Jeff Duntemann's Contrapositive Diary Rotating Header Image

Paying by the Page Turn

Amazon’s Kindle Unlimited (KU) book subscription system has been a laboratory of unintended consequences since it launched in July 2014. If you don’t subscribe or don’t know how authors are paid, my 4-part series on it may be useful. I’ll summarize very briefly: Each time a work available on KU is borrowed and at least 10% of it is read, the author is paid from a payment fund shared by all such borrows in a given month. The amount of money in the pot changes from month to month, as do the number of borrows. So the payment per qualified borrow changes from month to month. It’s been converging on $1.30 for some time. The length of the work doesn’t matter: Read 10% of a 150,000 word novel, and the author gets $1.30. Read 10% of a 1000-word short story…and the author gets the same $1.30. (For another another few days, at least. Stay with me.)

Care to guess the unintended consequences? Authors of novels pulled their works from KU or never opted in to begin with. Authors of short stories suddenly started making significant money. Authors of flash-length erotica (basically, isolated sex scenes) began making a great deal of money. And scammers began posting the same (very short) story on multiple author accounts, and Wikipedia articles as original works.

I could have guessed all of that except maybe the erotica, since I don’t read erotica. I had actually begun turning my individual short stories and novelettes into separate ebooks, figuring that $1.30 was way better than the 35c that 99c ebook shorts earn.

Alluva sudden, wham! Everything changes.

On July 1, a whole new KU payment system comes into force. The new system essentially pays authors by the amount of the book read. Read the whole book, author gets X. Read half the book, and author gets X/2. Read 10% of the book (perhaps because it was so bad you wanted to throw your Paperwhite at the wall) and author gets X/10. In general terms, when you read some arbitrary number of pages, author gets a pro-rata per-page payment. This is true (and evidently the payment will be the same) whether the book in question is a kids’ bedtime story, a romance novel, or a calculus textbook.

As in the current system, the per-page payment changes every month, depending on the size of the money pot and the number of pages read during that month. The two big variables are the per-page payout and the number of pages in the book.

Wait a sec…pages? In an ebook?

Yup. And this is something completely new. Amazon has addressed the fact that ebooks are not divided into pages by creating the Kindle Edition Normalized Page Count (KENPC) algorithm. As best I understand it (details are sparser than we’d like) the KU servers will examine each book posted by an author, and impose a standard page layout on the book’s text in a buffer. (It will not actually change the layout in the published book.) It will then count how many “pages” exist in the book when expressed by the KENPC algorithm. I have seen no reliable description of what will go into this standard layout. It’s obvious that they’re trying to keep people from padding out margins or tweaking fonts to turn less text into more pages. They’re also trying to equalize the differences between devices with vastly different screen sizes. KENPC takes into account photos, tables, and technical art somehow. Again, details are sparse. However, I’m happy just knowing that they’re going to some effort to make a page on one device more or less equivalent in terms of content to a page on another device. I’ve seen some grumbling about page metrics for children’s books, but since that’s a genre I have no experience in whatsoever, I can’t say much. It does seem a little unfair that a 30-page kid book will only earn what 30 pages in a 500-page novel earns.

Pages will only pay off the first time they are read. Reading a book a second time on the same borrow will not generate any additional revenue. Nor will going back to reread a chapter generate additional revenue. Swiping/tapping rapidly through a book will not pay. Some sort of timer runs while a page is displayed, and if the page isn’t displayed long enough, the page will not be considered read. Countable pages begin with the book’s starting point, so dedication pages, review excerpts, and indicia will not be paid.

Now, what can authors expect as a per-page payment? Nobody knows yet. People are guessing somewhere between .8c and 1c per page read. We’ll find out soon.

Any system like this is a basket of unintended consequences. These are the ones that immediately occur to me:

  • Authors of art-heavy children’s books will bail.
  • A lot of that flash-erotica will vanish. (This may be an intended consequence.) Or maybe not. A nickel is a nickel.
  • More previews of other books will appear at the end of a book.
  • Reference books will bail. This may include computer books, which are rarely read from cover to cover.
  • Page-turners will dominate. Difficult books (fiction or nonfiction) will bail.

This last point bears discussing. Some books are bought to be seen in buyers’ hands or (more often) on their coffee tables. As Megan McArdle points out, Thomas Piketty’s book Capital in the 21st Century is purchased a great deal more than it is read. I think this is true of a lot of literary fiction as well. Authors will have to understand that they’re no longer selling books. They’re not really selling pages, either. They’re selling page turns. To make money on KU going forward, each page will have to compel the reader to move on to the next page, and repeat until EOB.

This is bad news for James Joyce. This is good news for George R. R. Martin. And, I suspect, me.

It may also be bad news for writers who just don’t know what they’re doing. To pay by the page-turn, Amazon will have to report how many pages were turned. How much detail those reports will provide is still unknown. It would be terrific to know how many pages were read per title rather than in aggregate across all of an author’s titles, but I don’t think Amazon will be doing that, at least not right away. However, if you have ten 300-page books on KU and get paid for 67 pages, the reader base is telling you something.

I suspect that this is a fully intended (if unstated) consequence: to improve the readability of the material on KU. Fistfights break out frequently over whether readability and quality are strongly correlated. This is the dotted line where literature is separated from fiction, especially genre fiction. But consider what KU is: a mechanism allowing maniacal readers to get all the books they can read for ten bucks a month. If you’re a normal human being, Finnegan’s Wake will take you most of a month to bull through, and you can get ratty copies for a penny plus shipping online.

No, it’s going to work like this: If you can keep a reader up all night with your hard SF action-adventures, you can make money on the new KU. Write page-turners, and put previews of all of your page-turners in the backs of all of your page-turners.

That’s certainly what I intend to do. I will make money. Watch me.


  1. great unknown says:

    Wonderful idea – putting a preview of the book in the back. It might modify my habit of reading the first few pages of a book, then the last few pages, and then – over a longer period – filling in the middle in random jumps.

  2. TRX says:

    That’s how my wife reads a book. ebooks drive her nuts because it’s harder for her to skip back and forth.

    Our readers are ancient and primitive, and require extensive scrolling to move far; they don’t page.

    1. Most of the problems I’ve had with ebooks in the last ten years has had nothing to do with the books themselves. The readers have been eye-strainers, the formats unsupported by true WYSIWYG editors until very recently, the rendering engines pick and choose what elements of CSS they will render, the reader UIs are primitive in the extreme, etc. etc.

      Some of these problems are beginning to be solved. I had expected them all to be solved a long time ago. The miracle is that ebooks have become as much of a force in publishing and reading as they have.

    2. jic says:

      I also tend to skip back and forth, and since I consider it a bad reading habit that makes it harder to finish books, the fact that it’s difficult to do when reading ebooks is actually an advantage to me.

  3. Tom Roderick says:

    A very long time ago I was given a speed reading class as a high school graduation present. It was a nice present and I suspect, at that time, a rather expensive one. This was an actual class room course that met twice a week for about six weeks as I remember it with “homework”. Although I did well in the course and was told I was able to read 10,000 words per minute or more I really didn’t believe it.

    However, when fall came and I was an overloaded freshman engineering student I did use the techniques to cram for some exams in the non-engineering electives. Courses in Political Science and Sociology are examples. It amazed me that I was able to take the exams and regurgitate the “right” answers and did well in those electives. I am not saying that I really learned anything, however.

    Latter it was very useful in learning a new technology that I had to get up to speed on in a hurry. In that case the speed reading just provided the outline onto which a deeper study could be quickly hung.

    I have NEVER used the technique to read any kind of fiction. Fiction is read for pleasure and there is little pleasure in speed reading. I find it very taxing and draining. I don’t know if others have had this same experience but it would be interesting to see what a 10,000 word per minute reader would do to the Amazon algorithm.

    1. jic says:

      I never studied speed reading, but I did become very good at skimming during college. It was great for getting through textbooks quickly, but was terrible for reading fiction. I had to deliberately try to slow down, and I was so ‘successful’ that I started to have problems finishing books. It took me years to get any sort of balance back, and I’m probably still a little slower than I should be.

      1. David Lang says:

        There is a huge difference between skimming and speed reading. Speed reading training done properly includes tests on the content that you have read, so it’s not speed at the expense of comprehension. It also includes vocabulary training so you know the meaning of more words without having to look them up.

        When you start reading, you sound out each word, when you get to the point that you no longer need to do that, not only does your speed drastically increase, but so does your comprehension and retention.

        As I describe it, you spend less of your brainpower on the mechanics of reading, so you have more available to spend on the contents. We also have a finite attention span, and the more reading that you can do within that span, the more you are able to combine everything that you read to get the meaning out of it.

        This is very much like typing. If you don’t know how to type, you are so distracted by the keyboard that it’s hard to get thoughts down. As you get faster, you pass a point where they mechanics of typing disappear for you and the words flow from your brain to the screen. You don’t expect that a professional typist who is banking away at 70 wpm on the keyboard is going to be making a horrific level of mistakes compared to the person doing hunt-and-peck with two fingers, so why do you assume that the person who knows how to read fast is going to miss the meaning and subtleties of the work they are reading?

        Speed reading is the same thing. Depending on how you read (a letter at a type, a phonetic unit at a time, a word at a time, or a sentence at a time) you will have a fairly predictable cap on your reading speed. Somewhere around 200 wpm reading there is such a cap, and when you get beyond that to the 300+ speeds, you will find that you not only finish faster, but understand it better and remember it better.

        I’ve taken a couple different speed reading classes, the first in 4th grade, the second in college. I tested at ~1200wpm at the end of the college class (slightly beyond the 1000 wpm the test was calibrated to max out at), and I have friends who read about twice as fast as I do. I also have other friends who struggle to read at ~50 wpm. It’s very clear to me who gets the most meaning out of material they read, even just giving each just one pass through the material rather than giving each equal time.

        1. Tom Roderick says:

          The speed reading class I had did cover exactly the things you mentioned. One of the most important was in learning that you did not have to “hear” the sound of the word to know the meaning. This is the limiting factor that caps most people’s reading speed. Sound occurs in the time domain and you can only compress it so far.

          There were two other “tricks” which, with LOTS of practice, do seem to work. The eye can focus on an area much larger than one word. It tends to be an elliptical area several centimeters for both the major and minor axis of the ellipse. In some ways it is like being able to read several lines in parallel and have them sorted out in your mind. Also, you can read left to right and then right to left so you don’t have to scan the page as the old CRT tv’s did but can use the retrace to read the next few lines going the other way.

          Having said all this probably explains why I never liked to use any of these techniques for pleasure reading. The level of concentration needed is intense and it just wasn’t at all fun. Also, in fiction I think I ALWAYS lost some of the subtle nuances that the writer worked so hard to put in the work. Speed reading fiction, at least as I was taught, is really not fair to the author! However, it did seem to work pretty well for assimilating a large amount of technical information very quickly and the retention was not too bad, but could be greatly improved by making just a few short notes after reading a section or topic.

          I do think, however, that having had this course my “normal” reading speed did increase considerably since I was able to still use the concept of sight to meaning without having to hear the sound in most of the things I was reading. The one huge exception to this is, of course, poetry, which without the sound is nothing.

          1. TRX says:

            > The eye can focus on an area much larger
            > than one word. It tends to be an
            > elliptical area several centimeters

            I’m pretty sure that’s linked to age.

            Now that I’m in my mid-50s I’ve run into several situations where my focal area / attention spot is noticeably smaller than it used to be. I notice it mostly at the computer, where I only use a small part of the triple-monitor desktop, and when driving, where I’ve noticed both a tendency to target fixation and a loss of peripheral attention. I had my eyes checked a few months ago; there are no problems with my eyes, just what my brain is doing with the information.

        2. jic says:

          “There is a huge difference between skimming and speed reading.”

          Which is why I specifically said that I never studied speed reading. I was talking about something that was related to the issue at hand, not exactly the same.

  4. TRX says:

    When I was young I was tested at just under a thousand words a minute. The test involved reading various blocks of text and answering questions about what I’d read.

    Nowadays I’d be surprised if I go much over 200 for fiction, less depending on what kind of nonfiction.

    Back then I was learning, so I vacuumed everything down uncritically. Now I have a fairly detailed model of the world; everything new gets inspected, compared to the model, considered, given an arbitrary weighting of accuracy, and carefully slotted into storage. Nowadays I’m prone to long pauses while I think.

    1. That’s a beautifully phrased description of how I read as well, though I’m sure I do over 200 WPM for fiction. Nonfiction depends entirely on the topic, the quality of the writing, and how much context I already have. Tough topic, bad writing, no context means a slow read.

    2. TRX, my experience models yours in many ways. When young I was also a sweeper, gathering almost everything possible and filing it away. During engineering school, I learned quite a lot about critical thinking/reading and became more discerning about what I read, even in the professional literature.

      Now, with the internet-of-all-things, I read widely but am reluctant to believe much of that content (casting it as opinion until validated). I still prefer books, but am learning to use ebooks as a lighter alternative (although there are problems with ebooks as references).

      The model analogy is excellent. I also tend to pause for lengthy periods of time to reflect. I also tend to write more (in my journals, mostly).


  5. jimf says:

    It is interesting to see Amazon experiment with different methods…it seems as if they have a good shot at making the latest version work. I personally would not subscribe to KU as I am pretty selective on what I buy, vs which ebooks I check out from the library…I hope you get paid on KU library books too.

  6. […] algorithm seemed kind of gonzo: The same amount for 1,000 words as 100,000 words? As of July 1, it’s now about pages read. We can quibble about the per-page payment, but my spreadsheets tell me that at current rates, an […]

Leave a Reply

Your email address will not be published. Required fields are marked *