I lost faith in algorithms a long time ago. The most revealing part of Amazon isn’t the one-star reviews. It’s the questions.
“Why was I recommended this book?”
That single line says everything. People are using Amazon to complain about Amazon. The irony would be funny if it weren’t so depressing.
I see the fallout from this regularly in my bookshop. Readers come in exasperated because digital recommendations have led them astray. They bought one thriller for their mother’s birthday and now their feed is convinced they are obsessed with serial killers.
The problem is simple: the algorithm doesn’t understand intent.
It notices what you buy but not why. It records the data but not the feeling. It doesn’t know the difference between “I loved this” and “I’m never doing that again.”
When someone tells me a book made them cry, I learn something no algorithm ever could. I discover they read to feel, not just to know. That insight can’t be automated. It comes from listening, from conversation, from years of talking to readers about what moves them and why.
A five-star rating tells you nothing about whether someone was weeping, laughing, or hurling the book across the room. It’s just a number, and numbers don’t tell stories.
The feedback loop that shrinks your world
Even the scientists admit it now. Researchers at DeepMind have found that recommendation systems trap users in ever-narrowing bubbles. The more accurately an algorithm predicts your tastes, the less you are allowed to surprise yourself.
Accuracy becomes a prison.
I run several book clubs, and members often tell me, “I’d never have picked this book
myself.” That is the whole point. I’m not trying to confirm their existing tastes. I’m trying to expand them. The joy of reading lies in discovery, not repetition.
To do that, you need a human touch. You need someone who understands that a story isn’t just a product. It’s an emotional experience that can comfort, provoke, or change you.
While the internet doubles down on algorithms, something remarkable is happening in the real world. More and more independent bookshops are opening here and in the USA.
That’s not nostalgia. That’s a correction. Readers are voting with their feet. They are
choosing trust over automation, conversation over code.
The same thing is happening in my own world. Our book clubs keep growing. Members don’t leave; they bring their friends. Because once people rediscover the pleasure of being understood, not just targeted, they don’t want to give that up.
It’s proof that the most powerful recommendation engine is still a human one.
Here’s the thing: I’ll never recommend a book just to make a sale.
That is the fundamental difference between what I do and what an algorithm does.
Algorithms are built to maximise transactions. I’m trying to build relationships.
If I choose not to recommend something, it’s usually because I’m thinking long-term, about the trust of the reader who will come back next week, next year, or even next decade. That is how loyalty is built.
Algorithms can’t replicate that. They optimise for clicks. I optimise for connection.
So when I read those Amazon reviews, “Why was I recommended this?”, I see more than
confusion. I see a quiet rebellion.
It’s a sign that readers are waking up to what they’ve lost: serendipity, surprise, and the
magic that happens when a real person puts the right book in your hands at exactly the right time.
Serendipity isn’t dead. It has simply found its refuge again, on the shelves of independent
bookshops, where humans still curate with care, intuition.
