Expert and business books

How to write a book that gets clients

August 16, 2026 · The StoryProp team

A book that gets clients is designed to be used, and the money follows the using. A Comprehensive Study of Business Book ROI — an October 2024 survey of 301 self-selected published nonfiction authors led by Josh Bernoff with Dr. AJ Marsden — found more than 70% reporting an increase in leads or gigs after publication; in Bernoff's later analysis of the same data, 47% of the revenue authors attributed to their books came from something other than the book itself, rising to 69% among the authors whose books generated $25,000 or more. The object on the page does the work your first conversations do — proving you understand the reader's situation better than they can state it themselves, showing how you think when the information is incomplete and the stakes are real, letting them decide whether they want you near their problem — before anyone has booked a call.

That makes it a design problem, and the decisions are few: how much of the method to give away, who exactly you are writing to, what a case study is for, and how to end so the next step is obvious. Settle them before you start, because the object is not small. The same study puts median writing time at ten months — twelve for traditionally published authors, six for self-published — and the Editorial Freelancers Association's 2026 rate chart puts business and marketing ghostwriting at 50 cents to a dollar a word, roughly double its band for general nonfiction, which puts a 50,000-word book at $25,000 to $50,000 if you hire the writing out.

The shape of the return is the argument for the design. Counting only the authors who reported income in a given stream, the study's median increases run to $50,000 in consulting and $30,000 in speaking fees against $3,300 in direct book sales — and copies sold turned out to predict neither ROI nor whether an author met their goals. Thirty-one percent reported no financial increase in any category. The sample is worth naming, since every number here rests on it: a non-probability sample of 301 authors recruited through the four author-service vendors that sponsored the study — a hybrid publisher, a ghostwriting agency, a PR firm and a thought-leadership consultancy — so it leans toward authors who paid for professional help, and each revenue figure is the author's own attribution rather than an audited number. It is also the closest thing to hard evidence this category has; the more specific numbers circulating elsewhere come from firms selling book services. Nothing about publishing makes the return automatic; the design is the part you control.

Give the method away, all of it

The standard advice for a lead generation book is to give away the what and keep the how: state the principles, withhold the execution, leave the reader convinced and incapable. But readers can feel a withheld book. They register the example that ends before the hard decision, the framework whose stages are described but never operated, and they conclude one of two things — that the author has no method, only a vocabulary, or that the author has one and is dangling it. The first costs you their attention. The second costs their trust, the only thing the book was there to build.

The fear underneath the advice is that a complete book removes the reason to hire you. It doesn't, because what a practice sells was never the information. It is judgment applied to one particular situation, which the page cannot supply, never having met your reader's board. It is the outside view, which nobody holds about their own organization. It is sequencing under real constraints, the political ones especially. None of that travels in prose, however generous. The category's most copied example is Gino Wickman's Traction, which prints the whole of the EOS operating system — the meeting rhythm, the scorecard, the accountability chart — in enough detail to run without him, while the paid work is carried by a network of several hundred certified implementers.

A consultant who diagnoses failing product launches can publish the whole diagnostic: her questions in order, the answer patterns that mark each failure mode, the two modes that look alike at first and how to tell them apart. Some readers will run it competently, and they were never going to hire her anyway. The rest meet what the page cannot hand them: they are inside the organization being diagnosed. They know the answer to the third question and cannot say it out loud in that building.

Write for one reader who could hire you

Most business books are addressed to a category — founders, leaders, operators, teams — and read exactly as generic as that address implies, because nothing specific is true about a category. Write to one person instead. Not a marketing persona with a stock photograph and a list of pain points, but someone with a Tuesday: the operations director at a mid-sized manufacturer who inherited a scheduling mess she did not create, has twice been promised that software would fix it, and has the standing to try one more thing before the subject turns radioactive.

The useful refinement is temporal. Write to her as she was the week before she would have called you, not as she is deep into an engagement. That week she lacks your vocabulary, has not accepted your framing, and is braced for someone to tell her the mess is her fault. Fixing that moment settles everything downstream: which terms you define, which objections earn a full page rather than a footnote, which favorite advanced material has to go because it answers a question she will not have for another year.

Aiming this precisely repels readers, on purpose. The manufacturing examples lose the software founders; the frank chapter about what this costs a team in its first month loses anyone shopping for painless. Good. A book that fits one situation exactly is unmistakable to the people standing in it, and the conversations that follow start several steps in, since whoever calls has already agreed with the premises. One constraint keeps it honest: the book must still pay off for a reader who will never hire you. If a chapter's value depends on the reader becoming a client, it is not a chapter.

Case studies that teach instead of boast

The default case study has a familiar shape — the client's situation, the complication, the arrival of the author, the resolution, a closing line about results. It looks like a story and does none of a story's work, because it seats the reader in the audience. Nothing is taught; you are demonstrated. A reader deciding whether to trust a method does not need to watch it succeed in a room they weren't in. They need to see it operated. The 2025 Edelman-LinkedIn B2B Thought Leadership Impact Report, which measures published thinking broadly rather than books specifically, found 73% of decision-makers consider it a more trustworthy basis for judging a firm's capabilities than marketing materials and product sheets, and 86% say a consistent body of it makes them more likely to invite a firm into an RFP — which is the whole mechanism a book is built to run, and the reason it pays to position yourself as an expert by showing the work rather than the outcome.

A teaching case study is built differently. It opens with the information you actually had, in the disorder you had it, not the tidy version assembled once you knew the answer. It states the hypothesis you started with and why it was wrong, because being wrong early is the normal condition of the work, and hiding that is what makes a method look like magic. It stops at the decision point, names the option you rejected and what it would have cost. It gives the client their share of the credit, including the manager who was willing to be embarrassed in a meeting.

Include at least one case that went badly. A method with no failure modes reads as advertising, and readers discount its successes accordingly; a method with one honest failure in it becomes falsifiable, which is the precondition for being believed. That chapter is also where your judgment shows most, since anyone looks competent narrating a win. The ethics of client stories — anonymizing by function rather than name, disclosing composites, when to ask permission in writing — are their own subject, worked through in writing a book as a consultant.

Say where the method gets hard

Every method has joints where it can be applied wrong: the step that goes badly when a team is already under scrutiny, the diagnosis that misleads if the data was curated on its way up, the sequence that must be reordered in a family business. Name them. Say plainly which parts of the work are mechanical and which turn on a judgment call, and describe the judgment call closely enough that a reader can recognize one while standing in front of it.

This is the engine of the design, and it works because it is not an argument. You are not claiming the work is hard; you are showing where it is hard and letting readers measure themselves against it. Some will decide they can handle it and be right; others will find the wall exactly where you said it would be, and then they know two things — the size of the job, and who described it accurately beforehand. The book should also contain something a reader can run: a diagnostic, a checklist, an exercise that fits in an afternoon. The ones who get stuck know precisely where, and a specific question is how an engagement begins.

One habit protects all of this: a revision pass that checks every chapter still pays off on its own terms. You cannot see the drift while drafting, because each sentence felt reasonable at the time — the caveat that ends by noting most teams need outside help, the chapter close gesturing at what an engagement would add. Individually they are small; together they retune the book, and the ask lands harder when it arrives once, in its own place, than when it is spread thin across nine chapters. StoryProp is AI book writing software built on a permanent project record, so the rules you set in chapter one stay written down where the agent reads them before drafting, and a passage that drifted can be pointed at and revised without disturbing the pages around it.

The last chapter, and the page after it

The final chapter owes the reader a first move they can make on Monday without you — complete in itself, worth doing on its own terms, not a teaser for a larger program. A book that ends by making the reader more aware of their need has left them holding a problem. A book that ends by making them more capable has earned the right to be handed to a colleague, which is how most readers reach books like this at all.

Then, after the argument is finished, one plain page about how you work: who you work with, what an engagement involves, how to reach you, and a direct invitation to start one. Write it in the teaching voice, like everything else, and write it once, in full. Readers assume an author takes work; the page exists for the reader who finished the book at midnight wondering whether you take on work like theirs, and leaving it out fails exactly that reader. Make it durable: these books have long tails, and someone who finds yours years from now cannot act if the trail has gone cold.

None of this is a compromise between being useful and being commercial; it is the same act seen twice. In the ROI study, 68% of authors said the book made them more credible to prospects and clients, and 89% agreed that writing it was a good decision — 72% strongly. The reader who hires you is the one who read a book that respected them, tried the method on their own situation, got somewhere real with it, and then hit the wall you told them about, in the place you said it would be. Write the book that gets them to the wall.

The decisions that make a book like this work — one reader, the whole method on the page, a case study that stops at the hard call — are easy to set in chapter one and easy to lose by chapter nine. StoryProp keeps them as records it drafts against, so the book you designed in chapter one is still the book you are writing in chapter nine.

← All posts