OLD POST ALERT! This is an older post and although you might find some useful tips, any technical or publishing information is likely to be out of date. Please click on Start Here on the menu bar above to find links to my most useful articles, videos and podcast. Thanks and happy writing! – Joanna Penn
A few weeks ago my friend Joel Friedlander sent me an email (partly shown left) that he had been sent by Amazon which recommended my novel Pentecost next to John Connolly whose supernatural mysteries I really love to read myself.
I was absolutely overjoyed by the news! I put it on Twitter and also received the tweet below showing that Pentecost is being shown in the list of books that other people have bought. Subsequently I have seen it come up myself when I've gone in to look at other fiction books of a similar type.
This is evidence that the Amazon algorithms are kicking in and they are starting to market Pentecost themselves. These algorithms are top secret and no one knows for sure how to ensure this type of thing happens, but here's what we can infer.
- Rankings get you noticed. Pentecost has been in the Religious Fiction rankings since Feb 7th when it launched. It goes up and down but basically has been in the Top 40 or so for that bestseller category and around 2500 overall in the Kindle store. It has also been in the Top 100 of the same ranking in the UK but that site seems to run entirely separately from the US store.
- Ratings and Reviews make a difference. Pentecost has 32 reviews as I am writing this with 4.5 star rating on Amazon.com. This is more than a whole lot of big name authors and clearly Amazon pays attention when readers are enthusiastic about a book. (THANK YOU so much if you have left a review – it makes such a difference! If you love a book, the author will be very grateful for a review!)
Why is this important?
Amazon have an amazing amount of information on readers. They target you with offers that you often find irresistable! They know the types of books you read and offer you similar ones. This is gold for authors as it means that Amazon can reach people with your book who would never have heard of it otherwise.
Why do Amazon do this?
Clearly it's not for your benefit! They are a business and want to make more money so their algorithms look for the books that are selling in specific genres and they try to sell more of them. It's an entirely business point of view but it benefits authors who are selling well – regardless of the publisher! That is a very key point because I can't hope to get into a bookstore near you with my novel, but I can hope that you'll get an email from Amazon recommending my book (if you currently buy books in this genre anyway).
As an author, have you seen evidence of this recommendation algorithm and do you have any ideas on how it works? As a reader, do you pay attention to Amazon's recommendation emails?
Oli Hille says
Actually why don’t you just email me directly: oli.lifestyle@gmail.com
Oli
Jennifer says
Thanks! I have 3 ebooks up on Amazon but am new to internet mkt., lists etc. and I’m learning about the power of reviews from a few posts I’ve read recently. It would be nice to know who’s buying our books but I guess that’s not possible unless they leave a review or contact you? Any tips on how to handle this with Amazon and Barnes & Noble?
Jennifer
Joanna Penn says
Hi Jennifer, all you can do is try to get people to come back to your site and sign up for your next book – then at least you can email them!
So I have a link at the back of my book to http://www.ProphecyNovel.com where they can sign up for the next book. Maybe only 1 in 100 actually take this step but at least it’s something!
David Baird says
I’ve had a few books published and have published several myself. Fascinated by the new possibilities. But baffled.
Would somebody please explain to this ignoramus what the hell is an algorithm and what it has to do with writing/selling books? I thought I had left that stuff behind when I left school!
Greg Scowen says
Hi David,
I’m an author and a web developer/programmer, so have the pleasure of understanding (and now explaining – poorly) an algorithm.
There are few algorithms at work in the Amazon recommending system. These should include the reviews, ranking, other’s bought, after looking customer’s bought and recommendations.
Let’s start with the easy one. Reviews. The algorithm is something like:
sum of review scores /# of reviews
So if a book has 10 reviews mixed from 1 star to 5 star, you take the total stars of all reviews and divide it by the number of reviews. So simply an average, but an algorithm none-the-less.
Ranking presumably works by taking an average of how many books are sold per day. It would be nice to know if this ranking algorithm goes back over all days the book was sold or if it just looks at a defined period, say the last year.
The Other’s Bought feature (people who bought this also bought…) must be stored in a huge database that records all sales and matches books with others that people bought with them. This would also be weighted somehow, so that the most frequently bought books come up first.
The After Looking Customer Bought algorithm would work much the same way as the others bought algorithm. Amazon is recording what people end up buying after looking at a specific page. That means I can hope that people who buy my book first go and look at the same famous book, so one day I might come up on that page.
The tricky one, as discussed here, is the recommendations engine. I assume the algorithm takes into account the ratings and ranks found through the other 4 algorithms above and then creates a new ranking of recommended books for every title in Amazon.
That algorithm might look like this then (a complete guess/example):
For book ‘The Hunger Games’ by Suzanne Collins (B002MQYOFW) when creating a rank with Pentecost:
((Average Rating / 5) * # of reviews) – We know the books is rated well
+
(# of books in Genre – current genre ranking) – So we get a higher figure for better ranked booked
+
# of times Pentecost bought with The Hunger Games
+
# of times Pentecost bought after viewing The Hunger Games
=
The Recommendation Ranking for Pentecost with The Hunger Games.
Something like that is an algorithm. Like I said, this one is just what i came up with during the 5 minutes it took to post this. The algorithm used by Amazon is likely a hundred times more complex and based on months of think-tanks and fine-tuning.
The simple answer to your question is: an algorithm is an equation used to determine the ranking of – in this case – books for use in the recommendation features of Amazon.
David Baird says
Got it!
Thanks for taking the trouble to explain this, Greg.
I appreciate it.
Now all I’ve got to do is produce a best-seller and watch the algorithms demonstrate what a success it is.
Greg Scowen says
Glad to help, David.
In the same boat as you with producing that best-seller.
Joanna Penn says
Wow! Thanks Greg – you are a real star for going into depth on that one. I saw David’s comment and was a little scared because I couldn’t explain it very well. You’ve done a great job. Thanks again!
Greg Scowen says
That’s what I’m here for, Joanna. Glad to finally actually offer something back to your Blog. I’m always jealous of the effort you put in, I am just too lazy to pull it off. It may be because I am working on websites all day in my day-job and find it hard to get inspired in the evenings.
At least, that’s my excuse and I’m sticking to it.
J S says
Here’s the description of how Amazon’s engine works:
http://www.quora.com/How-does-Amazons-collaborative-filtering-recommendation-engine-work
Not any real method to ‘game the system’ or SEO as it’s primarily based on sales of similar books. The best hope is ensuring the category your book is placed in is ‘good’. But there are many opinions on what is the best category for a particular novel.
Karen Inglis says
Thanks to all for info on Amazon insights. I’ve been watching my children’s book, The Secret Lake, rise and fall in Amazon rankings over the last 3 months – nothing like Joanna’s impressive stats (hats off as ever, Joanna!) but it is slowly making is own little waves. In particular, as it is now semi-regularly getting bought alongside other books – including Billionaire Boy and War Horse – it seems to be riding in their slipstream – which is wonderful! I even had a message today that it’s ‘often bought’ with Micael Morpurgo’! I’m hoping that all of these little things will continue to feed off each other to help the sales build over the coming months…
Happy Christmas to you all – and, Joanna, best of luck with Prophesy launch!
Karen
Stu Ayris says
Lovely article and great work. I was just interested in the line when you said that the Amazon recommendation engine was kicking in – how do you know when that process occurs? I’d be really interested! Cheers!
Joanna Penn says
Hi Stu,
As per the screenshot at the top of the post, someone emailed me what Amazon had sent them which recommended my book, and also it appears on the lists that other people bought etc. So you only know by people telling you and evidence of sales e.g. if you start getting really big spikes, it’s not likely to be your efforts, but Amazon’s that make the difference.
I hope that helps, but basically, nobody knows for sure 🙂
Jim Kukral says
The person who can crack it will make millions! 🙂 We’re probably going to see Amazon Ranking Firms soon, like search firms.
Oz says
I am planning to write about a story much like “huck finn”, but do stories like that have a market in amazon?
Michelle Hughes says
Great Stuff. I always say that everyday is a learning experience for self-published authors… running across this article added to my knowledge, for that I wanted to say thank you.
Truly,
Michelle Hughes