Kindle Unlimited

[TheArtfulDodger]

Hello:

During a previous conversation about Kindle Unlimited, it was mentioned that the authors of books acquired in this way did not earn royalties based upon the sale (I was very surprised at this) but rather upon our reading of the books. As I have quite a few acquired in this program, I thought I would inquire as to exactly what the basis of payment actually was.

Several possibilities come to mind:
1. Based on how long the book is actually open in the reader.
2. The last page reached in the book.
3. The book being marked as completed, regardless of how long the book was actually open in the reader. I am hoping this is the actual situation.

None of these may be correct so I would appreciate the actual correct algorithm used by Amazon in determining the royalties. Thanks in advance for your help.

[Quoth]

Only Amazon really knows. To avoid the system being gamed, going to the end of book, or just having it open doesn't count.

This is really nothing to do with Calibre as you don't own KU ebooks.

If you are concerned about the authors/publishers, then buy the ebook.

[DNSB]

Originally Posted by TheArtfulDodger

Hello:

During a previous conversation about Kindle Unlimited, it was mentioned that the authors of books acquired in this way did not earn royalties based upon the sale (I was very surprised at this) but rather upon our reading of the books. As I have quite a few acquired in this program, I thought I would inquire as to exactly what the basis of payment actually was.

Several possibilities come to mind:
1. Based on how long the book is actually open in the reader.
2. The last page reached in the book.
3. The book being marked as completed, regardless of how long the book was actually open in the reader. I am hoping this is the actual situation.

None of these may be correct so I would appreciate the actual correct algorithm used by Amazon in determining the royalties. Thanks in advance for your help.

Kindle Unlimited is a loan not a purchase program so quite a few people will pick up a book, read a few pages and then DNF the book so paying a royalty based on the loan would not be fair to authors whose books are being read.

The closest to reality would be your #2 with quite a few other constraints. As near as I can tell, the algorithm pays based on the read page counts. There is a limit on how short a time the average time to read a page can be to keep robots from flipping pages as fast as possible. From reports from one source, it appears that reading each page in the same time is also considering a sign of a robot being used. The page counts are based on the word count based on ~200 word per page (Kindle Edition Normalize Page Count KENPC). This was done to combat authors who were using very wide margins and large line heights to generate more pages from the same number of words. From reports I've seen by authors, the payment per page read has been dropping for a while now to the point where some authors are leaving KDP and going wide. One item I read indicated that a book that used to earn $1.25 when fully read in KU is now earning about $0.78 so an ~38% drop in income.

#1? Nope. Kobo uses time reading for the Kobo Plus program. Something which hurts authors that I read since I have a fairly high reading speed so don't spend as much time reading a book.

#3? What would keep someone from downloading the book, going to the last page and then closing the book? That was often stated as the reason you saw a lot of books with the Table of Contents at the end so that the author was paid the full amount when the ToC was used. Which might be why Amazon closed that hole.

As for the actual correct algorithm? Good luck with that. Amazon gives out as little information as possible precisely to make gaming the system difficult. Some people have done experiments which seem to indicate it is a fairly complex system with multiple criteria to limit the effects of robot farms.

[nabsltd]

Originally Posted by DNSB

As for the actual correct algorithm? Good luck with that. Amazon gives out as little information as possible precisely to make gaming the system difficult.

I would think that all of the raw data needed would have to be stored on the client...otherwise authors would not get paid if you downloaded a dozen KU books then went offline for two weeks while on vacation. So, it should be possible to find on a jailbroken Kindle.

[DNSB]

Originally Posted by nabsltd

I would think that all of the raw data needed would have to be stored on the client...otherwise authors would not get paid if you downloaded a dozen KU books then went offline for two weeks while on vacation. So, it should be possible to find on a jailbroken Kindle.

Depending on how it is stored, it might be possible to take a look at the data. OTOH, given Amazon's (justified) paranoia, I would not expect decoding the data to be an easy task. After all, we are talking millions of dollars here since the KU monthly pool for most of 2023 has hung in at ~45-49 million US dollars.

Hmmm... A public key on the device for encrypting the data generated by Amazon with the private key available on Amazon servers. Possibly using a hash based on the serial number, owner name, etc. to link the two.

[Frogm4n]

The regular reading stats on a device are fairly detailed and can be see when doing a personal data request from Amazon. I don't use/have KU so I don't know if there are other additional stats collected for those specific ebooks, but the regular data is plenty detailed already. We just don't know the exact specifics of how the data is used to calculate KU reading by Amazon because that happens on their servers.