You can trust computing to turn what was simple into a system of Byzantine complexity.

Before the days of electronic books, the author’s royalties were relatively simple: a matter of contract between publisher and author. If you were lucky, your publisher gave you a generous deal, assuming that your book would only sell modestly, and they needed it for the catalogue. This became more complex with advances, book club deals, and so on, but was basic arithmetic and no more than one side of the paper for calculations.

Then along came Kindle and the iTunes Store, which of course publish under completely different forms of contract.

Amazon’s basic system has been to allow authors to opt for 35% or 70% royalty schemes, as detailed here. At first sight you would think that everyone would try for 70%, but it is not as simple as it seems. First the 70% option only applies to purchases made in certain countries, then the 70% is only paid after allowances for ‘delivery costs’ (calculated according to the size of the book in MB), and there are restrictions in book pricing, and the whole thing changes further when Amazon has promotional pricing. That is detailed here.

Not only that, but there are special arrangements for ‘sales’ through the Kindle Unlimited and Kindle Owners’ Lending Library schemes, as they are paid for by subscription.

Those used to pay according to the number of loans, much like a library royalty. But as of 1 July 2015, the royalty paid on such loans will depend on the number of pages read – a far more complex system outlined here. So if your book is borrowed a great deal, but few borrowers read much of it, your royalties under this new system will all but disappear.

Of course Kindle books do not have the same fixed pages as in printed books or some iTunes books: they reflow according to view size. So to cope with that, each book is going to be assigned a standard page count, based on an unreleased algorithm called the Kindle Edition Normalized Page Count (KENPC v1.0). Although Amazon says that non-text elements such as images, charts and graphs will count, it does not explain how that will be calculated.

Peter Wayner has already speculated in The Atlantic as to the effect of this new system, and the story even reached Slashdot.

However it is obvious that, without deploying sophisticated monitoring systems or asking comprehension questions afterwards, Amazon cannot possibly know which pages have been read, only which have been opened. At first sight this should make facile gripping page-turner fiction prosper under suscriptions, and drive out any books which are unlikely to be read from cover to cover. Far from being equitable across genres, it is likely to drive greater supply of pulp fiction. Perhaps we will even see books written to exploit the KENPC in the same way that ‘search engine optimisation’ does for commercial websites.

Not content with changing the publishing business, Amazon seems determined to change literature as a whole. For the better? I think not.