About That Algorithm…

21 June 2012

From Indie Jane:

I’m going to state up front that I haven’t drawn any solid conclusions by the end of this post, but the information contained below about more changes to Amazon’s algorithm is important for indies to know.

Price matters.

One of the advantages indie authors have over our traditionally published counterparts is that we can choose to sell our ebooks very cheaply. Setting a price of $.99 has proven not only a great deal for readers, but writers still earn 35 percent of the list price, which is much more than what authors of traditionally published ebooks receive.

The $.99 price point has also been a wonderful tool for breaking into a large market and competing successfully against established names. In fact, the $.99 price is part of the strategy I used to propel Absolute Liability to the Top 100 last summer. It is less likely to work now though, and this frustrates me a bit because it removes a tool from my toolbox.

What’s changed?

Amazon’s algorithm. [Insert scary music]

Read the rest here:


- Julia Barrett

Amazon, Ebook/Ereader Growth, Ebooks, Kindle, Pricing, Self-Publishing, The Business of Writing

28 Comments to “About That Algorithm…”

  1. Let’s face it, Amazon is in this for the money. And if along the way they make money from Indie publishers the doors are open. Let’s start with free books. How much is Amazon making on them, zero on the free books, and something on follow on sales, if any. How much do they make on $.99 books, $.69, and on $1.99, $1.39. On $2.99 and higher you have an advantage of choosing 35% or 70% (if you are curious how this election works check this link https://kdp.amazon.com/self-publishing/help?topicId=A30F3VI2TH1FR8#1-1_royalties_calculated) In any case Amazon is trying to find the sweet spot between prices and volume, maximum return. We do the same thing when we adjust our book prices.
    Changing the ranking algorithm based on the cost of the book could be reasoned in several ways.
    First, Amazon agreed with trad-publisher to level the playing field between $.99 and $9.99 eBooks. (Sinister, but not impossible)
    Second, Amazon ranks books by how much money it makes not how many books are sold. A $.99 eBook at 100,000 unit-sales earns Amazon $69,300. A $9.99 eBook at 30,000 unit-sales earns Amazon $104,895. The $9.99 book is more valuable to Amazon.
    Third, Amazon wants to make more money, and it indirectly forces us to raise the prices.
    Fourth, this could be a pre-amble to discourage the free giveaways. This makes financial sense, if you give free books away and then sell very little or nothing after the freebies stop, Amazon makes no money (you don’t either, but that’s not Amazon’s problem.) Maybe the free giveaways have run its course.
    Fifth, …anyone can think of another reason?
    Bottom line, we got to stay on our toes and adapt. I suggest you read Indie Jane’s entire article about the algorithms, very illuminating.

    • DG, I’m pretty sure sales rank is determined my number of sales, not dollars sold, because every time I drop my price to $.99, I sell more books, move up in the ranks, but make sweet nothing.

      • Here’s the thing that isn’t made clear in this post–bestseller rank hasn’t changed. Bestseller rank is still determined purely by number of sales.

        What changed is the popularity list ranks. In some ways, the popularity lists are more important than the bestseller lists. And since the start of May, the price of your book is a factor in how Amazon calculates its popularity lists.

  2. I love how it never occurs to people that the reason gimmicks stop working is not necessarily because some big force intentionally stopped it….

    Most of the time, gimmicks stop working because the customers get tired of them.

    Yes, Amazon adjusts its algorithm regularly — but as Blaine pointed out, as long as the book is not free, the sales ranking is based purely on sales.

    So if anybody sees their 99 cent price point not doing as well as it used to, that’s not the algorithm. That’s the customers getting wise to the tactic.

    FREE, on the other hand, does have an algorithm when a book moves from free to paid lists. Amazon experimented for a while to find the formula which accurately reflects user interest, and discounts the “Free bounce” by that much when the book goes back to paid.

    Always remember, when Amazon adjusts an algorithm, it’s not to advantage or disadvantage anyone, it’s to serve (and accurately reflect) the reader.

    When Amazon doesn’t like your price point, they change the deal they cut with you over royalties. They don’t need to mess with algorithms to do that.

    • My thoughts exactly. The post said the algorithm changed on May 3, but surely 99-cent promotions stopped being as helpful before that?

    • Inaccurate. On the popularity lists, $0.99 titles are no longer weighted the same on a per-sale basis as more expensive books. Therefore they’re given less visibility on Amazon’s storefront. Therefore they sell less. (And therefore their bestseller rank is worse, too, compounding the effect.)

      Maybe Amazon changed that algorithm because readers were growing tired of the $0.99 gimmick, if you want to call it that, but I haven’t seen any proof of that.

      They may have changed things to benefit the reader. They may have changed things to benefit themselves; a high-visibility $12.99 title earns them quite a bit more than a high-visibility $0.99 title. Ultimately, however, there’s no way to know whether a high-selling $12.99 title is “worth more” to the reader than a high-selling $0.99 title. Even if it is, we don’t know if that’s why Amazon made the change.

      All we know is that the algorithm did, in fact, change. Drawing conclusions about their motivations is extremely speculative.

      • Amazon has stated repeatedly that their primary motivation is always the customer. (In this situation, the reader.) The only logical conclusion to draw is that any change they make is motivated by the desire to satisfy the customer. Anything else would be speculation.

        • Amazon is a corporation, therefore it’s primary motivation is making a profit…it sees a large customer base and volume and customer service as the means to earning a profit.

          To believe that any large corporation has a “primary motivation” beyond making as much as possible is really naive.

      • I didn’t say that the other algorithms were affected, I said “sales ranking.”

        I also pointed out that when they change the algorithm, it is intended to more accurately reflect reader feelings.

        Amazon always experiments with new features to get the numbers right. Then they adjust the numbers. They are adjusting the numbers to reflect actual customer popularity, though — not to advantage or disadvantage anyone.

        This is the thing about Amazon, they know their stock-in-trade is accuracy of predicting customer interest. Price to purchase ratios are a factor in this, but Amazon does not game the algorithm to favor any particular price. That is, not out of “we make more profit on this price” — they simply set it so each price reflects the level of real interest shown by readers. So a purchase of a 99 cent book shows less interest than purchase of a $15 book.

        They make money off customer satisfaction, and if they change the algorithm, that’s the reason. If Amazon wants to push particular price points… they can do that directly. They can just set limits on our pricing, or offer incentives.

        When Amazon changes an algorithm, and your sales drop because of it (and frankly, nobody has the real data to judge this) then odds are your sales techniques were “pumping” your sales, and you just dropped to where you would be naturally if all things were equal. (Or as close to it as Amazon can make it.)

        • There is now evidence building that the Digital List Price and not necessarily the actual sales price is what’s being weighted in the popularity lists. For most indies, of course, list and sale prices are one and the same. Amazon imprints seem to be using this tactic to ride higher on the pop lists as the imprints have begun pricing most of their books with a $9.99 Digital List Price while selling at $2.99 – 4.99.

          Speculation on my part is this: I don’t believe the timing of this algorithm change is coincidental nor do I believe it has anything to do at all with a freebie or 99c glut or customer dissatisfaction. As publishers return to a wholesale model, Amazon will be free to discount books again within the DOJ’s guidelines. Thinner margins on heavily discounted books will have to be made up for in quantity. Giving greater visibility to Big 6 books with thin margins will result in more copies bought and better profit. Indies are collateral damage in all this, nothing more.

          I also believe that the algorithms set as they are will indirectly pressure ebooks to gravitate toward the sweet spot Amazon has been pushing all along: $2.99 – 9.99. Amazon has already implemented the direct way to entice authors/publishers to conform to that pricing by offering 70% royalties for the sweet spot and 35% outside of it. When agency pricing came along, there wasn’t an easy way for Amazon to continue to control pricing. Through the wholesale model, they can once again.

          Really, though, IMO it has nothing to do with customer dissatisfaction of seeing lower-priced indie books high in the pop lists (I think we would have heard a groundswell of complaints if that were so) or of how much Amazon does or doesn’t make off of a 99c book. There are concerns greater than where a few indies want to price their books at play here.

  3. The new thing (oh horrors!) making the rounds on some of my loops is that the algorithm is in part affected by the number of “likes” an author has on their author page, the magic number being 40 (don’t ask me. I’m not pubbed except for an obscure short story in an even more obscure anthology). But the “like me and I’ll like you” nightmare is popping up everywhere. Because I’m a perverse little soul, it makes me long for an “I loathe you and the horse you rode in on” button.

    • I probably shouldn’t chime in here, but I so love and admire your perversity!

      • Julia, feel free to chime. Your admiration is only because you don’t know the true depths of my perversity. In realms other than here in PG’s polite salon, the words “foul mouthed female-doglike-person” are flung in my direction with alarming frequency. Then again, other folks DO appreciate what I prefer to call my “frankness”. LOL!

    • There may be a magic number of “likes” — but I think it’s related to people who bought it (not just Liked it), and therefore having it showing up on other people’s Also Boughts. I only have 4 likes on one of mine, but the duology recently started showing up on (more popular, heh) books’ Also Bought lists, and I have noticed a definite uptick in the past few days. It doesn’t seem to be related to anything else I’m doing, or any reviews that are showing up (there aren’t any new ones, woe, woe, woe wallie woe), but it’s definitely present.

      In my case, the Also Boughts kicked in around 60-70 sales, overall. (At least, that’s when I got onto one I cared about, which is the AB for Bujold’s Curse of Chalion. ;) )

      I will say, getting the Dear Author review was probably extremely helpful in the long-run. Many of the books on my Also Bought (and now several vice-versa) are books they’ve reviewed — the review was good enough to indicate that “if you like X and are not thrown off by Y, you might look at this,” and the site is popular enough that there were a small-so-far-but-significant-number-of people who did check it out.

      Another helpful thing may’ve been when Curse of Chalion was on sale for 2.99 and I raved about it to everyone I knew — so there was some crossover of “people who know Beth and bought her book also bought this other book.” :)

      But, in the end, I think I wrote a competent book, and it is finding its market, and while those two factors above probably helped… They’re not gimmicks.

      • Actually I’ve talked to an Amazon person — the main thing “likes” do for a book is look pretty on the sales page.

        “Likes” are used in the user algorithm: if you “like” something, it indicates the kinds of books you’re looking for. It effects the person clicking the button more than it effects the book.

        Any activity on the page can give it a popularity/relevance boost (not just the like button but the visits to the page, and the fact that you clicked on anything while there shows higher “interest” — negative behavior, ironically, will do the same thing, so a “loath” button would help the book the same as a “like” one does). So a massive “like” campaign could possibly give you a little boost, mostly by side effect.

        However, unusual activity like that can be measured and adjusted for.

        • I think what they were talking about was not likes on any book page, but likes on the AUTHOR page. Not a clue what good that would do, but they seemed to think it would have some sort of impact.

  4. When I was younger I spent a long time trying to gimmick Google. The thing is, I found that SEO writing could make you a lot of money, but sooner or later Google got smarter and that money went away.

    If I had spent that time learning to write better articles that humans actually wanted to read I think in the long term I would have done better.

    Gimmick hunting is fun but it is self defeating ultimately.

  5. I knew it! I speculated that price should be taken into consideration a while back. This is the first time I have seen any “confirmation.”

    I know that since I raised my prices, I have sold fewer books but made more money. My ranking fluctuates wildly, however, and right now…well, let’s not talk about it :).

    I have seen ZERO bounce from free promos. When my books go on promo, I hit high, usually #2 in Action/Adventure and in the top 100. But when the promo ends, I go right back to where I was before. I might gain exposure, but I also might be giving away sales.

    I am contemplating taking my books off of the “select” program since I get very few Prime downloads. I understand that…their free download is better spent on a more expensive book.

    “Select” might be a great avenue for a debut (it was for me) but returns seem to diminish rapidly. I think I might be better off reopening other channels for the existing books and only using “select” for new titles.

  6. As in any business venture- and selling your book is exactly that- channels of distribution are the key. Visibility is the key. The carrot of the $0.99 price point or the free days is morphing into another marketing gimmick that “naive” newbies may try, but seasoned authors will abandon. WHAT the next “big thing” will be is yet to be determined. Kudos to the authors who first vetted this info on the Kindle boards and shared with others. Great post and comments.

  7. My humble opinion based on some personal experience. .99 is less affective because it is all too common now. Free is less affective because it is not attracting loyal readers, like the .99 model. People who download free books rarely actually read them, they just downloaded because it was free. There is no value in the. Kim for them because they just “grabbed” it.

    People who only go as high as .99 or 2.00 will not buy anything over that and rarely will continue reading a writer’s work if it ever rose to a reasonable price point, like $5.00 or even $6.00. They are the people you see in the bookstore that only go to the bargain bin for books. Nothing wrong with that, but I believe that is why the free promo books go right back down to where they were before.

    I gave away my first book in celebration of the release of the second. It was an experiment that I learned from. People who don’t and admittedly never would read fantasy, downloaded it. The book received about 100 downloads in two days. The carryover? Negligible. Virtually no sales of that or subsequent books.

    Now that time has passed, and I have more books available, sales are slowly building.

    • One of the issues we face is how to price a novella or a short story. As a reader I would feel ripped off – actually I have felt ripped off – if I spend say… $2.99-$4.99 for what turns out to be a five page book.
      I’m not comfortable charging a reader more than $.99-$1.99 for a very short work
      Now, should I price my 60K works at $4.99? I don’t know yet. I tend to hover around $2.99-$3.99. Maybe I will increase the price to $4.99-$5.99 and see what happens.

    • I’ve done research over the past few months that indicates customers are downloading free books in the same numbers now as they were back in February. It’s not readers “getting wise to freebies” that are slowing post-free sales but a combination of seasonal sales plus two changes in Amazons algorithms that determine how much visibility a book gets on the popularity list and the KOLL.

      Books that can push through and gain a large number of free downloads still enjoy a post-free sales bump. Books that are sale priced to 99c still enjoy a bump up the ranks. You’ll still find 99c books on the pop lists because quantity weights them there despite other factors weighted against them.

      What’s changed, due in large part to new algorithms, is the amount of work the author/publisher must do to ensure a book gains the needed visibility up front. The strategy of merely pricing a book at 99c or offering a book for free and expecting readers to flock to it is no longer supported by the algorithms (although, yes, the model is still supported by a certain customer segment). Understanding WHY the algorithms work the way they do and HOW they work helps us better understand WHAT model Amazon’s internal promotional tools and recommendation engines will favor.

  8. The price points I’m going with are .99 for short stories, 2.99 for novella. (my personal definition of novella is 20-30k words) and $4.99 for my 80k+ books. We’re all experimenting, but in my heart that seems a fair price without devaluing my work.

    • Those are pretty much my thoughts as an unknown author. Pricing higher than 5 bucks would be a sign of hubris, but pricing a full length novel at a buck would be a sign of desperation.

  9. [...] my money the most riveting discussion of the questions raised in “About that Algorithm” is in this post at The Passive [...]

Sorry, the comment form is closed at this time.

Page optimized by WP Minify WordPress Plugin