Forget doom scrolling. Escape Love Island. With a nudge or two from AI, you could switch your entertainment addiction to a format that’s a few millennia older and far more nourishing: the humble book.
“AI could fundamentally change the way we market and sell books,” says Sara Lloyd, global head of AI at Pan Macmillan, one of the Big Five publishers in the UK and home to international author brands such as Ken Follett, Julia Donaldson and Elton John. “Machine learning and automation, now combined with the power of natural language processing, are opening up a world of possibilities that will enable the relatively smaller marketing resources of most publishers to punch well above their weight.”
It’s a punch that lands pretty hard. Business is booming for British publishers, despite continued reports that the toxic cocktail of social media and streaming will soon make their legacy product obsolete. Nielsen BookScan reports that the British book market achieved record sales of $2.27 billion (£1.83 billion) last year, the highest sales since records began, a coup fueled by blockbuster releases from Britney And Prince Harryhis memoirs Spare part is the fastest-selling non-fiction book in history. In contrast, 2023 was a tough year for both Twitterstorms and boxsets, with media executives claiming we had “Top television while Elon Musk continued his consistently impressive work Refueling X.
Backlist monetization
In this environment, the emergence of sophisticated, generative AI-powered marketing tools could be just the push publishers need to win back heads – and pounds – from their shinier competitors, and they have a powerful stealth weapon at hand: their backlists. Founded in 1843, Pan Macmillan has two centuries of old award winners, former bestsellers and underrated gems just waiting to be rediscovered BookTokers. The other four major companies – Penguin Random House, Hachette, Simon & Schuster and HarperCollins – have similarly extensive archives that they can use.
“Over 100 million [book] Titles were published in English alone,” says Nadim Sadek, the founder of Shimmer AI, a London-based “AI-native” advertising platform that has received sign-ups from more than 120 publishers since its launch six months ago, including both members of the Big Five as well as smaller independent publishers at a rate of around one per working day. “No more than 5% of them were ever advertised. Accordingly, they are largely under-monetized.”
Courtesy of Molly Flatt
Matchmaking flavors and titles
Shimmr uses a three-step process to turn a simple PDF or ePub file into a read you can’t resist. First, “the Analyzer” uses a variety of large language models (LLMs) to define the structural patterns and psychological values of the work, or “BookDNA.” Then “the generator” uses a series of diffusion models and other image generators to produce images and copies that are closely tied to that book DNA. Finally, “the deployer” shares the content across media channels in which readers with corresponding psychological dispositions are active.
“Reward models ensure that the campaign continuously optimizes itself and looks for the most effective combinations of BookDNA, images and text in different channels and for different target groups,” explains Sadek. “By matching them with more psychologically appropriate books, readers also enjoy more fulfilling reading, which positively reinforces their reading (and buying) habits.”
The results are already impressive. “Within just six months of our launch, Shimmr has doubled the ROAS (return on ad spend) of the average Google benchmark,” reports Brooke Dobson, Shimmr’s chief revenue officer, “and we are on track to exceed that by at least to improve threefold.” Shimmr’s self-optimizing system. Additionally, titles on Shimmr see an average increase in sales of 35%.”
It’s no surprise, then, that the company has attracted some big-name publishing investors and advisors, including former Macmillan CEO Richard Charkin and former Penguin Random House executive Makus Dohle.
Balancing innovation and ethics
But mainstream industry leaders are still cautious about admitting to experimentation with AI, and for good reason. Dozens of authors, including John Grisham and George RR Martin, are currently suing OpenAI and Meta for training LLMs on copyrighted works without permission or compensation, while counterfeit AI-generated content supplements the already meager earnings of authors and illustrators (on average, about $8,700 (£7,000) a year, at last count from author licensing body ALCS).
Lloyd, so far the only UK specialist publishing manager with AI in her title, is keenly aware of the tension between innovation and ethics. “Publishers must grapple with two parallel demands: on the one hand, they must innovate and harness the benefits of AI to make us more effective in selling our authors’ works; on the other hand, they must lobby and vigorously push for the use of AI to protect their copyright protected material is respected, acknowledged and compensated.”
Sadek also wants to highlight how Shimmr protects the creators it serves. “Books are hermetically separated from the training sets [our] models,” he emphasizes. “Each execution is built from an initial single pixel informed by the BookDNA. No pre-existing images are used as a starting point.”
However, Lloyd is also optimistic that bookstore leaders need to proactively embrace and shape AI. “Macmillan (and our parent company, Holtzbrinck Publishing Group) recognized the reach and influence of AI – and the power it is likely to have in the future – and recognized that a centralized role and cross-functional perspective was critical,” she explains. “The disruptive potential of AI means the publishing industry cannot risk falling behind.”
So next time you’re looking for a new series to enjoy and instead buy a decades-old novel that fits your tastes incredibly well, take a moment to celebrate. They are helping the publishing industry beat AI at its own game.