researching. The day real people receive an offer from you.” Research, in this method, is measured against sending.

The stop rule: one sentence and its evidence

Here is the rule the book gives for when to stop, as it sets it:

Each W ends when you can write one decision sentence for it, with its strongest evidence named beside it.

It has two halves, and each one does a job.

A decision sentence is a choice written plainly enough that it could turn out wrong. It commits to something you could act on this week. “There seems to be real interest in this area” fails: nothing could prove it false, and nothing follows from it. A sentence that names a particular kind of buyer with a particular problem passes, because you can act on it and the world can disagree with it.

The evidence beside it keeps the sentence honest. It makes you say what the decision rests on, and it lets anyone, including you later on, see how much weight it can hold. The book grades every finding on what it calls the evidence ladder, with money people already spend at the top; how that grading works is the subject of the essay on whether anyone would actually pay for your skill. Here, the point is narrower: you name the strongest thing you found, and you put it next to the sentence it supports.

This is what the rule looks like in the worked example that runs through the book from its first chapter. Once the research there turned toward money, its WHAT came to rest on these lines:

Decision: Companies of 20-80 people that are hiring several people
  at once lose weeks to slow starts, because nobody has built the
  system. Some of them already pay outsiders to build it.
Confidence: high on the problem; medium on what they pay.
What would change this: finding that every brief comes from one
  sector.

From the book’s chapter A1: the worked example’s WHAT, decided. The example is onboarding. Yours is whatever someone has paid you for. The people in it are imagined. The method is real.

Read it against the rule. The first sentence is the decision: a particular kind of company, a particular problem. The second names the strongest evidence beside it: money already moving, because some of those companies already pay outsiders to do the work. Directly above it in the same file sit the evidence lines: contract briefs from companies that size, and consultants selling the work at a fixed price on their own pages. Nowhere in that part of the file is there a market-size estimate or a survey, and the sentence doesn’t need either. It can be acted on, its confidence is stated, and the last line says what would overturn it.

When the sentence won’t come

Sometimes you sit down to write the sentence and it won’t come. The book’s instruction for that moment is one line: “If you can’t write the sentence yet, name the single missing fact and fetch it once: one more search, or one more run of the agent.”

The two limits in it are the point. Single means the gap is one fact you can name. A general sense of not knowing enough gives a search nothing to look for, so that gap never closes. A named fact does: whether anyone outside your own industry pays for this, say, or whether the people paying are the ones you assumed. One search can answer a question like that.

Once is the other limit. A second and third pass at the same gap tends to bring back the same evidence in new words, and each pass feels like progress. The book makes a related point about the prompt in its first chapter: “Rewording it and running it again gives you more guesses, not better ones.” After the one fetch, you try the sentence again. If it comes, write it. If it doesn’t, the book’s next instruction applies.

Decide, and write the confidence down

The next instruction is the one that makes the rule usable: “Then decide at whatever confidence the evidence supports, and write that confidence down. High, medium or low, with a reason.”

A confidence level is what lets you decide before you’re certain without pretending you are. It records, in a word anyone can read, how far the decision can be trusted. The worked example does it in one line, “high on the problem; medium on what they pay”, and splits its confidence in two because the file was surer of one half than the other. Under it sits a line that turns a decision into something checkable: what would change this. The example’s answer is specific, “finding that every brief comes from one sector”, so the decision comes with the evidence that would overturn it, named in advance.

Written down, the confidence is also a note to your later self. It says which decisions were taken on strong evidence and which on thin, so you can see what to recheck first when anything changes.

Stopping before you’re certain means some of these decisions will be wrong, and the book doesn’t pretend otherwise. It has you write every decision down beside its confidence and what would change it, and Section D, the last of its four sections, “is for when the answers come back, silence included.” A decision taken at medium confidence is one you have already said might not hold. Some of them won’t.

Why waiting for certainty is the trap

The book’s reason for all of this fits in two sentences, and they’re worth reading whole:

“Deciding with your confidence written down, and moving on, is the skill. Deciding only when you’re certain is the trap, because in a market nobody is ever certain, and waiting for certainty is how research eats the weeks meant for selling.”

Nobody is ever certain because the fact that would settle it doesn’t exist yet. It’s a reply to an offer you’ve made, and research, by its nature, comes before the offer. So the last stretch of uncertainty can’t be researched away. It can only be named, written beside the decision, and carried into the sending.

In the book, this idea belongs to one chapter: A7, “The five Ws: find out what your market will pay for”, the last chapter of Section A. The chapter after it is B1, “Decide: turn the research into a direction.” So the answer to the question at the top of this page is a point you can see when you reach it: five decision sentences written down, each with its strongest evidence and its confidence beside it. That is enough.

This book can’t promise you a client, a sale or an income. Whether anyone says yes is a stranger’s decision, and nobody can sell you a stranger’s decision.

You have a skill someone has already paid for. This is a book and a team of AI agents that take you, on evidence you can check, to answering “so what do you do?” with a price in the sentence — and to a live page and messages you’ve already sent. Run its first step on your own work, free, before you buy.

The book’s contents, from the start

The essay above is one idea from the book. Below is the front of the book’s contents; the full list is one click away, and so is the free first step.

The Offer Method · Version 1 · Contents

Before you begin

  • An opening

  • What this is made of

  • What you’ll walk away with

  • What this can’t promise

  • What it’s for

Section A — The machine, and your free will

  • A1. What you’re holding

  • A2. Two roads to the same place

  • A3. Road one: set up the machine

  • A4. Road two: the browser kit

  • A5. Meet your team

  • A6. How to run it and stay in charge

  • A7. The five Ws: find out what your market will pay for

The book’s own contents page, cut after Section A.

Free emails: new essays, open questions, what changed

You’d get essays like this one as they’re published, the questions we can’t yet answer, and each version’s note of what changed, sent only when there’s something to report.

Also from the book, taught in full

Do I need another course before I start?

Only when a question about your own work needs one. A course or a tool earns its place at that moment, and the question comes first.

Would anyone actually pay for my skill?

Look for money already moving to solve the problem your work solves. A friend’s “I’d buy that” and likes on a post count for little, however many you have.