Testing a cover with readers
There is one question you must not ask, and it is the only one most authors ask: do you like it?
Friends and family answer that question socially. They are not evaluating a cover; they are managing a relationship with you. The answer is warm, fast and worth nothing. Worse, it feels like validation, so you stop testing and ship.
A cover has a job — tell the right reader, in about a second, at about the size of a postage stamp, what kind of book this is and whether it is for them. So test the job, not the taste.
Ask questions that have wrong answers
Every useful question is one you could fail. Show the cover small, for a second or two, then take it away and ask:
What kind of book is this? The single most valuable question. If they say "some sort of thriller?" and you wrote a cozy mystery, you have found a real fault before it cost you anything.
Who is it for? Age, reader, occasion. A cover that promises a gift book when you wrote a serious history is mis-addressed.
What do you think it is about? You are not looking for the plot. You are looking for whether the impression is in the right area.
Would you expect it to be funny or serious? Tone is the thing covers get wrong most quietly, and readers can answer it instantly even when they cannot say why.
Can you read the title? Ask them to say it back. Not to spot it — to read it, at the size you showed it.
Then the one that predicts behavior rather than opinion: put your cover in a grid of eight real covers from your category and ask which three they would click. No explanation, no context, a few seconds. This is the closest thing to the actual moment of purchase you can stage, and it is brutal in a useful way. The thumbnail test explains why that grid is the honest setting.
Who to ask
Readers of your genre who do not know you. Both halves matter. Someone who does not read fantasy cannot tell you whether your fantasy cover reads as epic or grimdark, and someone who knows you cannot separate the cover from you.
Where to find them: a reader group for your genre, your newsletter list if you have one, a forum or subreddit where that genre's readers talk to each other, a librarian, a bookseller who stocks that shelf, or a stranger in a coffee shop who is holding a book like yours.
Not other authors. This is the trap that catches experienced writers. Authors critique design — they will tell you about the kerning, the stock photo, the drop shadow. Useful sometimes, but they are responding as makers, not buyers, and they will talk you into a cover that other authors admire and readers scroll past.
Not designers, unless you are asking a design question. A designer will improve your execution. Only a reader can tell you your promise is wrong.
How to show it
The setting decides the answer, and most authors rig it without meaning to.
At the size a reader will see it. Not full-screen on your laptop. Roughly a centimeter and a half tall, in a list.
On a phone. That is where the browsing happens, and a phone screen is less forgiving than any monitor.
In a grid. Never alone. A cover alone is compared to nothing; a cover in a grid is compared to its competition, which is what actually happens.
For a second or two. Show it, take it away, then ask. A cover that needs fifteen seconds of study has already lost the reader it was made for.
One version at a time for the job questions. Save side-by-side comparison for the grid click test. Asking someone to choose between your two covers turns them into a designer, and you will get taste back instead of recognition.
In the tools. Previews are free, unlimited and watermarked, so making three versions to test costs nothing but your time — and the cover check renders your front at the sizes stores really display it. Mockups put the finished wrap on a physical book, which is what people respond to most naturally when you hand them a phone. Start a book → · about the tools
Reading the answers
Collect responses before you react to any of them. The instinct is to fix whatever the last person said, and that produces a cover shaped by whoever spoke most forcefully.
Look for the pattern, not the loudest voice. One person who strongly dislikes your color is one person's taste. Five people who guess the wrong genre is a design fault, and it does not matter that none of them complained.
Count the wrong answers, not the objections. Your tally is: how many read the title, how many named the right genre, how many would have clicked. Everything else is commentary.
Discount anything about detail. "I'd make the author name bigger" is a preference. "I couldn't tell what it was" is data.
A split is information too. If half read it as romance and half as women's fiction, you are sitting on a boundary — which may be fine, or may be why the book is hard to categorize. See Comp titles: reading the shelf you want to be on.
When to stop
Stop when the answers are consistent and boring. When ten people in a row name the right genre, read the title, and describe the book roughly as it is, the cover is doing its job and further testing will only produce new opinions about taste.
That is also the moment to stop revising. More rounds after that point tend to sand off the one distinctive thing the cover had. Feedback, revisions, and knowing when a cover is done covers the rest of that decision.
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: Feedback, revisions, and knowing when a cover is done · More: The thumbnail test · Mockups: showing a book that does not exist yet