Pre-Publication Content QA Checklist

The document-integrity audit process used before any MMP manuscript goes to print/ebook conversion. This is a technical/structural QA pass — it does not replace an editorial read for content quality, but it catches a class of defect that’s invisible on a normal read-through and expensive to fix after approval (IngramSpark charges a $25 revision fee for any change after eProof approval).

Why this matters

Every defect category below was actually found on a real MMP manuscript during setup — none of these are hypothetical. A normal cover-to-cover read misses all of them, because they’re structural/metadata issues, not prose issues.

1. Placeholder & leftover-text scan

Search the full extracted text for:

• TODO, TBD, XXXX, lorem ipsum, [PLACEHOLDER, FIXME, COMING SOON (case-insensitive)
• Bracketed placeholder patterns: \[[^]]{1,60}\] — review every match. Legitimate matches (editorial insertions like [altar], math notation like [(3/2) + (4/3)]) are fine; anything that reads as an unfilled field is not.
• ISBN placeholders specifically — search for ISBN and check every instance has a real number, not [978-XXXXXXXXXX] or similar. Check that the label matches the actual format (a number under “Paperback ISBN” must actually be the paperback ISBN, not a mislabeled ebook one — this exact mix-up happened and needed correcting).
• Double spaces, duplicate consecutive words (\b(\w+)\s+\1\b).

2. Word-document structural integrity (if source is .docx)

• Blank/empty heading-styled paragraphs. Extract every paragraph with a Heading1/2/3 style and check for empty ones (no visible text but the style still applied). These are invisible in a normal read but will:
  • Show as blank entries in Word’s Navigation Pane
  • Produce blank lines if the document’s TOC field is ever refreshed (F9)
  • Risk inserting stray blank chapter breaks if the file is later converted to EPUB (ebook conversion tools typically split chapters on Heading1)
  • A real instance: 10 blank Heading1 paragraphs sitting before three of four Part dividers (the first Part had none — a clear sign of an accidental copy-paste artifact, not intentional spacing).
• TOC field / bookmark integrity. If the document has a real Word TOC field ({ TOC \o "1-2" \h \z \u }), cross-check every PAGEREF _TocNNNNNN field against the corresponding bookmarkStart with the same name — any PAGEREF with no matching bookmark is a broken TOC link.
• TOC page numbers should be strictly increasing when read in document order — a regression indicates a real pagination problem, not just a formatting quirk.

3. PDF-level integrity (for a rendered/exported PDF)

• Folio drift check: extract the printed page-footer number from every page and compare it against the page’s actual position in the PDF. Zero mismatches should be the result — any drift indicates a pagination bug from the export process, not a content issue.
• TOC-vs-content cross-check: for every TOC entry and its claimed page number, verify the heading text actually appears on that page (watch for false negatives caused by a heading wrapping across two lines in the extracted text — verify an apparent miss by hand before flagging it as real).
• Leaked field-code / conversion-artifact scan: search for Error! Bookmark not defined, Error! Reference source not found, raw LaTeX commands, raw Markdown heading markers (^#{1,6} ), or literal Markdown emphasis markers — none of these should survive into a finished PDF.

4. EPUB-specific checks

See the EPUB Conversion & Validation guide for the full technical checklist (EPUB3 compliance, metadata attribute rules, image size caps, zip structure).

5. Front/back-matter accuracy checks

• Cover file color mode, checked per slot. The print cover PDF must be CMYK; the ebook cover JPG must be RGB — they are opposite requirements for two different files, not a single asset reused twice. Open each file and check its actual color mode rather than trusting the filename or where it came from (a design tool’s “for screen” export is RGB, which is right for the ebook slot and wrong for the print one). See the Print File Specifications guide.
• ISBN consistency across every file for the same edition — the docx, the exported PDF, and the EPUB should all show the same ISBN with the same correct format label. Files drift out of sync easily when edited in parallel (a Word file and a derived PDF/EPUB can silently diverge if only one gets updated).
• AI-use disclosure consistency — if the book contains a printed “AI Use” note (e.g. disclosing that diagrams were author-drawn and AI-finished, or that AI assisted with research/editing), the retailer-platform AI-disclosure checkbox (IngramSpark, KDP, etc.) must say the same thing. A mismatch between what’s printed in the book and what’s declared to the distributor is a real inconsistency to catch before submission, not just a formality.
• Cover copy vs. manuscript claims. Read the actual front-cover tagline and back-cover blurb against the book’s own stated scope (e.g., its “What This Book Does—and Does Not Claim” section, if present). Marketing copy drifting into a stronger claim than the manuscript itself makes (e.g., a “teaches us” / direct-lesson-transfer framing on a book that explicitly disclaims that kind of claim) is a real defect, not a style nitpick — it sets the wrong reader expectation before they open the book.

6. What NOT to do during this pass

• Don’t fabricate placeholder review quotes/blurbs to fill an empty field — leave it empty until real ones exist.
• Don’t guess at undecided business fields (price, publication date, wholesale discount) — flag them for the author’s explicit input rather than picking a plausible-sounding default.
• Don’t treat this checklist as a substitute for an actual editorial/content read — it only catches structural and metadata defects, not writing-quality issues.

Reusable pass order for the next title

1. Extract full text (docx → plain text via python-docx or direct XML parse; PDF → per-page text via pypdf).
2. Run the placeholder/leftover-text scan (§1).
3. If docx: check blank headings + TOC/bookmark integrity (§2).
4. If PDF: check folio drift + TOC-vs-content cross-check + leaked-artifact scan (§3).
5. If EPUB: run the full EPUB checklist from the EPUB guide (§4).
6. Cross-check ISBN and AI-disclosure consistency across every file for the edition (§5).
7. Read cover copy against the manuscript’s own stated claims/scope (§5).
8. Only then move to final IngramSpark/KDP approval — remember post-approval revisions cost $25 each on IngramSpark.