Amazon's Algorithm Didn't Promote My Book. Here's What I Should Have Done Instead.

I believed the myth: write a good book, upload it to Amazon, and the algorithm will do the rest.

There are millions of readers browsing Amazon, and the platform's sophisticated recommendation system pushes my book to fans of romance fiction (my book's genre). There is this "Customers who bought this also bought…" magic that would introduce my book to the right people.

I just had to publish and wait.

So I waited.

And no prizes for guessing the algorithm never showed up for me.

My book sat in Amazon's digital warehouse like a product on the bottom shelf in the back corner of a massive store. Sure, it was technically available, but it was practically invisible.

I thought Amazon's algorithm was designed to help good books find readers, but I was wrong.

Here's what the algorithm actually does, why it ignored my book completely, and what I should have done instead.

The Amazon Algorithm Myth

The myth sounds logical: Amazon wants to sell books. Your book is for sale on Amazon. Therefore, Amazon will help sell your book.

This is technically true but functionally meaningless.

Amazon will help sell your book if the algorithm has evidence that your book will sell, and without that evidence, you're on your own.

The algorithm doesn't discover hidden gems, like a person rummaging through an overcrowded bookstore. Instead, it amplifies and pushes what's already working.

If your book is selling, Amazon promotes it more. If it's not selling, Amazon ignores it.

I was waiting for the algorithm to kickstart my sales. But the algorithm only kicks in after you've already generated sales yourself.

What Amazon's Book Algorithm Actually Looks For

Amazon's recommendation system isn't magic. Like all good platforms, the algorithm is math.

Before it promotes your book, it looks for signals that a book is worth promoting, such as:

• Sales velocity: Is the book selling consistently? How many sales in the last 24 hours, 7 days, 30 days?
• Conversion rate: When people view your book page, do they buy? Or do they click away?
• Reviews: How many reviews does it have? What's the average rating? How recent are the reviews?
• Click-through rate: When Amazon shows your book in search results or recommendations, do people click?
• Category ranking: Are you ranking well in your categories? Top 100 in a category triggers more visibility.
• Also-bought data: Are people who buy your book also buying popular books in your genre?

Why the Algorithm Ignored My Book (And Probably Yours Too)

My book had none of these signals, having sold only 10 copies in the first 90 days of publication. It meant the algorithm (which we need to remember is AI, not a person personally selecting what sells) had no reason to promote it.

Like most failures on Amazon, it had: 

• Zero sales velocity. I launched with no audience and no marketing. Nobody bought the book in the first 24 hours. Or the first week. Or the first month. The algorithm saw a flat line, a dead product that was not worth promoting.
• Zero reviews. No social proof with nothing to show potential buyers that it was worth reading. The algorithm won't push a book with no social validation.
• Bad conversion rate. The few people who did find my book clicked away without buying. Because zero reviews = zero trust. And the low conversion told the algorithm, "This book doesn't convert, don't show it to more people."
• Over-competitive categories. I picked categories I thought fit my book. They were too competitive. I was ranking #250,000+ in those categories. The algorithm only promotes books ranking in the top 100 of their categories. I wasn't even close.
• No also-bought connections. Since nobody was buying my book, there was no data on what else my buyers purchased. The algorithm couldn't connect me to related books.

I was invisible to the recommendation engine.

What I Should Have Done Instead

Waiting for the algorithm was my fatal mistake. I was waiting for something completely out of my control, and not in my favour, to work in my favour and give me what I wanted. It was never going to happen, but foolishly, I thought I knew better. 

Some of it was hope, of course. Yet, in reality, most of it was publishing fatigue, sprinting to the publishing deadline without realising I was in a longer promotional marathon. 

Luckily, it was my first book, and I won't repeat these errors when launching the second book. 

Here's what I will do next time (and what I should have done the first time):

Strategy 1: Generate sales velocity yourself in the first 30 days

The algorithm needs evidence that your book can sell. You have to create that evidence by ensuring readers buy from you the moment you launch. 

Before the launch, I would:

• Build an email list (even 50–100 people) - which I'm already doing on Substack
• Build relationships in reader communities - like I'm currently doing on Facebook within reader groups (it helps I'm already a reader and engaging with these people as part of my passion)
• Line up advance reviews through conversations with reviewers and using paid review services

On launch week, I would:

• Email the list immediately - I would hype up the launch, letting them know the wait is over
• Post in relevant communities (without spamming)
• Run targeted ads to drive traffic
• Offer a limited-time discount to boost sales

The goal during that first week would be modest. I would aim to close 20–50 sales in the first week, which should be enough to demonstrate velocity.

That's just one option. Let's see what else I can do (and what you can do for your launch). 

Strategy 2: Optimise for conversion before you drive traffic

As part of my desperate sales strategy, I ran ads on Amazon. Hopefully, I thought, I could undo my mistake and drive sales that way. However, I was already setting myself up for failure because my ad pointed readers to a book page with zero reviews. 

Of course, nobody bought. Like any good product or landing page, you need to convert the traffic you drive. If I run ads next time (which I likely will), I will: 

Launch with at least 10–15 reviews. Use ARC services. Send advance copies. Do whatever it takes.

Write a compelling book description. Not a summary but a sales pitch. I would make people want it based on the description.

Choose the right categories. Pick winnable categories where I can rank the top 100, and avoid the most competitive ones.

Price strategically. Consider launching at $0.99 to build velocity and reviews. Even though the idea of giving away the book hurts, it needs readers before it makes money. 

Strategy 3: Win in narrow categories first

I picked broad, competitive categories like "Romance Fiction" and "Women's Fiction."

I was competing with thousands of books. Ranking #200,000+. To say I was invisible was an understatement, which is the case for all writers in my situation.

I should have picked niche categories where I could rank in the top 20, since the algorithm promotes books that rank well in their categories.

An example of this (with different categories, of course):

• Instead of "Self-Help," pick "Self-Help > Creativity" or "Self-Help > Motivational > Time Management"
• Instead of "Fiction > Thriller," pick "Fiction > Thriller > Psychological" or "Fiction > Thriller > Crime > Serial Killers"

The goal would be to win in a small pond first, build visibility there and then expand to the more competitive categories.

Strategy 4: Create also-bought connections

The "also bought" algorithm is how most readers discover new books. If someone buys Book A (a bestseller in your genre), Amazon shows them Book B (your book) if other people bought both.

But, like everything else on Amazon, you need to create that connection.

How? I would get people who already own popular books in your genre to buy your book, too.

• Target ads to readers of specific bestselling authors
• Post in fan groups of similar books (carefully, not spammy)
• Offer your book free or discounted to readers who already love your genre
• Ask your avid reader friends and family, in your genre, to do you a solid and buy your book

Once enough people buy both, the algorithm connects them.

The Cold Hard Truth About Selling Books On Amazon

Here's what I hate saying, but I think we all know to be true. Amazon doesn't care about your book.

Amazon cares about showing customers products they'll buy. If your book demonstrates it can convert, Amazon will show it to more people. If it doesn't convert, Amazon won't waste its traffic on you.

The algorithm is not designed to help unknown authors. It's designed to maximise Amazon's revenue by promoting products with proven demand. 

This doesn't make Amazon bad, by the way. It's a business after all, and we have to work within their business model if we want to use their selling features. 

The Algorithm Will Help - But Only After You've Helped Yourself

The algorithm is not your enemy, but it's not your friend, either. It's a tool, and like any tool, it only works if you use it correctly.

You can't just upload a book and expect the algorithm to do your marketing for you. You have to create the initial signals - sales, reviews, conversion - that tell the algorithm your book is worth promoting.

Then, and only then, will the algorithm amplify your efforts.

I will never make this mistake of waiting for the algorithm to save me. 

Next time, I'm creating the signals myself, and then I'll let the algorithm do what it does best: amplify what's already working.

---

I write about the emotional and practical reality of being a writer - drafting, doubt, discipline, and publishing while still figuring it out.

Mostly for people who write because they have to, need to, want to | https://linktr.ee/ellenfranceswrites