Turning a finished print book into an ebook
You have a printed book. It is done, it looks right, and now you want the ebook. The file sitting on your desk is the print PDF, so that is the file you feed to a converter.
That is the wrong move, and it is the first move most people make.
A PDF has appearance, not structure
A print PDF does not contain your book in any sense a reading system can use. It contains instructions for placing marks on a fixed page: this glyph at this coordinate, that glyph 4.2 points to the right of it. Nothing in the file says "this is a chapter title" or "this word is emphasized." The meaning was consumed in the act of setting the page.
A converter reading that file has to guess, and it guesses badly. What comes out has a recognizable set of faults.
- Line breaks frozen in place. Every line of the printed page becomes a hard break. On a phone the text wraps early and ragged, mid-sentence, in a pattern that follows the shape of a 6 × 9 page nobody is holding.
- Hyphens stuck mid-word. Your justified print setting hyphenated words at the end of lines. In the PDF those are just hyphen characters. They survive the conversion and now sit inside words in the middle of a reflowed line: "compre- hensive."
- Italics lost or scattered. Emphasis in print is a different font resource, not a tag. Converters often drop it entirely, or apply it to a run that starts and ends in the wrong place.
- Running heads and page numbers stranded in the text. The book title, the author name and the folio were drawn on every page. To a converter they are simply more text, so they land in the middle of your prose, once per printed page, forever.
- Structure gone. No real headings, so no navigation. Chapter openings become an ordinary paragraph in a larger size, which builds nothing a reader can jump to.
Every one of these is fixable by hand. Fixing them across 300 pages takes longer than starting again from the right file.
The right source is the manuscript
The print book was set from something. Find that: the .docx, the Scrivener project, the Word file your editor returned. That file still has the structure the PDF threw away — chapter headings as headings, italics as italics, paragraphs as paragraphs, and no hyphenation, because hyphenation is a decision made at typesetting time and not before.
Build the ebook from the manuscript, not from the print output. If the manuscript has drifted — corrections made at proof stage that never went back into the source — reconcile it now, once, and keep it as the single copy that is true. Preparing a manuscript that typesets cleanly covers what a clean source looks like.
In the Interior Formatter. It reads a .docx for structure only — chapter starts, italics, bold, scene breaks — and produces the print PDF and the EPUB 3 from that same source, so neither file is ever made out of the other. Start a book → · about the Interior Formatter
What changes even from a good source
A clean manuscript gets you a sound file. It does not get you a finished ebook, because several things in a print book mean nothing in a reflowable one.
- Page cross-references. "See page 214" points at nothing. Rewrite as a chapter reference, or make it a link.
- The index. A list of page numbers is dead weight. Either convert the entries to links or drop the index and rely on the reader's search. Does your nonfiction book need an index? covers the judgment call.
- Front matter. A half title exists because a printed book needs a page before the title page. Blank versos exist because leaves have two sides. Neither has a reason to be in an ebook. Trim to the title page, copyright page and anything a reader actually reads. What goes in the front of a book, and in what order has the full order.
- The table of contents. In print it is a list with page numbers. In an ebook it is navigation, generated from real chapter structure, and the device builds its own menu from the same data.
- Images. Each one needs alt text, and each one needs to be sized for a screen rather than a press. A figure that sat at a fixed spot on a page will now land wherever the text carries it.
- Back matter. The also-by list, the newsletter signup, the review request — all of it becomes clickable. That is the one place the ebook beats the printed book outright. See The back of the book is where your next sale is.
When there is no digital source at all
For an older book, the manuscript may genuinely be gone. Then you scan the printed pages and run OCR, and you treat the result as a draft, not as text.
OCR is good now, and that is exactly the danger. It rarely produces obvious garbage. It produces plausible words: "com" for "corn," "modem" for "modern," "he" for "be." Those pass a spell check and they pass a skim. A full read against the printed pages is not optional, and italics, small caps, accented characters and anything in a table will need checking separately.
Budget it as a proofreading pass on the whole book, because that is what it is.
The rule that makes this disappear
Produce both from one manuscript in the first place. The print PDF and the EPUB are siblings, not parent and child, and neither is ever derived from the other. Do it that way and the conversion problem never exists — you just make the ebook.
Make yours
Everything here is built into the tools, and they are free to use: name your book, bring your manuscript or your artwork, and work as long as you like. You only pay when you want the files without the preview watermark.
Next: Why your ebook is enormous, and what to do · More: EPUB and print PDF are two different books · Links in an ebook, and where they should point · Bringing a book out of Google Docs, Pages or Scrivener