The Same Six Problems: What Turns Up in Almost Every Manuscript

After 38 years and 1,100+ titles, the surprise is not how different books are. It’s how similar the problems are. Six of them turn up in almost everything that crosses my desk, from first novels to books that went on to sell in the hundreds of thousands, and none of the six is about talent. They are structural habits that are close to invisible from the inside, which is the only reason they survive into a finished draft. Here they are, with what each one looks like and how to test your own manuscript for it.

Key facts

  • 1 The book starts before the story does
  • 2 It isn’t about what it spends its pages on
  • 3 Things happen but nothing changes
  • 4 One point, made three times in three costumes
  • 5 Everyone sounds like the author
  • 6 The ending explains what the book already showed
  • Why they survive They are invisible from inside the manuscript
  • All six are Structural, so they are fixable

1. The book starts before the story does.

The most common structural problem in manuscripts, by a wide margin. Somewhere between page 20 and page 90 is a sentence where the book actually begins, and everything before it is the writer arriving.

It takes a recognizable shape. Chapter one establishes a world, a history, a family, a set of conditions. It is frequently well written, which is part of the trouble. Then something happens that changes the situation, and the reader who made it that far discovers what they’ve been waiting for. The material before it isn’t wrong. It’s context delivered before anyone needed it.

The reason it happens is mechanical rather than mysterious. You had to write your way in. Nobody starts a book knowing where it starts, so the first 50 pages are you working out what you’re writing. That work is real and necessary; it just isn’t the book.

The test: find the first moment where the situation changes and cannot go back. Now read the book starting there, and ask what from the earlier material you genuinely miss. Usually the answer is three or four facts, and those can be placed later in a paragraph each, at the moment they’re needed. Occasionally, you miss nothing at all.

2. The book isn’t about what it spends its pages on.

Ask a writer what their book is about and you get one answer. Count the pages and you get a different one. The gap between those two answers is where a great deal of structural editing happens.

A memoir described as being about recovering from illness spends 200 of its 300 pages on a career. A business book pitched as a method spends most of its length on the author’s biography. A novel the writer describes as being about two sisters devotes its energy to a village.

Neither answer is necessarily the wrong one, and that’s the interesting part. Sometimes the stated subject is right and the pages need to follow it. Sometimes the pages are telling you what the book actually became while you were writing it, and the stated subject is left over from the proposal. Discovering which is one of the most valuable conversations in a structural edit.

What is never right is leaving the gap open. A book divided against itself asks the reader to hold two purposes at once, and they will resolve it themselves, usually by deciding the book is unfocused.

The test: write the page count next to each chapter, then write the one thing that chapter is for. Read the list without the manuscript. The list is what your book is about. If it isn’t what you would have said, you have found the thing to fix, and you have found it in an afternoon rather than a year.

3. Things happen, but nothing changes.

This is what people mean when they say a middle sags, and the diagnosis is almost always more specific than “pacing.” Events are occurring, but the situation isn’t moving. A reader who can’t see the situation moving loses the thread of why they are reading, even when each individual scene is good.

It shows up in fiction as a run of chapters where characters do things yet the balance of power, knowledge, or need at the end of the run is what it was at the start. In nonfiction, it shows up as a sequence of examples all making the same point at the same level, so the argument accumulates evidence without advancing.

The mechanism is straightforward. Middles get written over the longest stretch of time, when the ending isn’t yet fixed, so material gets added because it’s good rather than because it changes anything. Each addition is defensible on its own. Together, they form a plateau.

The test: for each chapter, write one sentence beginning “by the end of this chapter, X is now true and was not true before.” Where you can’t finish the sentence, or where the answer repeats the previous chapter’s answer, you have found the plateau. Combine, cut, or escalate.

The fix is usually not to add drama. It is to reorder, so that what a character learns in chapter fourteen arrives in chapter nine and changes everything after it.

4. The same point, three times, in three costumes.

A writer makes a point in chapter two, worries it didn’t land, and makes it again in chapter six wearing a different scene. Then again in chapter eleven. Each version is well made. Together, they teach the reader that the book repeats itself, which is the most efficient way to lose their attention.

In nonfiction, it’s nearly universal because the chapters were often drafted separately and each one restates the premise. In fiction, it appears as three scenes demonstrating the same quality in the same character: three occasions on which the father is unreliable, none of which costs anything more than the last.

The reason it survives is that you never read your book the way a reader does. You wrote chapter two in March and chapter eleven in October, and you have never once read them within twenty minutes of each other. A reader reads them on consecutive evenings, and they notice.

The test: list every scene or section by what it demonstrates rather than what happens in it. Sort the list. Anything appearing more than once is a candidate. The rule I use: keep the version with the highest cost to the character or the strongest evidence for the argument, and cut the others outright rather than trimming them.

That last part matters. Three half-scenes are worse than one whole one, and the instinct to shorten all three instead of cutting two is how books end up with a flat middle in which nothing quite lands.

5. Everyone sounds the same.

Every character speaks in the author’s rhythm, at the author’s education level, with the author’s vocabulary and sense of humor. In nonfiction, every case study and every interviewee arrives sounding like the writer’s own prose.

It is subtle enough to survive many drafts because the dialogue isn’ bad. It’s well written. It’s all well written in the same way, and the effect is a book in which the reader can’t tell who is talking without the attributions, and in which the argument never encounters a genuinely different voice.

The usual cause is that you hear all of it in one head, which is yours. Distinctiveness has to be built deliberately: not accents, which age badly and are usually a mistake, but sentence length, vocabulary range, what a person says when they’re uncomfortable, what they refuse to discuss, whether they answer the question asked.

The test: take 20 lines of dialogue from across the book, strip every attribution, and read them cold. Can you assign them? Better: ask someone who has read the book to do it. If they can’t, the characters are being distinguished by their name tags rather than by their speech.

For nonfiction, apply the same test to quoted material. If everyone you interviewed sounds like you, you’ve been paraphrasing them into your own voice, and you’ve lost the corroborating power of a real person saying something in their own words.

6. The ending explains what the book already showed.

The last chapter tells the reader what it all meant. Two or three pages, sometimes a whole chapter, in which the writer states the theme that the preceding 300 pages already demonstrated.

It’s there not for vanity but anxiety: you’ve spent years on this and you can’t bear the possibility that someone finishes it and misses the point. So you say the point.

The cost is precise. A reader who has understood is now being told what they understood, which is faintly insulting. And a reader who hasn’t understood isn’t helped because a stated theme is a much weaker thing than a demonstrated one. You spend the last pages of the book converting something the reader felt into something they were told.

The test: delete the final few pages and read the new ending. Then ask a fresh reader who hasn’t seen either version what the book was about. If they can tell you, the explanation was unnecessary. If they can’t, the problem is earlier in the book and the final pages were covering for it, which is diagnostic information you didn’t previously have.

There are a few caveats. Literary fiction, memoir, and certain traditions (reflective essay, braided essay) legitimately use concluding meditation. If the reflection adds something beyond restatement (emotional resonance, expanded scope, new frame), it survives the test. If it doesn’t, it fails.

In complex speculative fiction, dense academic argument, or when crossing cultural frameworks, explicit articulation can be functional rather than compensatory. Again, the test catches this—readers who actually benefit from the signpost will show it when you run the diagnostic.

The same applies to the first paragraph of a nonfiction chapter that announces what the chapter will argue, then argues it, then summarizes what it argued. One of those three is doing the work.

Why can’t you see these in your own book?

Because none of them is visible from inside the manuscript, and that’s a fact about how reading works rather than a failure of skill or attention.

Three mechanisms, and they compound:

You can’t read what you wrote, only what you meant. Your eye supplies intention. Where the page is thin, memory fills it in, and the gap between what’s on the page and what’s in your head stays invisible to precisely one person.

You never read it at reading speed, in order, in a few sittings. You read it in fragments, out of order, over years. Every one of the six problems above is only detectable across a continuous read, which is the one thing you can’t give your own book.

You know how it ends. Suspense, confusion, and the pacing of information are all invisible to someone who already knows. You can’t experience your own withheld information as withheld.

That’s the entire argument for an outside reader, and it’s why the value of one doesn’t decline as you get better. Experienced novelists have exactly the same blind spot; they have simply learned to expect it. What changes with experience isn’t that you stop having these problems, it’s that you stop being surprised to find them.

How do you test your own manuscript for all six?

One afternoon, one document, no reading of the actual prose. The reason this works is that all six are structural, and structure is visible in a summary in a way it isn’t visible in the text.

Build a single table, one row per chapter, five columns:

  • Page count. Reveals problem 2, where the pages disagree with the stated subject.
  • What this chapter is for, in one sentence. Reveals problem 4, repetition, once you sort the column.
  • What is true at the end that wasn’t true at the start. Reveals problem 3, the plateau, wherever the cell is empty.
  • Whose voice dominates. Reveals problem 5, if the answer is always the same.
  • Could the book start here? Reveals problem 1, and the answer is usually yes for at least one chapter that isn’t chapter one.

Then delete the last few pages and see whether anything is lost. That’s problem 6.

This will take a handful of hours on a full-length manuscript but it finds more structural trouble than a month of rereading because rereading puts you back inside the prose where none of it is visible. If the table shows you nothing, that is genuinely useful information: your structure is probably sound and your money should go to a line edit rather than a structural one.

If you want the same analysis done by someone who has run it a thousand times, that is what a manuscript evaluation is.

What does this say about talent?

Very little, which is the point I most want to leave you with.

I have worked on more than 1,100 titles, including 70 that became bestsellers. Those books had these problems too. Not fewer of them and not milder versions: the drafts that became successful books arrived with openings in the wrong place, plateaus in the middle, and endings that explained themselves. What separated them wasn’t that they were born without the six. It’s that somebody found them and the writer did the work.

The reason this matters practically is that writers routinely read a structural note as a verdict on their ability. It isn’t. It’s a report on a draft, and drafts are supposed to have these problems because they are the natural product of writing a long thing over a long time without being able to see it whole.

The genuinely useful distinction isn’t between writers who have these problems and writers who don’t. It’s between manuscripts where they have been found and manuscripts where they haven’t. The first kind gets published and the second kind gets rejected with a note saying “didn’t quite hold my attention,” which tells the author nothing about which of the six was responsible.

Frequently asked questions

Do published books have these problems?

Their drafts did. Almost every book I’ve worked on had one if not all of these issues to varying degrees. The difference between a published book and an unpublished one is very often just that somebody caught them in time.

Which of the six is most damaging?

Starting too early because it costs you readers before any of the other issues get a chance to. An agent, a bookseller, or a browser makes a decision in the first pages, and a book that hasn’t begun yet loses that decision no matter how good page 90 is.

Can I fix these myself?

Often yes, once you know they are there, which is the hard part. The chapter table above finds most of them without any outside help. What you can’t do alone is see them in the first place while reading normally, which is why the table works and rereading doesn’t.

Is this the same for fiction and nonfiction?

All six occur in both, with different faces. Repetition is more common in nonfiction because chapters get drafted separately. The plateau is more common in fiction because middles are long. The wrong opening and the explaining ending are equally common everywhere.

My beta readers didn’t mention any of this.

They wouldn’t, usually. Readers report experience rather than diagnosis: “it was slow at the start,” “I lost interest around the middle,” “the ending felt a bit much.” Those three comments are problems 1, 3, and 6, reported accurately in the only vocabulary a reader has.

What if my manuscript has all six?

That’s a normal draft, not a bad one, and it’s the most common result of the chapter table. Fix them in the order they appear on this page because each one changes the material for the next: sorting out what the book is about will resolve some repetition on its own, and fixing the opening frequently fixes the plateau.

Does this mean I need a structural edit?

If the table shows several of the six and you can’t see how to fix them, yes. If it shows one or two and the fixes are obvious, do them yourself and spend the money farther down the sequence. Either way, you now know which conversation you are buying, which is more than most people know when they hire an editor.

Did you recognize your manuscript in more than one of these? That’s the normal result, and it’s a description of a draft rather than a judgment on a writer. If you’d like the same analysis done properly, my manuscript evaluation exists for exactly this, and my substantive editing is where the fixing happens. Or just tell me which of the six you found and I’ll tell you what it usually means.

error: