How to Create a Framework for Your Business Book (From Patterns You Already See)

Most advice on business books says you need a framework. You do not have one. You have fifteen years of client work, a way of seeing problems that clients pay for, and no four-box diagram with a clever name.

The temptation is to invent one this weekend: five words that spell something memorable, arrows between them, book built around it.

Experienced readers can tell. A framework made for the cover tends to have stages that overlap, an order that does not hold, and one element that exists only to complete the acronym. A framework found in your client work has none of those problems, because it describes something you already do.

If you already teach a named model, read how to turn a framework into a book instead; if your method is a service you deliver, writing a book around your methodology fits better. The main site’s short guide to turning expertise into a book framework sets out the principle. This is the workshop version, with the tests we would run.

Step 1: Write up 10 to 20 client situations, not principles

Do not start by listing what you believe. Ask an expert for principles and you get the tidy conference version. Ask about a specific client and the real method comes out.

User researchers rely on the same idea. The critical incident technique, which Nielsen Norman Group traces to a 1954 paper by John Flanagan, asks people to recall “a time when a behavior, action, or occurrence impacted” an outcome, rather than to describe their work in general (Nielsen Norman Group). The article also recommends collecting both positive and negative incidents. You want both.

Pull 10 to 20 clients from the last few years: wins, stalls, and at least two or three engagements that went badly. Fewer than ten and the model bends around your two favourite clients. For each one, write a card of about 150 words. Use initials or a code, not names; these are working notes, and anything that later goes in the book needs the consent and disguise covered in using client stories and case studies in a book.

The “first thing you noticed” line matters most. It is where your diagnostic expertise lives, and experts leave it out because it feels obvious. If writing the cards stalls, extracting your knowledge before writing a book covers interview prompts that get it out of your head.

Step 2: Sort the cards until a pattern shows

Print the cards or put them on a wall. Then sort them into groups without deciding the groups first. This is a version of affinity diagramming, which Nielsen Norman Group defines as “organizing related observations, ideas, concepts, or findings into distinct clusters” (Nielsen Norman Group). Two pieces of its advice apply directly: start grouping notes “based on content before deciding on a theme label”, and “don’t try to force ideas together.”

Sort several ways: by what you noticed first, by where the turning point came, by what went wrong in the failures. You are looking for one of two things.

A repeating sequence. The same moves, in roughly the same order, across most cards. Clients had to do A before B made sense, and when they skipped A, B failed.

A repeating tension. The same trade-off, or the same two variables, deciding what you did. Every card turns on, say, how urgent the problem was and whether the leadership team agreed on it.

If neither shows after three or four sorts, see the last section of this article.

Where software fits at this stage

If your client history lives in proposals, session notes and talk transcripts rather than in your memory, finding the cases is the slow part. Guided software is one option for that first pass. WritebookAI runs a source analysis on PDF, Word and text files and pasted notes and shows what the material covers and what is missing, and a verification interview asks you about the gaps and conflicts. It will not find your framework for you or decide which client cases are representative, and you check everything it reports against what actually happened.

Step 3: Decide what kind of framework the pattern is

Most advice treats “framework” as “a numbered list of steps”. That is only one shape. Match the type to the pattern you found, not to whichever looks best in a keynote.

Type	What it claims	Fits when your cards show	A well-known example	Sign you have forced it
Sequence	Do these stages in this order	The same moves in the same order, and skipping one breaks the next	GTD’s five steps: capture, clarify, organise, reflect, engage (Getting Things Done)	Clients regularly enter at stage 3 or do two stages at once
Matrix / 2x2	Two variables sort every situation into one of four types, each needing a different response	A repeating tension between two things you assess	BCG’s growth-share matrix, built on market share and growth rate (BCG)	One quadrant is empty in your cases, or the axes are not independent
Pillars	Several things must all be present; order does not matter	Failures each trace to a different missing ingredient	We found no single standard example; common in leadership books	Pillars overlap, or one pillar holds most of your stories
Diagnostic	A set of questions that locates where the problem is	You spend the first meeting working out which of a few problems the client really has	Many consulting audits work this way	The questions do not lead to different actions
Maturity model	Named levels an organisation or person moves up through	Clients cluster at recognisable stages of capability	CMMI, whose levels run from Initial to Optimizing (CMMI Institute)	Clients sit at different levels on different dimensions

Two opinions. The 2x2 is the most overused shape in business writing because it looks rigorous; if you cannot name a client in each quadrant, you do not have a matrix. The diagnostic is the most underused, and suits consultants well, because diagnosis is often what senior buyers pay for.

Step 4: Check the parts are distinct and the order is real

Now draft the model as plain labels on one page. Then run three quick tests against your cards.

1. The one-home test. Go back through every card and put each key moment under exactly one part. If a moment fits two parts equally well, those parts overlap. Merge them or redraw the line between them.
2. The swap test (sequences and maturity models only). Swap two adjacent stages and ask whether the cards still make sense. If they do, the order is your teaching preference, not something the work requires. That is fine, but then it is pillars, not a sequence.
3. The count test. If you have more than seven parts, a reader cannot hold the model in their head. Find the two that are really one.

Our rule: a stage earns its place only if at least two cards show getting it wrong caused a visible problem.

Step 5: Run it against the cases where it did not work

This is the step that separates a framework from a slogan. Take the cards where your approach failed, stalled or needed something unusual, and ask what the draft model says about each.

You will get one of three results:

• The model explains the failure. A stage was skipped or done badly. Good: that failure is now a chapter’s best story.
• The model is silent, but the case was outside your scope. The client needed a therapist, a lawyer or a different product. That becomes a boundary you state in the book (“this is not for you if…”), which senior readers trust more than a model that claims to work everywhere.
• The model is silent and the case was in scope. Something is missing. Add the part, or accept that the model is narrower than you hoped and say so.

Keep a short list of exceptions you cannot yet explain. Do not hide them.

Step 6: Name it plainly, then use it with clients

Name the model after it survives the tests, not before. A plain name that describes the outcome or the action (“the four stalls”, “the handover sequence”) is easier to say in a meeting than an acronym, and easier for clients to repeat to colleagues. If you catch yourself choosing a word because it starts with the right letter, the acronym is now designing the method. The main-site guide warns that forcing ideas into a neat acronym “can remove the nuance that made the expertise valuable”. Naming in more depth, including trademark checks, is covered in how to turn a framework into a book.

Then use it in your next five to ten client sessions before you write a chapter. Every time you add “well, except…”, note it: each is a missing part or a boundary. A month of this is cheaper than rewriting three chapters.

A worked example: a sales consultant with no model (hypothetical)

Imagine a consultant who helps B2B software founders with stalled sales pipelines. She has worked with about forty companies and has never named what she does. She writes 16 client cards.

First draft. She assumes her method is a sequence: define the buyer, quantify the problem, map the decision, secure a sponsor, close. Five neat stages.

The swap test breaks it. Her cards show clients entering at every stage. One company had a clearly defined buyer and a quantified problem but no sponsor. Another had a sponsor but had never put a number on the cost of doing nothing. The order was how she taught it, not how the work happened.

Re-sorting by “first thing I noticed” shows the real pattern. In 13 of 16 cards, her first meeting was spent working out where deals were stalling, and the answer was one of four places: no single named buyer, no cost of inaction the buyer believed, no next step the buyer owned, and no internal champion once the deal reached procurement. Each needed a different fix. That is a diagnostic, not a sequence.

The failure cards add a boundary. Three cards fitted none of the four stalls; in each, the product did not solve a problem buyers would pay for. The book will say plainly that if all four checks pass and deals still die, a sales book will not fix it.

The name is plain: the Four Stalls. Each stall becomes a part of the book, with a self-assessment question at its head. She tries the questions on her next six clients and rewrites two that founders kept answering with “sort of”.

Notice what she did not do: keep the five-stage sequence because it was already on a slide.

When not to force a framework

Some books are better without one, and we would rather you wrote a strong book of that kind than a weak framework book.

An argument book. Your main contribution is a belief that changes how a reader sees a problem: that most leadership training fails because it ignores the manager’s own manager, say. The book’s structure is the case for that belief. A diagram would shrink it. How to write a thought leadership book covers this shape.

A field guide. Your expertise is judgement across many situations that do not share an order or a tension. Chapters organised by situation (“when a co-founder wants out”, “when your best client asks for a discount”) can serve the reader better than an invented model linking them.

A collection of hard cases. Some senior advisers are most useful showing how they thought through ten difficult engagements. The reader learns the pattern from the cases.

Our test: if after Step 2 you have sorted your cards three different ways and nothing repeats in more than half of them, do not force it. Organise the book another way; organising years of experience into book chapters and how to structure a business book lay out the alternatives.

This week: ten cards, one sort

Do not try to design the framework this week. Write ten client cards using the checklist above, including at least two failures, and do one sort by “first thing I noticed”. That is two or three hours of work.

If a pattern shows, carry on with Steps 3 to 6. If it does not, write ten more cards and sort again before deciding the book needs a different shape. Either way, you are building from what you actually do.