On June 15th, Amazon announced that they will change the reward system of their KU borrows from 1st July onward. In short, in the past , 1 borrow of a book would earn the author a payout of the ‘KU pot’. So if the pot was $5.000.000 and there were 3.333.333 borrows, each borrow would earn 5.000.000/3.333.333=$1.5
With the new system, Amazon will count the ‘pages read per month’ and calculate the monthly Payment Per Page Read (which I will call PPP). Let’s assume as an example that the PPP is $0,02. If a customer borrows a 100 page book, the author will get paid 100x$0.02=$2 . This is, IF the customer reads all 100 pages that month. If he only reads 10 pages, the author will get only $0.20. Notice that if a customer reads the complete book but spread out over several months, the total $2 will be spread out over several months.
The complete announcement can be found here.
As we all know, ebooks don’t have pages. So Amazon had to come up with some kind of formula that determines how many pages a book has, taking into account all the different devices, font types, font sizes, line spacing etc. To make this as fair as possible, Amazon has defined a standard Kindle Edition Normalized Page Count (KENPC). The only thing that is known about it at this moment of writing is , that KENPC has been defined. The full text is here. Unfortunately, they haven’t specified in detail what the font size, font type, spacing etc is used for the KENPC. So we will have to guess that for the moment.
Another thing mentioned in the full text from Amazon, is that they will take images, graphs and charts into consideration in their KENPC formula. (without any further detail). I guess that this means : if an page has a chart on it , it will count as X lines of text in the calculation of the KENPC.
The whole purpose of the exercise seems to be, to reward the actual work of an author. As a quick example: With the “old” KU payment system, an author that gets 10 borrows a month for his 500 page novel, would earn the same as the scam artist who also gets 10 borrows of his 20 page scamphlet. With the new system, and if we assume a payout of $0,02 per page read, the successful author of the novel would now get paid 10x500x$0,02=$100 (assuming that the customers read his whole book in the same month), while the scam artist would only get 10x20x$0,02=$4 . And if the customer doesn’t get further than 50% in this scamphlet, it would even be reduced to $2 . Seems fair enough.
This indeed is a first step into the direction of ‘pay for performance’. But there remain a number of questions unanswered.
Before I get to these questions, let’s analyze a bit what may happen, using the data of the past that we have.
First , let’s see what the KU pot, KU payout and total number of borrows was from August 2014 onward.
(The numbers in yellow are future estimates. The nr of future borrows is calculated with the TREND function)
In graph form we get:
We see that the borrows are still growing every month. And to keep the payment to a ‘reasonable’ level, the KU pot is also growing every month. But the KU payout seems to stabilize somewhere around the $1.40 level. Notice that there is no reason why the nr of borrows in the future should change, because there is nothing that changes for the readers. So we may fairly assume, that IF the old KU payment system would have continued, the average KU payout would have been around $1.40 .
Now what will be the average payout per book in the new system? The first thing that we would have to know is what the number of pages are in the ‘average’ book. Is it 50 pages? 100? 200? 400? Of course Amazon knows this number, but I don’t. So how to get a feeling for that number? The best I’ve come up with, was to select totally random 50 books and look at the ‘page count’ that is shown for that book. I realize that there might be some flaws in this. First, the number of pages on the current detail page of a book is not necessarily similar to the new KENPC. Second, I assumed that sales will be completely random over all books on Amazon. Which is probably not the case. Novels, which are typically a couple of hundred pages, sell more than non-fiction ‘how-to’ books which in general have much fewer pages. But since I don’t have these numbers, I decided to go for a complete random selection.
For these 50 books I calculated the average number of pages over all books, the average number of pages of all ‘short’ books, and the average number of pages of all ‘long’ books. Short being, less than the overall total average and ‘long’ being above the overall average. The results were:
Overall average length : 225 pages
Average length of short books : 70 pages
Average length of long books : 400 pages
Now using an estimated KU payout we can make a 2 dimensional table with horizontally the number of pages per book and vertically the % of the book that will be read. This gives something like
How to interpret this data? For example, if in the old KU payment system the KU payout would have been $1.40 per borrow, the new PPP (Payout Per Page) for a book with 200 pages that is read for 80% would be $0,009 .
If we take the same table for a more realistic KU payout of $1.40, we get
And if we do the same for a more pessimistic payout of $1.35, we would get
The green areas are the areas of which I think that they are the most plausible. If we study the above a bit more in detail, we can conclude that the expected PPP will probably be somewhere between $0,005 and $0,05. Which is still a pretty large range.
There is another way to look at it. If we assume a certain number of pages per book, we can make a 2-dimensional table with horizontally the KU payout and vertically the % read. This gives for example for a 70 page book.
And for a 100 page book
And for a 200 page book
Now if we look at these 3 tables, we see that the average PPP is somewhere between $0,008 and $0,045. And even with an upper limit of $0,023 if we look at the 200 page book. So if I had to guess ONE number for the future PPP, I would set it at around $0,025.
Now how will this impact the average author? Well, first problem : what is the definition of an ‘average author’. I don’t know. But instead, we could define a couple of typical authors. For this exercise, I have imagined 5 typical profiles:
1. The short story writer. Writes books with less than average pages (40-80 pages)
2. The long story writer. Writes books with more than average pages (200-300 pages)
3. The scam artist. Writes useless books. As short as possible. (20-30 pages)
4. The reference book writer. These authors write dictionaries, quote/joke collections, reference articles etc. These books are typically used only to lookup something from time to time.
5. The children books writer. These books have lots of images. This one is a bit tricky because we have no idea how the images will be counted for in the KENPC
For all 5 profiles, I have defined 3 levels: The successful one, the average one and the poor one. For example, the successful long writer, gets 1000 borrows a month and are read for 100%. The average long writer gets 100 borrows which are read 100% and the poor long writer gets 10 borrows which are read for 60%. This gives 15 different profiles. Now we can plug in some figures for these profiles and compare their payment with the old KU payment system and the new one. For an estimated KU payout equivalent of 1.40 and an average PPP of $0,01 we would get the following table:
In this case, only the writers of long books and the writer of good reference books would benefit from the new system. Their revenues could almost double. The clear losers are the scam artists. Which is of course a good thing. But the short story writers would also get less.
If we assume a slightly higher PPP of $0,02 we get the following table :
In this scenario , almost everyone wins except for the scammers who lose up to 90% of their revenue. For the good long writers, their revenues may get multiplied by 4 !
We could make two more simulations , similar to the above, but with a more conservative page count per book.
The result is similar to the table above, but the gain for the long writers is more modest . But still 43%!
For the PPP of $0,02 we get
Again, similar as above , but now the increase for the good long writers is ‘only’ 180% . Or almost tripled by 3.
If we know the PPP and the corresponding KU payout in the ‘old’ system, we can calculate what the ‘number of pages read’ should be to break even. In other words, if your book has fewer pages than the break-even point, you will earn less than what you would have earned in the old KU system. And vice versa. This gives us the following table :
In other words, if we assume that the payout will be equivalent to the old system ($1,4 per borrow) then depending on the PPP, we would get
which means that at a PPP of around $0,02 books with less than 70 pages would earn less in the new system as compared with the old one. Given the data above we could guestimate the break even point at around 70-100 pages.
To summarize , I would guess that the expected PPP would be around $0,02. The writers of good long books will be the big winners. Their revenue may increase anywhere from 50% to triple. The good short story writers will also benefit but their gain will not be as spectacular as the long story writers. Whatever the scenario, the scam artists will be the big losers.
As stated in the beginning of this article, there are still a number of questions that remain unanswered.
-How long will Amazon keep paying the author?
As we have seen, a customer may buy a 1000 jokes collection of 300 pages and read every month only a couple of pages. Will Amazon keep paying pages that are read 1 year, 2 years, 5 years from now?
-What is ‘a read page’?
That’s a pretty tricky question. It takes maybe 5 minutes to read a full page of a novel, but only 5 seconds to ‘read’ a page with an image. Now I cannot imagine that Amazon will go as far as measuring the time spend on a page. So my guess is : a page opened is a page read. Of course, because pages read will only be taken into account when that data is recorded on their servers, it means that pages read offline will only be accounted for whenever (and if), the device is connected again to the internet and gets synchronized. And of course, read pages will only be recorded once. Even if the customer rereads your book 10 times 
-How will charts/images/graphs me counted?
I already addressed this question above. For the moment there is no information whatsoever how these images will be counted for in the KENPC.
Let’s see mid August, when we get the first KU/KOLL payment calculations ,how far I was off.
Did you like this article? Then please like my author page.
Want to get my articles mailed to you? Subscribe to my list on the top right of this page.
