In our digital world, data is becoming increasingly important, both for creating content and finding content. Major companies use algorithms for these purposes. However, this strategy is not new. 
For books, data has always played a role in marketing. Best seller lists are compilations of data, created based on algorithms, and some people have been taking advantage of this for years. Jeffrey Sussman, who works in public relations, said that JFK’s father paid people to buy copies of his son’s Profiles in Courage from bookstores that he knew reported to the New York Times best seller list. This propelled the book onto the list and made it a success.
Since then, the algorithms have changed, but the strategy remains the same: get on the list and become a best seller. Below are some findings of a few site’s algorithms and what helps to sell books on them.
KDP Select
For a while it seemed that the fastest way to earn money by self-publishing was signing up for Amazon’s KDP Select program, which gives Amazon exclusive rights to the book for 90 days at a time. In exchange, the author gets 5 days to promote the book for free, and the book is included in the Kindle Owner Lending Library. The payout is around $2 each time the book is borrowed. In the beginning, authors who took advantage of the freebie days could see their books shoot up the Freebie list on the website. Then, after the promotion, the buzz would spill over and turn into actual sales, which would help propel the book up the Paid best seller list on Amazon. Being on the paid list attracted more attention, readers, and sales, which kept the book on the list and helped sell more copies. Success fueled success, and a few authors became best sellers.
But a while back, Amazon changed its algorithms (see Elle Lothlorien’s explanation in “Hang Up Your Pimp Costume, Kids: Why Free Book Promos on Amazon Don’t Work Like They Used To (No Matter How You Spin the Numbers)“) and the freebies stopped being an easy way to sell books. It still works, but it’s a lot harder now to find the same level of success. James Moushon on Self Publishing Review and Martin Crosbie on Indies Unlimited give a list of tips and sites to use to run a promotion on Amazon. Additionally, it may also help to run an ad on the site Kindle Books and Tips.
(As a side note, Booktrope was able to turn a free book into a best seller using similar techniques, but the book was available on both Kindle and Nook.)
Interestingly, Elle Lothlorien later wrote another article, “THING 3. Prostitute Your Book: The Art and Science of a Becoming a Successful Free Book Pimp on Amazon,” which details a breakdown of exactly what it takes to succeed. Basically, there’s a magic number of 20,000. Read the post to find out the formula for attaining that magic number. However, there is speculation that computer algorithms could eventually replace book reviews. Already computer software exists that can categorize and analyze art, and eventually there could be a software that uses machine learning to give trusted reviews of books–which would probably change Lothlorien’s formula.
Metadata and Amazon Algorithms
Joanna Penn wrote a helpful book, How to Market a Book. An excerpt is published on Jane Friedman’s site, and it goes over how to sell more books by optimizing metadata. Metadata is data about the book, such as description, author name, title, etc., and you can read more in depth posts about metadata here.
Penn offers advice for how to pick categories, find keywords, and use them in a book’s metadata to increase sales. She also recommends reading David Gaughran’s Let’s Get Visible to understand how Amazon’s algorithms work and make the best use of them.
An in-depth article that explains how to use keywords and pick the right categories is Penny Sansevieri’s “Keys to Understanding Amazon’s Algorithms.” Another is Jin Lee’s “The Secret to Become an Amazon Top Seller.” She also advocates researching keywords and having strong metadata. But she emphasizes that it’s more important to sell hundreds of copies in one day, or more specifically, aim for 300 sales.
Many other experts have written about how authors with more books tend to earn more money. In most cases, these authors have series and build a fan base that’s ready to purchase each new book in the series. But it’s interesting to think about from a metadata/algorithm perspective as well.
David Hill posted on Singularity Hub “Patented Book Writing System Creates, Sells Hundreds of Thousands of Books on Amazon.” A professor from INSEAD, Phillip Parker, “created a computer system that can write books about specific subjects in about 20 minutes. The patented algorithm has so far generated hundreds of thousands of books. In fact, Amazon lists over 100,000 books attributed to Parker, and over 700,000 works listed for his company, ICON Group International, Inc.”
The books are sourced from a database of information, and content is broken down into a formula. It costs about 20 to 50 cents to produce a book, but some of the books sell for $795. Even though not all the books sell well, there are enough sales that Parker makes a nice profit. I think in this case it helps to have so many books under one name.
Other Metadata and Algorithm Uses
Algorithms are working and more prevalent than ever when it comes to content creation. Netflix, which has been releasing its own series/shows, decided to renew Hemlock Grove for a second season based on its data from algorithms. Janko Roettgers on paidContent pointed out that Hemlock was meant for a niche audience so it was not heavily marketed. But the fact that the show has been renewed proves there are viewers.
Google’s search algorithms also heavily rely on meta tags. According to Matt Cutts from Google, meta keyword tags don’t do anything, but meta description tags and other meta tags “are worth paying attention to.”
“The meta description is really handy, because if we don’t know what would make a good snippet, and you have something in the meta description tag that would basically give a pretty good answer–maybe it matches what the user typed in or something along those lines, then we do reserve the right to show that meta description tag as the snippet. So we can either show the snippet that might be the keyword in context on the page or the meta description.”
“Now, if the meta description is really well written and really compelling, then a person who sees it might click through more often,” he says. “So if you’re a good SEO, someone who is paying attention to conversion and not just rankings on trophy phrases, then you might want to pay some attention to testing different meta descriptions that might result in more clickthrough and possibly more conversions. So don’t do anything deceptive, like you say you’re about apples when you’re really about red widgets that are completely unrelated to apples. But if you have a good and a compelling meta description, that can be handy.”
In addition to understanding meta tags, Karen Lotter on Self Publishing Advice Blog wrote that there are four things indie authors need to understand about Google search: web crawling and indexing, page rankings, search, and Google policies. See the chart above for other ideas on how to rank high with SEO.
What do you think? Let me know in the comments if you have any other tips for improving metadata and SEO!
[…] Ricci presents Marketing with Metadata and Algorithms posted at Digital […]