There is a particular silence that follows the moment you attach a manuscript and hover over the send button. Everything you meant to do is still in your head; everything you actually did is in the file. The gap between the two is where submissions quietly fail — not because the writing is weak, but because a chapter repeats a point made forty pages earlier, a quotation has no source, or an image belongs to someone else. A structured edit before submission is not vanity. It is what lets an editor read your argument instead of your mistakes.
Cover the rest of the manuscript and read only the chapter titles and subheadings. A reader should be able to see the shape of the argument from that page alone: where it begins, where the evidence sits, what the conclusion rests on. If the sequence only makes sense because of explanations buried in the text, the contents page needs work. So does the structure.
Summarise each chapter in a single sentence, with no adjectives and no "and". Write the sentences out in order and read them as a list. If two of them say nearly the same thing, one chapter is surplus. If one reads like "assorted material about X", that chapter has no job yet and will read as filler. Watch, too, for the promise made in the introduction: a book about a town's industrial decline should not save the explanation of that decline for the epilogue.
Direct quotations need a source; so do distinctive arguments, figures and paraphrases that are not your own thinking. Check that each note gives enough information for a reader to find the source, and that page numbers are right — do not trust memory here. Reference lists and notes should match each other in both directions, with nothing in the bibliography that is never cited.
Style matters less than consistency at this stage. Choose one system, apply it throughout and be prepared to convert it later. A publisher's house style will almost certainly differ from yours, and no editor will hold that against you. Inconsistency inside your own manuscript, however, suggests the text has not been read closely.
Make a permissions log before you submit, even if you have not cleared anything yet. Include:
Rights holders can be slow to reply, hard to trace, and sometimes charge a fee, so start early. The rules around fair dealing and copyright vary by territory, so if a piece of material matters to your book, ask the publisher how they would like it handled and take professional advice where the answer is not clear.
Non-fiction wobbles when the narrator is unstable. Look for shifts between "I", "we", "one" and "the reader", and decide which you want. Read a page from chapter two next to a page from chapter nine: do they sound like the same person? Check the places where you hedge out of habit — "it could be argued that perhaps" — and either commit to the claim or explain why you are holding back. That is a different thing from genuine uncertainty, which readers respect.
Pick one term for each key concept and use it consistently; swapping between "staff", "employees" and "workers" invites readers to look for differences that are not there. Note your decisions on spelling and punctuation — UK or US spellings, serial commas, how you treat numbers, dates, job titles and italics for titles — and keep quotations in their original spelling. A one-page style sheet saves hours of correction and tells an editor you are careful.
Do not reread your own draft and call it checking. Work backwards from the source instead: for each name, date, figure and quotation, go to the document, interview or dataset and confirm it. Prioritise the claims your argument depends on — a wrong date in passing is embarrassing, but a wrong central figure undermines the book. Anything you cannot substantiate should be cut, softened, or flagged in a note to the editor rather than left to chance.
Search the manuscript for "currently", "recently", "now", "last year" and "in recent years". Each one is a small time bomb. Check statistics against their publication date, confirm that institutions, job titles and place names are still correct, and consider whether a point tied to a news cycle will still read in two years. Where a figure will age, give the date alongside it.
Word count, file format, whether chapters go in one file or several, how notes are supplied — publishers ask for these things for practical reasons, and following them is the easiest mark of professionalism available to you. Beyond that, tidy the basics: consistent heading styles, no manual spacing or stray formatting, page numbers, double spacing, and a title page with your contact details.
Check the attachment is the version you edited, the address is right, and the covering note reads as though a person wrote it. If the guidelines ask for a synopsis, chapter outline or author biography, send all three. Name your files clearly — surname, title, date — rather than "final final v3".
Read the first chapter aloud; you will hear repetition and clumsy rhythm that your eyes have stopped noticing. Read the whole manuscript once on a different device, where the layout looks unfamiliar enough to wake you up. Ask one trusted reader for the three places they got bored or confused, and act on them. Then leave it a day, check the guidelines one final time, and send it. You will not feel certain — nobody does — but a manuscript that has been through these ten checks arrives as a considered piece of work, and that is what gets read.
Photo: cottonbro studio / Pexels
