Last month I bought a novel with a 4.7-star average and eleven hundred reviews. The top review said they had stayed up until three in the morning because they could not set the book down. I made it to page forty. I have already forgotten what happened in those forty pages, which is its own kind of review.
This was not the first time. I have been burned on Amazon, on Audible, and on YouTube, by reviews that claimed the book was amazing and turned out to be anything but. It was somewhere around the fifth or sixth burn that I noticed that something in me had changed. I had stopped browsing. I had stopped taking chances. I was rereading books I already owned and already loved, because the only reviewer I still trusted was me.
That is a strange place for a reader to end up. It is a stranger place for a guy who runs a book app to end up. So let me walk you through how book discovery actually works in 2026, and then I'll tell you what I did about it.
The pipeline
Here is how a book acquires three hundred five-star reviews before a single buyer has paid for it.
The publisher sends out advance reader copies. ARCs, in the trade. (Self-published authors do this too. I am one. No stones will be thrown from inside this glass house.) The ARC goes to a reader who receives the book free, in exchange for "an honest review." The honest review arrives on launch day. It is almost certain to be four and half to five stars. It is nearly always this way, because the reader would like to keep receiving free books, and readers who post three-star ARC reviews are not contacted again. Nobody writes this rule down. Nobody has to.
There is a disclosure line at the bottom of these reviews. "I received a free copy in exchange for my honest opinion." I have come to read that sentence the way you read "paid actor's/ model's" at the bottom of a diet ad.
Then come the paid reviews, which are against the rules of every platform that hosts them, a fact that has inconvenienced the paid-review industry about as much as a No Soliciting sign inconveniences a door-to-door salesman. They arrive in batches. They are enthusiastic in a general way. They describe the book as a page-turner without mentioning a single page. If you have spent any time reading reviews at all, you have read hundreds of these, and some of them may have convinced you to buy.
The video problem
BookTube and BookTok were supposed to fix this. Real people, on camera, giving you their honest opinions. For a while, they did.
Then the publishers found them, and the sponsorship deals followed, and unboxing a publisher's ARC box is now its own genre with its own following. A creator who delivers a harsh review is likely to loose sponsors. Same rule as before. Still unwritten.
Plenty of creators are honest. I believe that. The trouble is that from where I sit, on my couch, holding my credit card, an honest opinion and a paid opinion look and sound the same. I know because I have acted on both, and some of them went from Amazon box to Goodwill box in less than an hour.
The fandom problem
There is a newer version of this, and it is worse, because it does not require anyone to be dishonest.
A large influencer announces a book. Sometimes they wrote it. Sometimes they merely blessed it. Their followers, who number in the millions and who would preorder a dirty tissue if it came from their favorite influencer, preorder it. Preorders all count toward the first week of sales, which is the week the bestseller lists measure, and so the book debuts on the New York Times list before most copies have been opened. The New York Times list counts sales, not quality.
Then the second wave arrives: ordinary buyers who saw "New York Times Bestseller" on the cover and assumed, reasonably, that it described the book rather than the author's follower count. Their reviews show up two or three weeks later, and they are all the same, barely contained outrage in a paragraph. At least one recent bestseller was declared more or less unreadable by the people who bought it on merit instead of loyalty, and by then the medal was already printed on the cover, where it will stay forever.
Of the bad review options, this is the worst sort to see through. The five-star reviews from the faithful are sincere. Nobody paid them. Nobody sent them an ARC. They meant every word. They are just not reviewing the book. They are reviewing the person.
The reread shelf
So I retreated. Of the last twenty books I finished, fourteen were rereads. I told myself this was comfort reading. It was risk management.
There is nothing wrong with rereading a book you love. There is something wrong when a reader who wants new books, and has the money for them, cannot buy one with any confidence, because its next to impossible for him to discern real hype from the artificial or biased.
What I built
I run Sage Chimp and I am a real reader, a true fan of books, so I built the feature I wanted as a reader: review separation.
When you look up a book, Sage Chimp reads the reviews and sorts them into two piles. Pile one is promotional: the ARC reviews with their disclosure language, and the launch-day pileups from accounts that praise in suspicious unison, whether they were paid to or would have done it for free. The generically enthusiastic go in there too. Pile two is what's left. People who paid money and took the time to share their real thoughts, including the cranky ones. Especially the cranky ones. A three-star review that says "great premise, sagging middle, I skimmed the last hundred pages" is worth more to me than fifty launch-day praise filled fluff reviews.
Then you get both scores, side by side. The promotional score and the human score. Sometimes they agree. When they don't, the gap tells you more than either number alone.
I tested this with real reviews until I was getting results I could trust. A recent read came out of those tests and, spoiler alert, I liked the book.
Nobody sent me a free copy.