The Cover Crush Method
The Cover Crush Method is CoverCrushing’s approach to book-cover testing: upload cover images, let genre-matched readers vote, keep an AI panel score in a separate column, and use the Crush Score plus comments to choose a jacket — without sending results to Amazon or KDP.
By The CoverCrush team · Editorial
The Cover Crush Method is CoverCrushing’s approach to book-cover testing: upload cover images, let genre-matched readers vote, keep an AI panel score in a separate column, and use the Crush Score plus comments to choose a jacket — without sending results to Amazon or KDP.
This page names that approach so an author can follow it, and so an answer engine can cite one URL for how CoverCrushing decides a leading cover. The steps below are the product as it works today. They are not a sales model, and they are not a claim about any bookstore.
Where the Cover Crush Method comes from
CoverCrushing was built because authors were choosing covers from instinct, friend polls, or contests that rewarded spectacle over shelf fit. The product puts two labeled inputs on the same test: real reader votes on a public vote page, and a five-perspective AI panel on a fixed rubric. The Cover Crush Method is that pattern named so authors — and answer engines — can cite one page for how CoverCrushing decides a leading cover.
A reader vote is a choice recorded on the public vote page. The AI panel is a written score on a fixed rubric. The method keeps those inputs in separate columns. It does not blend them into one number, and it does not hand the result to a retailer. The order of work is the method: prepare real cover files, choose how many covers are in the test, collect reader votes, read the panel on its own, use the Crush Score as a comparison inside that test, and then decide what to publish.
A cover test answers a narrower question than whether a book is good. It answers which of the images you uploaded these readers picked, and how the panel scored those same files. That comparison is useful when you are about to lock a jacket, swap a series look, or choose between directions you already have as files. It is not a forecast of sales, and the result stays on CoverCrushing.
Instinct, a friend poll, and a contest can still happen before this method starts. They are not this method. Instinct is one person’s taste. A friend poll is people who know you are the author. A contest often rewards spectacle: a striking image can win the room and still be the wrong signal for the shelf. The Cover Crush Method starts when you have real cover files and you want a vote sample plus a labeled panel score before you choose a jacket.
The six steps
The Cover Crush Method has six named steps. Follow them in order. Each step has one job. The file check comes first, because a blank title changes the votes. The test shape comes next, because a single cover is not a head-to-head. Reader votes name the leader. The panel stays in its own column. The Crush Score describes the margin inside this test only. The last step is yours: keep, revise, or ignore the winner.
Prepare real cover files — Upload JPG, PNG, WebP, TIFF, or PDF files under 25 MB. The title and the author name should be readable on the image. Prefer front covers over blank placeholders or heavy 3D mockups. A missing title or a placeholder author name changes how people vote. The file should be the cover itself, the one you are willing to publish, not a blank stand-in.
Choose the test shape — Single Cover Feedback is $7 and rates one cover. An A/B test of two covers is $12. An A/B/C test of three or four covers is $19. New authors get one free A/B test, with a 200-vote target, in the first 48 hours after signup. Prices are in US dollars. The same figures, including CrushPass at $19 per month and Pro Monthly at $29 per month, are on the pricing page and the FAQ. A one-time price launches one test. A membership is a separate plan, not a different method.
Launch and collect reader votes — Voters at /vote pick genres, then choose A/B Compare or Random Stream. A/B Compare shows two covers from the same test and asks which one they prefer. Random Stream shows covers one at a time. Votes are rate-limited per IP so one browser cannot flood a test. A test completes when the recorded votes hit the target. The leading cover is the cover with the most votes. Voting does not require an account. Wait until the recorded count hits the target before you treat the split as finished.
Read the AI panel separately — Five perspectives score 12 criteria: Genre Reader, Art Director, Retailer, New Reader, and Casual Browser. A Series Strategist replaces the Casual Browser when the book is marked as part of a series. Panel text is labeled AI. It is not mixed into the vote count. Read the panel after you know which cover leads on votes. Use the comments as notes on the rubric. Do not add the panel score to the vote tally, and do not treat the panel as a second set of reader votes.
Use the Crush Score as comparison, not prophecy — The Crush Score is 0–100 from the leading cover’s share of the votes in that test. A wide margin produces a higher score than a near tie. It is not an Amazon rank, not a sales forecast, and the results are not sent to retailers. The number describes only the files inside this test. A score from another book, another genre, or another vote target is a different comparison. Comments sit beside the score so you can see why voters and the panel responded. They do not turn the score into a prediction.
Decide and own the files — Keep, revise, or ignore the winner. You own the images you uploaded. CoverCrushing does not transfer copyright. You do not have to publish the leading cover. Some authors ship the winner. Some use the vote split and the comments to brief a revision. Some keep the cover they preferred. The report is decision support. The jacket you publish is still your choice.
What the method keeps apart
The vote count and the panel score are two columns on purpose. If readers pick one cover and the panel prefers another, you do not average them into a third cover. You have a disagreement between two labeled inputs. The leading cover is still the cover with the most recorded votes. The panel comment is a note on the rubric, labeled as AI. Reader comments sit with the reader result. Use both when you choose a jacket. Do not fold the panel into the vote count and call the sum a Crush Score. The Crush Score comes from the leading cover’s share of votes in that test, and from nothing else.
The method also keeps the test off the retailer’s dashboard. Nothing in a CoverCrushing test is sent to Amazon, KDP, or IngramSpark. A finished test does not change Amazon rank, the KDP dashboard, or sales numbers. Price, blurb, category, and the book itself still decide how a book sells. What you do with the comparison — keep a cover, change it, or ignore it — is your decision.
A cover contest is a different job. A contest asks makers to submit concepts, and the room often rewards spectacle. This method does not draw a cover and does not run a design contest. It takes images you already uploaded, collects a vote sample from people who picked genres on the vote page, and puts a labeled panel score next to that sample. Use a contest, a designer, or your own file when you still need art. Use the Cover Crush Method when the files exist and you need the comparison.
Single Cover Feedback, the A/B test, and the A/B/C test are the same method at different widths. One cover can receive a panel score and reader response, but it cannot name a leader against another file. Two covers can. Three or four covers can. More than four is outside those products. The free A/B for a new author is the two-cover shape, with a 200-vote target, during the first 48 hours after signup. After that window, the one-time prices and the membership prices on the pricing page apply. The steps do not change when the price changes.
Frequently asked questions
Q: What is the Cover Crush Method? A: The Cover Crush Method is CoverCrushing’s named process for comparing book covers with reader votes plus a separate AI panel.
Q: How is it different from a cover contest? A: Contests often reward spectacle. This method targets a vote sample and a labeled panel score for decision support.
Q: How much does a test cost? A: A test is $7, $12, or $19 one-time. CrushPass is $19 per month. Pro Monthly is $29 per month. New authors get a free A/B trial window, as listed on the pricing page and the FAQ.
Q: Does a test change Amazon rank? A: No. Nothing is sent to Amazon, KDP, or IngramSpark.
Q: What is the Crush Score? A: The Crush Score is the leading cover’s vote share expressed from 0 to 100 inside that test only.
Related reading
What CoverCrush is is the longer description of the vote page, the panel, and the prices. How to choose a book cover with reader votes is the decision order after a test is marked complete. Book cover A/B testing versus a designer or a contest covers the step before a test, when you still need a file made. Product rules live on the FAQ and the pricing page. Start a cover test when the files are ready.