Expert publishing blog opinions are solely those of the blogger and not necessarily endorsed by DBW.
How Amazon pays authors for work included in Kindle Unlimited (KU) made headlines across the inter-webs recently. Ann Christy’s post “KU Scammers on KU – What’s Going On” even made it on to the homepage of Hacker News. The discussion raises many interesting questions about what reading data Amazon collects and how Amazon uses reader analytics.
First, a little background: Amazon introduced KU, its all-you-can-eat ebook offering, almost two years ago, not long after Oyster launched its much lauded, but now defunct, ebook subscription service. Authors were initially compensated by Amazon based on the number of ebooks downloaded, but that system was being abused by some clever folks who realized that short books, such as novellas, would earn the same amount of money as full-length novels, and that splitting a full-length book into multiple books would optimized payouts.
Readers did not like this practice, so Amazon changed its policy and introduced “pay by page” in June of last year.
However, enterprising souls once again quickly discovered another loophole, as the way in which Amazon measures pages read for KU is not what one might think. Amazon uses the “last page sync” signal, which is a feature of Amazon’s Whispersync, to determine how far somebody reads. This data point can in fact be easily manipulated to a scammer’s advantage.
For example, one could place a link early in the book promising a $100 Amazon gift voucher, but that link could take the reader to the last page of the book. If the book has 10,000 pages, Amazon would now think that the reader has actually read 10,000 pages even if the “reading” took place in a matter of seconds.
Obviously that can’t be true, but Amazon’s algorithms don’t check that. Computer code is not imbued with “common sense.”
The first questions is, why Amazon is not looking at all “last page synced” data points and checking against the maximum possible reading speed? Doing so is not difficult, but certainly requires more effort. It seems on this point that the Amazon developer or development team took a shortcut and did not take into consideration how the system might be gamed.
Well, we can be sure they are alert to the situation now, as the problem, in technical terms, is not terribly difficult to fix. It’s worth noting, though, that Amazon does not have infinite resources, at least not in terms of developers, and there is a war for talent throughout the tech industry, especially out on America’s west coast. Thus, it might not be the highest priority for Amazon at the moment even if it might be easy to fox.
Amazon also has many, many other priorities, and its ebook marketplace is just one of many things the company does. So adding an extra layer to check whether the signal is realistic and stop scammers that way would be incredibly easy for Amazon. Being experts in reading analytics, we at Jellybooks are quite confident of that assertion.
Some commentators in the debate have asked if Amazon can check whether individual pages were read. It almost certainly can. The provider of an ebook reading application or device has much greater capabilities than a third party (like Jellybooks) with regard to extracting information, and one such piece of information is the “page-turn.” The majority of reading applications record that data point, and several also send it back to the app providers, though not always in real time (not every page-turn is reported back when it happens, but are instead sent in certain time increments, just like “last page sync” is sent in time increments or when the app closes).
Amazon almost certainly has that data, but it might not have that data in a consistent format. There is a large number of Kindle apps and devices in the market, and there is a legacy of older apps and devices that have been built over nine years. It is probably no coincidence that Amazon recently released a major software upgrade that was mandatory for all users.
We speculate that Amazon may not have used page-turn data, because it does not have it in a consistent format or not for all its users.
In addition, Amazon can check chapter opens and closes just as Jellybooks does. This would reveal in an instant if a large portion of the book is in fact skipped and allow them to adjust pay-outs accordingly.
Ultimately, if this much money is at stake, there will be scammers trying to get their hands on it. However, by making it economically unattractive to game the system, Amazon can make it fairer and more equitable for genuine and honest authors. The current system of “last page sync” is a very poor determination of how somebody reads a book and thus far too easily gamed.
This of course also raises the question of how accurate the “completion rates” are that Amazon reports to authors and select publishers. If it uses the same methodology as above, then this is highly suspect, as well. If we at Jellybooks could get our hands on Amazon’s data trove, then we would analyze and present it in much more depth using our “Candy for Publishers” service than Amazon itself ever would.
In Amazon’s defense, I should note that we have an unfair advantage in that we focus on nothing else but reader analytics and its many applications. Thus, we think much, much harder about the problem than a large ecosystem operator who has many other things and other priorities to consider might.
So let’s give Amazon a break and a couple weeks to fix this. We would still like to work on that data set from Amazon, though, for the benefit of readers, authors, agents and publishers. And we know you are open to that idea, Seattle, but as always, being in publishing is a game of patience.
We know you have been watching us, Seattle. You tell us so in every email. We are ready when you are!
My earlier writing on ebook subscription models:
Earlier posts in the data-smart publishing series:
• “The Internet of Bookish Things”
• “Reading Fast and Slow – Observing Book Readers in Their Natural Habitat”
• “Start Strong or Lose Your Readers”
• “What Books Have the X-Factor? Measuring a Book’s Net Promoter Score”
• “Men Are from Mars, Women Are from Venus, But What About Readers?”
• “How Does Age Affect Reading?”
• “8 Reasons Why People Buy Books”
• “Data Vs. Instinct – The Publisher’s Dilemma”
• “It’s the Cover, Stupid! Why Publishers Should A/B Test Book Covers”
• “Foreign Rights and Reader Analytics”
To get all the ebook and digital publishing news you need every day in your inbox at 8:00 AM, sign up for the DBW Daily today!