Last Updated on August 23, 2022 by Ben Oakley
Think of Amazon as a machine, artificial intelligence if you will, propped up by human drones employed to maintain a corporate image. The machine uses algorithms alien to the human mind, disconnected but at the same time vital to the lifeblood the machine needs to power it. Just like KENP page reads!
With millions of books published through KDP each year, there is no possible way for a team of humans, no matter how large to maintain the system. Thus, algorithms are used within the Amazon machine, an AI constantly working, learning, and shifting the goal posts.
One such algorithm is the ranking system on Amazon, its bestseller rank, subgenre ranks, whatever you need a ranking on its there, teasing you, tempting you with the best of the best, and laying down the gauntlet for those whose ranking sits comfortably over a million.
This algorithm is like a bitcoin address; unhackable, its transactions known only to the user, yet broadcast live to the world on a blockchain that thousands try and work out each day.
It is controlled only by the AI within the Amazon machine. The way it works is a mystery known only to the handful of humans who helped develop the algorithm.
Since KDP went live, writers and publishers have sought to crack the code, unveil the algorithms, and uncover the inner workings of the machine for one purpose – to use it to their advantage.
Amazon KDP Insights.
Sure, Amazon KDP has human gatekeepers but they are immaterial when it comes to the machine. But there is a misconception within the ranking algorithms that leads some self-publishers to believe that KENP page reads contribute towards the ranking figures on Amazon – they don’t.
Let me explain in detail because this seems to divide writers and publishers right down the middle, which is understandable as the truth of the situation is controlled by Amazon’s algorithms – it’s AI, the uncontrolled ghost in the machine.
Amazon KDP says, ‘Best Seller and Category Ranks are based on customer activity of your book relative to the activity of other books.’
It makes sense, right? If customers buy books from Author No.1 and not Author No.2, then the rankings of Author No.1 will go up, whereas Author No.2 will go down in a relative fashion.
‘Monitoring your book’s Amazon sales rank may be helpful in gaining general insight into the effectiveness of your marketing campaigns and other initiatives to drive book activity, but it is not an accurate way to track your book’s activity or compare its activity in relation to books in other categories.’
What that means is that Amazon sales ranks are constantly changing and if you’re using the ranking system as a marketing insight then it’s not always going to work how you think it would.
It appears from the Amazon KDP help page on sales ranking that various activity contributes towards ranking. Except, when you read that page, there is no information on exactly what contributes to ranking – and this is done on purpose.
Too many variables in rankings.
Despite Amazon being a machine, an inorganic system working overtime to monitor submissions through KDP, it is at the end of the day still a system. And where there is a system, there are secondary systems built around them to try and manipulate them.
Amazon will never tell you exactly what affects rankings because there are too many variables. An author could sell 10 books in one day, their best ever, but rankings won’t move because every other author in the world has sold 11 books on the particular day or hour.
The algorithm updates the rankings every hour – mostly, though if you’re refreshing the KDP royalties estimator every hour then you could experience delays in reporting. No system is perfect.
In terms of eBooks, which we’re mostly discussing here, only when a book appears in the library of a Kindle or a Kindle app, does the ranking reflect that sale OR borrow. KENP read does NOT affect the ranking of a book.
At the moment, I average around 300,000 KENP read a month in addition to the sales of my books. When someone makes the decision to BUY a book, that sale is reflected in the overall Amazon eBook rankings, and any category the book may be under.
Some categories get you a #1 bestseller tag in the 50,000 ranking range, others sit firmly below the 5,000 range. Your ranking in each category depends on the category you have chosen.
Playing the category game doesn’t get you anywhere, as the rankings to be in the top 10 in a category with just a couple of sales means nothing. Hitting the top 10 with one sale does not equate to hitting the top 10 in another category with 100 sales. Relativity is important in this game.
KENP affects royalties only.
But wait, why doesn’t KENP read affect rank? Simple – KENP read affects the Royalties Estimator, and more precisely the KENP Read page on the KDP Reports. If someone borrows a book using their Kindle Unlimited subscription, it counts as a sale.
Here’s another way of looking at it:
- KENP = $$$
- Book borrowed on KU = ranking movement.
Why does a borrow count as a sale? With Kindle Unlimited, you can borrow as many books as you want but only keep a certain amount on your bookshelf. If your bookshelf’s full then you must return a book to borrow another one.
A KU borrow counts as a sale within the algorithm because it puts that book on the reader’s Kindle or Kindle app. The same process is used when books are on a free promotion. Every book borrowed or added to the Kindle bookshelf counts as a sale, albeit affecting the free book rankings and not the paid ones.
Rankings don’t move on KENP reads, they move through a sale or a borrow which counts as a sale. You don’t get paid for a KU borrow even though that borrow affects the sales rankings on the Amazon store. You get paid when the reader begins reading your book.
Amazon are always one step ahead.
The reason that many authors and publishers tend to believe KENP affects ranking is because from the time a book is borrowed to the time it is read is generally a small time frame. Some people might read the first chapter straight away or read it in one go.
Going back to the machine – Amazon’s algorithms are world class but to build a ranking system that’s reflected through KENP is one step too far and makes no sense due to varying page lengths across various genres.
A KENP-based system would result in a flawed ranking system, as one person reading a 1,000-page book could equate to 100 people reading a 10-page book. Thus, one person would only have to read one book to rank at the same position where 100 people have read another book.
Oh, and if you’re thinking that means the KU or Prime borrows can be played for your benefit – forget it. Amazon says, ‘if we determine that your book’s purchasing or borrowing activity originates from accounts that are attempting to manipulate our services, it may result in termination of your account and loss of royalties.’
Remember, Amazon is a machine, its algorithms dictated by humans but run autonomously using artificial intelligence and programming that many of us will never understand.
And as much as we attempt to break down the code, Amazon are and always will be one step ahead. You’re either with them or you’re not.