A reader scrolling Amazon is not reading your cover. They are reading your genre off your cover, at the size of a postage stamp, in about a second, and then deciding whether you are their kind of book before they have read a single word of the title. Choosing a book cover by genre is really choosing what promise a reader reads in that one second, and most covers spend it on the wrong signal.
The wrong signal in contemporary romance is usually the naked man. You know the one. Oiled torso, no head, slightly aggressive lighting, a font that is trying very hard. Authors reach for it because it reads “romance,” and it does read romance. The problem is that it reads a very specific romance: the 2015 billionaire-alpha, dark-and-brooding, mass-market KU version. If that is the book you wrote, carry on. If you wrote a warm small-town second-chance story with a himbo and a dog, you have just promised a reader something your book does not deliver, and romance readers are feral about a broken promise.
A book cover by genre is read at thumbnail size
This is the part authors skip. You spend weeks staring at the full-resolution file on a big screen, falling in love with a detail nobody will ever see, and you forget that the actual reader meets your cover at 150 pixels wide, next to nineteen other books, while half-distracted. If the genre does not land at that size, nothing else about the cover matters.
So the first test is the dumbest and most useful one I know. Shrink your cover down to a thumbnail, line it up next to five current bestsellers in your exact subgenre, and look at the row. Does yours obviously belong? Or is it the one that looks like it wandered in from a different shelf? You are not judging whether it is pretty. You are judging whether a tired reader, scrolling at 11pm, would clock it as the thing they already know they like.
Because that is what a reader is doing. They are not hunting for something new and strange. They are reaching for the brand they always buy, the one they know they like, and they want to spot it fast. Your cover’s job is to look like the thing they already reach for before it tries to be clever.
Contemporary romance: the illustrated cover is doing real work
The shift toward clean illustrated and cartoon-style covers in contemporary romance is not a passing aesthetic. It is a positioning signal readers have learned to read. An illustrated cover tells them this is character-forward, lower on explicit heat, more banter than brooding, the kind of book they can leave on the coffee table. The photographic clinch cover tells them the opposite. Neither is better. They are pointing at two different readers.
Where authors go wrong is picking the style they personally find prettier instead of the one that matches the book inside. I have watched a genuinely spicy book fitted with a sweet illustrated cover, which pulled in readers expecting closed-door and left them annoyed, and I have watched a soft romantic comedy wearing a dark photo cover that scared off the exact people who would have loved it. In both cases the reviews were not really about the book. They were about the mismatch.
I had a client once whose early reviews kept saying the book was light on historical detail. She read that as readers calling her book bad. It was not. It was readers telling her she had promised a historical and delivered something closer to a romance that happened to be set in the past. The fix was not a rewrite. It was positioning, starting with the cover and the shelf it pointed at. Bad reviews that cluster around a single complaint are almost never a quality problem. They are a signposting problem, and the cover is the biggest signpost you own.
Upmarket and literary fiction: let an object do the talking
The further you move from genre romance toward upmarket, book-club, thriller and literary fiction, the more the signal flips away from people and toward objects. Look at the covers that dominate the thriller and book-club shelves right now: a single stark object, a lot of negative space, a large legible title, a restrained palette. Freida McFadden’s thrillers and Ken Follett’s big historicals both lean on that object-and-title restraint rather than a photographed face, and it reads instantly as “serious, propulsive, for grown-ups.” That restraint is itself the genre signal. A busy, over-illustrated cover on an upmarket thriller reads as self-published-and-unsure, fairly or not, and that read costs you the exact reader who buys hardcovers and tells their book club what to read next. The object cover is not minimalism for its own sake. It is the look that particular reader has been trained to trust.
The rule underneath all of it: polished, and promising the right read
If you take one thing from this, make it this. There is no single correct cover style. There is only “does it look professionally made” and “does it promise the book a reader is actually going to get.” Illustrated, photographic, typographic, object-led, none of those is right or wrong in the abstract. They are right or wrong for your specific subgenre and the specific reader you are trying to reach.
It is always positioning.
The cover is where that positioning becomes visible before a reader has read your blurb, your reviews, or your first line. Get the genre signal right and everything downstream gets easier, because the people clicking are already the people inclined to buy. Get it wrong and you spend the whole launch fighting a first impression you set yourself.
A 2026 caveat. If a cover reads as AI-generated, a chunk of your early reader community will quietly refuse to share it, and ARC readers in particular are getting caught in that crossfire, reluctant to post about a book because the cover looks machine-made. Whatever your own position on the tools, that is a real reader reaction that affects real release-week visibility, so it belongs in the conversation with your designer.
What to actually brief your designer
The line I want to be clear about: how your cover gets designed is between you and your cover designer, and a good one is worth every cent. This post is not about teaching you to design. It is about what you hand them before they start, because the brief is where the genre signal is either set or lost.
A brief that sets the right signal usually covers:
Your exact subgenre, named precisely. Not “romance” but “small-town contemporary romance, low-angst, open-door.” Not “fantasy” but “cozy fantasy, found family, gentle stakes.” The narrower you name it, the better your designer can point the cover at the right shelf.
Three to five comp covers you are deliberately standing next to. Pull current, well-selling books in your exact subgenre and say “this neighbourhood.” This does more work than any adjective. It is the same comp thinking you should already be doing for your blurb and your categories.
The tone and heat level, signalled by style rather than skin. Tell your designer whether the book is sweet, steamy or scorching, and let them translate that into illustration-versus-photo, palette and type, instead of defaulting to how much flesh is on display.
Title legibility at thumbnail size, stated as a requirement, not a hope. Ask to see the cover at 150 pixels wide before you approve anything. If you cannot read the title and clock the genre at that size, it is not finished, no matter how good it looks full-screen.
Where the cover has to live. Amazon thumbnail, a print spine on a shelf, a tiny circle on a BookTok roundup, a newsletter header. A cover that only works at full resolution is a cover that fails in every place a reader actually meets it.
The quickest check you can run today
Pull your cover up as a thumbnail next to five bestsellers in your subgenre. If yours is the one that looks like a different kind of book, you do not have a taste problem and you do not need to start over from scratch. You have a brief to write. That row of thumbnails, and the single genre it is all quietly promising, is the brief.