Most indie authors don't ship bad books because they edited too little. They ship late — or never — because they edited too much. The difference between a polished manuscript and an over-polished one isn't quality; it's fear dressed up as professionalism. And the modern editing toolkit, powerful as it is, has made the problem dramatically worse. When you can run your prose through ProWritingAid, Grammarly, Hemingway, and then a paid developmental editor, the temptation is to keep cycling until every sentence feels "perfect." But perfection is a moving target, and chasing it has a real cost: months of lost momentum, a growing backlist that never materializes, and readers who would have loved Draft 7 never getting the chance to see it.
Here's what the tool-obsessed editing culture won't tell you: the returns diminish fast. Dave Chesson of Kindlepreneur — someone who is openly dyslexic and has relied on editing software for over a decade — makes a revealing distinction in how he actually works. Quick communications skip the tools entirely. Blog posts and book drafts get a second pass through ProWritingAid. But even he, someone with a genuine neurological reason to over-check, draws a line. He doesn't run everything through every tool. He matches the editing effort to the stakes of the piece. That instinct — knowing when something is done enough — is exactly what most indie authors with one to three books under their belt haven't developed yet. They treat a 70,000-word novel the same way they treat a query letter, applying maximum scrutiny to every line, every word, every comma, long past the point of meaningful improvement.
The new AI layer makes this worse, not better. Tools like ChatGPT and Claude can now restructure your paragraphs, suggest alternate phrasing, flag pacing issues, even mimic the tone of your favorite bestseller. That capability feels like a gift, but it introduces a hidden cost: each AI-suggested revision creates a new decision point, and each decision point is another opportunity to second-guess yourself. Meanwhile, the publishing industry is actively watching. Literary agents are now asking whether AI was used in submissions. Some publishers run contracted manuscripts through detection software like Pangram, which operates on probabilities rather than certainties. The Authors Guild launched a Human Authored Certification program — $10 per book for non-members — that permits AI for grammar and spell-check but draws the line at AI-generated text. It runs on the honor system. The point isn't that these guardrails are airtight. The point is that they exist because the industry recognizes a real tension between tool-assisted editing and authorial ownership. Every additional AI pass on your manuscript nudges you closer to a gray zone that didn't exist three years ago, and the anxiety about that gray zone becomes yet another reason to delay publication.
So when should you actually stop? I'd argue the answer is brutally simple: stop when the changes you're making no longer alter the reader's experience. Fixing a typo alters the experience. Catching a plot hole alters the experience. Rewriting a perfectly functional sentence because ProWritingAid flagged it at 98% readability instead of 99%? That's not editing. That's procrastination with a progress bar. A useful heuristic: if your last full editing pass changed fewer than two things per chapter that a reader would actually notice, the book is done. Not flawless. Done. Those are different things, and confusing them is the single most expensive mistake indie authors make with their time.
The compounding math favors shipping. An author who publishes three solid books in eighteen months will almost always outperform an author who publishes one immaculate book in the same window. Readership loyalty in indie publishing is built on volume and consistency, not on the absence of a single awkward adverb on page 214. Your backlist is your business. Every week you spend on revision number fifteen is a week you're not drafting the next book that could double your audience.
Here's your concrete next step: open your current manuscript, go to the last chapter you edited, and count the substantive changes from your most recent pass — not cosmetic tweaks, but changes that would genuinely alter how a reader understands or feels about the story. If that number is two or fewer, close the file, write your publish date on a sticky note, and put it where you'll see it every morning. The book is ready. You're the one who isn't, and that's a problem editing will never fix.