Joe: What if I told you that the two most successful novels of the last two decades, Fifty Shades of Grey and The Da Vinci Code—one a steamy romance, the other a religious thriller—are, according to a supercomputer, basically the same book? Lewis: Wait, what? How is that even possible? One is about a very specific kind of… interior decorating, and the other is about Tom Hanks running through Paris with a different haircut. They couldn't be more different. Joe: On the surface, you're absolutely right. But underneath, they share a hidden DNA, a secret rhythm. That wild idea comes from a book called The Bestseller Code by Jodie Archer and Matthew L. Jockers. Lewis: The Bestseller Code. I like the sound of that. It’s like they’ve found the cheat codes for literature. Joe: Exactly. And what's so cool is the authors themselves are a perfect pairing for this question. Archer comes from the traditional publishing world, she was an editor at Penguin. Jockers is a data scientist from the Stanford Literary Lab. It's art meeting science head-on. Lewis: Okay, so you have the literary insider and the tech wizard. What did they do, lock themselves in a room and argue? Joe: Pretty much, but with a supercomputer. They fed thousands of novels—bestsellers and non-bestsellers—into an algorithm they built, a "bestseller-ometer," to see if it could learn to spot a hit. And their model can predict a New York Times bestseller with about 80% accuracy. Lewis: Eighty percent? That’s terrifyingly high. Publishing houses have been trying to do that for centuries with, you know, human brains. Joe: And they often get it spectacularly wrong. Which brings us to the first big idea: the problem this algorithm is trying to solve.

Lewis: Right, because for every blockbuster hit, there are stories of the ones that got away. Joe: Exactly. Think about it. J.K. Rowling’s Harry Potter was rejected by twelve publishers. Kathryn Stockett’s The Help was turned down by sixty agents. Sixty! John Grisham, the king of the legal thriller, was rejected by at least sixteen publishers. The traditional method is based on gut feeling, market trends, an editor's personal taste... and it's clearly fallible. Lewis: It’s more art than science, and sometimes the art is just a wild guess. So this bestseller-ometer is supposed to replace that gut feeling with cold, hard data? Joe: That's the provocative idea. The machine isn't "reading" for enjoyment. It's not sitting back with a cup of tea, pondering the prose. It's scanning for thousands of features: the frequency of certain words, the complexity of sentences, the use of punctuation, the emotional arc of the plot, the dominant themes. It’s a "distant reading" of 20,000 novels at once. Lewis: Hold on, but isn't art subjective? How can a machine measure 'good writing'? That feels a little… dystopian. Like we’re just going to start getting novels optimized for algorithmic approval. Joe: That's the crucial distinction the authors make. The algorithm isn't measuring "good" or "literary." It's measuring "bestselling." It’s identifying the patterns that a mass audience, for whatever reason, finds unputdownable. The two are not always the same thing. Lewis: Ah, okay. So it’s not a critic, it’s a bookie. It’s just calculating the odds of commercial success. Joe: A very, very good bookie. And this has been controversial. Critics of the book worry that if you crack the code, you just encourage people to write to a formula. You kill originality and get a thousand clones of The Da Vinci Code. Lewis: Which, to be fair, sometimes feels like what’s already happening. But I’m fascinated. If the machine isn't looking for what a literary critic looks for, what is it looking for? What’s in this secret code? Joe: Well, that’s where it gets really interesting. Because the ingredients for a bestseller are not at all what you’d expect.

Lewis: Alright, I'm ready. Lay it on me. What did the machine find? What's the secret code? I'm guessing it's just a formula for the optimal number of explosions and love triangles. Joe: You'd think so. The common wisdom is "sex, drugs, and rock and roll" sells. But the algorithm found the exact opposite. Those topics are statistically less likely to be in a bestseller. In fact, the presence of explicit sex actually goes down in the bestselling category. Lewis: No way. After the whole Fifty Shades phenomenon? How can that be? Joe: We'll get to Fifty Shades in a second, because it’s the perfect example of the rule. But first, the algorithm found three main ingredients that are far more important. The first is theme. The single most dominant, most predictive theme in a bestseller is, of all things, human closeness. Lewis: Human closeness? Not espionage, not murder, not magic? Just… people connecting? Joe: People connecting. The machine found that books about relationships, family, belonging, and community are consistently at the core of bestsellers. It’s not about the setting—a law firm or a dystopian future—it’s about the human drama within it. The authors call John Grisham and Danielle Steel the "godparents" of the bestseller for this reason. They are masters of grounding their high-stakes plots in relatable human connection. Lewis: That makes a strange kind of sense. You come for the courtroom drama, but you stay for the characters' relationships. Okay, what's the second ingredient? Joe: The second is plot. And this brings us back to your question about Fifty Shades of Grey and The Da Vinci Code. The algorithm found that bestsellers have a very specific emotional rhythm. A "perfect curve." It’s a three-act structure with a perfectly symmetrical rise and fall of emotional tension. Lewis: So it's like a pop song. It's not about the specific lyrics, it's about the verse-chorus-verse structure that gets stuck in your head. Joe: That is the perfect analogy! The machine plotted the emotional highs and lows of thousands of books, and found that Fifty Shades and Da Vinci Code, despite their wildly different subjects, have an almost identical emotional heartbeat. The pacing of conflict and resolution, the distance between a moment of despair and a moment of triumph, is mathematically regular. It’s this rhythm that creates that "page-turner" feeling. It’s addictive. Lewis: Wow. So E.L. James and Dan Brown are, in a way, master composers of emotional rhythm. That’s a much more generous take than I’ve heard before. Okay, theme, plot... what's the third ingredient? Joe: Style. The machine found that bestselling style is incredibly consistent. It’s informal and highly conversational. Bestsellers use way more contractions—words like "it's" and "don't." They use simple, direct language. They ask more questions. And they almost never use words like "very." Lewis: Why not 'very'? Joe: The authors quote the classic writing guide The Elements of Style, which calls 'very' a "leech" that sucks the life out of prose. Bestselling authors tend to use stronger, more direct words instead of qualifying weak ones. It’s about confidence in the language. Lewis: So, the secret code to writing a bestseller is to write about relatable human connection, with a plot that has the rhythm of a pop song, and in a style that sounds like you’re talking to a friend. Joe: In a nutshell, yes. It's less about a shocking premise and more about flawless, almost invisible, execution of these fundamental elements. Lewis: Okay, this is blowing my mind. But it leads to the big, scary question. If we have the code… can a computer just write a bestseller now? Are human authors about to be out of a job? Joe: The authors dedicate a whole section to that. And the answer, they found, is a resounding no. The machine revealed one more thing: the ghost in the machine. The one thing it can’t replicate.

Lewis: The ghost in the machine. I love it. So what is it? What’s the one thing the algorithm can’t crack? Joe: The author. The unique, inimitable human being behind the words. The best case study for this is what happened with J.K. Rowling. A few years after Harry Potter, a debut crime novel called The Cuckoo’s Calling by an unknown author named Robert Galbraith came out. It got good reviews, but didn't sell much. Lewis: I remember this! The big reveal. Joe: Exactly. A reporter got a tip that Galbraith was actually Rowling. To prove it, they sent the book to a computational linguist, who ran it through a stylometrics program—basically a cousin of the bestseller-ometer. The program analyzed the book's "linguistic fingerprint." Lewis: Linguistic fingerprint? What’s that? Joe: It’s the idea that every author has unconscious habits of style. It’s not about the big words, it's about the little ones. How often you use the word "the" versus "a." Your preference for commas over semicolons. The average length of your sentences. These things are almost impossible to fake. And sure enough, the computer confirmed it. Galbraith's fingerprint was a perfect match for Rowling's. She had been unmasked by an algorithm. Lewis: That's incredible. So even when she was actively trying to sound like someone else—a male, ex-military crime writer—her own unique style bled through. Joe: It’s like your DNA. You can change your clothes and dye your hair, but your core code is still there. And that’s what the algorithm can’t generate. It can recognize the pattern, but it can't create the unique fingerprint. The authors tested this by looking at early AI-written novels. Lewis: Oh, this is going to be good. What does an AI-written novel sound like? Joe: Hilariously bad. They mention a 1952 experiment by computer scientist Christopher Strachey, who programmed a machine to write love letters. One of the outputs was: "HONEY DUCK, MY PRECIOUS LOVE SEDUCTIVELY PANTS FOR YOUR CRAVING. YOURS IMPATIENTLY, M.U.C." Lewis: (Laughing) Okay, I think romance authors are safe for now. "Seductively pants for your craving" is not going to be a bestseller. Joe: Not yet, anyway. The point is, the machine can assemble words in a grammatically correct way, but it has no lived experience. It has no heart. It doesn't know what love or loss or "human closeness" actually feels like. It can identify the ingredients, but it can't cook the meal. Lewis: That’s a great way to put it. So the machine can be a great editor or a great analyst, but it can't be the artist. It can tell you if the song is catchy, but it can't write the song. Joe: Precisely. The human author, with their messy, unique, inimitable fingerprint, is the ghost in the machine. And for now, that ghost is irreplaceable.

Lewis: Wow. So this whole journey into the data, into the algorithm, it doesn't end with a magic formula. It ends with a deeper appreciation for the human element. Joe: That's exactly it. I think the book's biggest revelation isn't a formula for hits. It's a mirror. The algorithm shows us what we, as a culture, are collectively drawn to. And it's not shock and awe, or intellectual complexity. It's stories of human connection, with a rhythm that matches our own emotional pulse. Lewis: The machine holds up a mirror and shows us our own heart. It’s not about making books more robotic; it’s about understanding what makes them more human. Joe: And it reveals that the most successful stories are the ones that master the invisible architecture of emotion. The ones that feel effortless are often the most exquisitely constructed. Lewis: It makes you wonder, if this algorithm can decode our collective reading taste, what else can data reveal about our hidden desires? It’s both exciting and a little unnerving. Joe: It is. And it’s a great invitation for all of us. The next time you pick up a book you just can't put down, take a second to look under the hood. See if you can feel that steady, rhythmic beat. See if you can spot the invisible threads of human closeness holding it all together. Lewis: I definitely will. It’s like being given a new lens to see stories through. Joe: So, we encourage everyone to think about their own favorite page-turners. Do they fit the Bestseller Code? Let us know what you find. Share your thoughts with us online. Joe: This is Aibrary, signing off.