You have a folder of book ideas. It might be a notes app, a notebook, or a document you've added to for years. It holds topics you know and care about. If you're honest, though, you can't say why a stranger would go looking for any of them. That's the quiet trouble with choosing a book idea from the topic side. You pick the subject that excites you most, outline it, and start writing. Months later you're deep in chapters, and the question you skipped on day one is still waiting: who is stuck, on what, and how would they describe it? A topic-first book idea is a guess with a tidy outline on top. Good writing doesn't fix a guess. You can write a clear, careful book about something nobody was trying to solve. There's a better order. Start with the reader's problem, in the words readers already use. Then decide whether you have a book idea worth your time. I call this the Pain Gap method, and it has four moves: Collect observable reader problems. Capture the exact wording. Separate what each observation suggests from what it doesn't prove. Narrow to one problem worth solving first. It won't tell you what will sell. It will replace "I think people want this" with evidence you can point to and weigh honestly. Free download: The Pain Gap Book Idea Scorecard. Capture the reader problems you uncover and turn them into book candidates. The scorecard's Reader Pain Evidence Bank gives every problem you find a place to live — the exact words, where you found them, and what they do and don't prove — so you can score up to three book ideas side by side before you commit. A Book Topic Is Not a Book Idea (topics vs. reader problems)Where Reader Problems Actually Show UpReviews and reader questions as observable languageWhat a complaint can and cannot tell youHow to Capture Problems Without GuessingHow to Tell a Real Problem From a Passing ComplaintNarrow the List to One Problem Worth Solving FirstYour Next Step: From Problem to Outline A Book Topic Is Not a Book Idea (topics vs. reader problems) A topic is an area of interest: sleep, grief, remote management, personal finance. Topics tell you where your knowledge lives, but they describe your side of the table. Nobody wakes up needing "a book about sleep." They wake up exhausted and can't work out why they're still awake at 3 a.m. A book idea has a different shape. It names a reader, a stuck point, and a change. Here's the contrast, using illustrative examples only: Topic (your side)Reader problem (their side, illustrative)Sleep"I'm tired all day but wide awake at night."Remote management"My team is online and I can't tell who's struggling."Personal finance"I earn decent money and still don't know where it goes." Those lines are made-up samples to show the difference. They aren't research findings. The Pain Gap is the distance between where a reader is and where they're trying to get, described in their own terms. The book lives in that gap. If you can't describe the gap, you have a topic, not a book idea. Try this test on each item in your folder: "A reader who ___ is trying to ___ but ___." If the only way to fill it in is with generalities, you've found a topic. Most folders are full of them, and that's normal. The rest of this article is how to fix it. Where Reader Problems Actually Show Up People state their problems all the time, mostly to each other and mostly unprompted. You don't need special tools to see them. You need to look where people already talk: Book reviews on existing titles in your area Reader questions: Q&A sections, comment threads, questions left for authors Complaints about books, courses, tools, or the situation itself Forum and community posts in niche or professional groups What these share is that they're observable. You can read them, quote them, and link to them, and someone else could look at the same thing and check your reading. That's the standard. If a problem exists only in your head, or in "everybody knows," it doesn't go on the list. Reviews and reader questions as observable language Reviews of existing books are useful because readers write them right after trying to use the book. Look for sentences that start with "I wanted," "I was hoping," "I still don't know how," or "it never explained." They show what the reader expected and where they felt left on their own. Reader questions work differently. A question in a forum thread is a problem in its rawest form, before the person has tidied it into something polite. Notice the situation they describe, what they've already tried, and the words they pick. Pay attention to vocabulary. You may call something "habit architecture" while a reader says "I can't keep anything going past week two." (Illustrative, again.) The reader's wording is the language a book has to meet them in, so keep it exactly as written. What a complaint can and cannot tell you A complaint is useful and limited. It can tell you that a real person had this frustration. It shows the words they used, what they expected, what they'd already tried, and where existing material may have left something unanswered. It cannot tell you how many other people feel the same, or whether anyone would choose a book to fix it. It can't say whether a book is the right format, or whether the complaint is about a weak book rather than an unmet need. It also doesn't show whether this person is typical of anyone. So hold onto this rule: a reader question or complaint is a clue, not proof that a book will sell. I won't give you a number of complaints that equals a market, because no honest number like that exists. What you can do is collect clues carefully, record what each one does and doesn't show, and decide with your eyes open. How to Capture Problems Without Guessing Guessing sneaks in when you summarize. You read a review, think "readers want more practical steps," and write that down. Now you've recorded your interpretation, and you can't tell later which part was theirs. Capture the evidence first and your reading second. For each observation, record five things: The exact words, copied as written. Don't tidy or paraphrase. Where you found it: the source type, plus a link or title so you can return. What it suggests: one sentence, labeled as your reading. What it doesn't prove: one sentence. Most people skip this step, and it matters most. The problem it points to, written as a stuck point rather than a topic. Here are two sample entries. Both are illustrative, invented to show the format. Sample entry A (illustrative) Exact words: "I've read three books on this and I still don't know what to do on Monday morning." Where: Review of an existing book in the category Suggests: Some readers may find explanation plentiful and a first action missing. Doesn't prove: How many readers feel this, or that it applies beyond this one book. Problem it points to: Not knowing what to do first Sample entry B (illustrative) Exact words: "Does anyone else feel like they're doing everything right and it's still not working?" Where: Community forum post Suggests: A reader may be following advice, seeing no results, and wondering if the fault is theirs. Doesn't prove: That others in the thread share it, or what the person actually tried. Problem it points to: Unclear why correct-seeming effort isn't producing results These are not real quotes and not evidence of anything. Keep your own entries in one place, such as a single document or spreadsheet, so you can compare them later. How to Tell a Real Problem From a Passing Complaint Not every complaint is a problem worth a book. Venting about a bad day, a shipping delay, or a typo isn't the same as a reader stuck on something they need to get past. Run each entry through these questions: Is it about the reader or the product? "This book was boring" is a verdict on one book. "I can't figure out how to start" is about the reader's situation. Is it specific? "I manage six people across three time zones" carries more than "management is hard." Have they already tried something? Someone who has read, tested, and is still stuck is describing a more durable problem than someone who just noticed a question. Does it show up in different places, from different people? The same stuck point in reviews, forums, and questions, phrased differently by different people, is stronger than one person repeating themselves. I won't put a number on "enough." Count what you have and say plainly how thin or thick it is. Does it exist without a book to complain about? If people raise it with no title involved, it's more likely a standing problem than a reaction to one book. Is something at stake? Time, money, stress, a relationship, a goal. Stakes help explain why someone would go looking. These questions help you rank problems against each other. They don't forecast anything. A problem that passes all of them is still a well-supported clue, not a guarantee. Narrow the List to One Problem Worth Solving First By now you have a pile of entries and probably several problems. Group the entries by the problem they point to, not by topic. Then pick up to three candidates to compare. Any more and you're back to a folder. For each candidate, ask: Can I state it in the reader's own words, using quotes I actually collected? How varied is the evidence, and how thin? Think sources, people, and phrasing. Can I credibly help with this, through experience, results, or knowledge I can demonstrate? Does it fit in one book with a clear before and after, or is it several problems stacked together? Is a book a reasonable way to solve it, or would a checklist or a quick answer do? Put the candidates side by side and let the answers decide. You aren't choosing the problem you predict will sell best, because you can't know that. You're choosing the best-supported problem you're able to solve well. Park the others. They aren't rejected, just later. Then write the one-sentence version: "This book helps [reader] who [situation] get past [stuck point] and reach [change]." If that sentence feels shaky, go back to the evidence, not to the thesaurus. Your Next Step: From Problem to Outline You started with a folder of topics. You've now collected observable problems, captured the exact words, separated what each one suggests from what it doesn't prove, and narrowed to one. That's a book idea with a reason behind it, plus a record of what you know and what you don't. The next step is building the outline around that single problem. Every chapter should either move the reader across the gap or not belong. I cover that in the next piece: How to Outline a Nonfiction Book Around One Reader Problem If it helps to keep the evidence and your candidates in one place, the Pain Gap Book Idea Scorecard from earlier is built for that. A plain spreadsheet works too, and the method is the same either way. One last reminder: everything you've gathered is a set of clues. Hold it loosely, and be honest with yourself about what's thin.