Feedback, revisions, and knowing when a cover is done
At some point you have a cover you mostly like and no idea whether it is any good. This is the stage where most self-published covers get worse, because the author asks the wrong people a question that has no useful answer.
The question is not "do you like it." Nobody's taste predicts whether a stranger will click. The question is "what kind of book is this, and would you pick it up?" — and that one works on anybody, including someone who has never heard of you.
Who to ask
Readers of your genre. The only group whose reaction means anything. They have seen two hundred covers in your category this year and have working instincts about what belongs there, even if they cannot explain them. Find them in genre reader groups, review communities, or the comments under books like yours.
Not friends and family. They are answering a different question — whether they like you — and they will say it looks great. This is not dishonesty; they cannot see the cover without seeing you behind it.
Not other authors, first. Authors are useful later, on execution: the type is too tight, the spine is off, that face is dated. But they read a cover as designers, not as shoppers, and they will give you craft notes when what you need is a market read. Ask readers first, authors second.
Strangers are fine. Better than fine — a stranger's guess about what genre your book is, made in one second with no context, is the single most valuable piece of feedback you can get.
How to ask
Show it small. Send a thumbnail, not the full-size file. If you show a 1,900-pixel cover, you get feedback about details nobody will ever see. See The thumbnail test.
Show it beside other books. Put your cover in a grid of nine real covers from your actual sub-category and ask which one they would open. A cover shown alone is graded against nothing; a cover in a grid is graded the way it will really be encountered.
Ask what it says the book is. "What genre is this? What's it about? Who is it for?" If they guess literary fiction and you wrote a cozy mystery, you have found a problem no amount of polish fixes.
Never explain the design first. The moment you say "the lighthouse represents her isolation," you have given them information no reader will have, and their answer is now worthless. Show it. Say nothing. Let them talk first.
In the Cover Maker. The cover check renders your real cover — title, author and all — at the two sizes stores display, plus a grayscale panel, so you can hand people the version they would actually see rather than the one on your screen. Start a book → · about the Cover Maker
Reading contradictory answers
You will get conflicting feedback. That is normal, as long as you know what to do with it.
Look for the pattern, not the vote. One person disliking the blue is taste. Five people out of eight failing to read the title is data. Count how many independent people hit the same thing.
Separate reaction from prescription. People are reliable about what they felt and unreliable about how to fix it. "It looks cold" is worth acting on. "You should make it orange" is not — it is their attempt to solve your problem with none of your constraints.
Discount taste, keep function. Whether someone likes a typeface is noise. Whether they could read it is signal.
Design by committee is the specific trap. If you try to satisfy every note, you end up with a cover that offends nobody and attracts nobody — the compromise is always toward the generic, because that is the only place all the opinions overlap. Collect feedback widely, then decide alone. One person makes the call. That person is you.
The finishing checks
When you think it is done, run these five. They are pass or fail, not matters of opinion.
- The title reads at thumbnail size. About 160 px tall, seen from normal distance, by someone who does not already know what it says.
- It survives grayscale. Desaturate it. If the title disappears, you were relying on hue contrast instead of value contrast, and it will disappear in a search result too — Color and contrast.
- It sits correctly in its genre. Beside the top covers in your sub-category it looks like a neighbor, not a visitor — Genre conventions: what your cover has to promise.
- Type is inside the safe area. Everything that must survive the trim sits about 0.25 in inside the trim line, and artwork runs a full 0.125 in past the outside edges — Bleed and safe area, explained once.
- The spine matches the real page count. Not an estimate. The finished interior's page count, on the printer and paper you are actually using — How to calculate spine width.
Pass all five and the remaining questions are taste. Then order one physical proof and look at it in daylight before you announce anything — Order the proof. Then read it in daylight.
The honest part
There is a point where a cover stops improving and starts merely changing. You move the title up four pixels, then back down. You try a warmer blue, then the old blue. The version from Tuesday was fine and so is this one.
That point arrives earlier than most authors believe, and staying past it costs months the book spends unpublished while you rearrange things nobody will notice.
If it passes the five checks, reads clearly in its category, and you are no longer finding problems, it is done. The real feedback — the kind that comes from readers with money — only starts once the book exists.
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: Order the proof. Then read it in daylight · More: Choosing a title that can be found and remembered · Mistakes that mark a book as self-published