Posts Tagged ‘Tech Humor’

Thoughts on Kindle Unlimited and Scribd

Friday, July 3rd, 2015

Some of you may know that Amazon changed the terms of its subscription service, Kindle Unlimited (KU) such that payments due to authors with books in KU are calculated in a different manner than previously. If you’re a reader and you subscribe, you can read all you want for $9.99 a month. With the single limitation, so far, that you can have up to 10 books on your “shelf” at once. To get book number 11, you have to read or release one of those books.

With the Kindle Unlimited subscription you can access hundreds of thousands of Kindle books and thousands of audiobooks with Whispersync for Voice. You can keep up to ten books at a time and there are no due dates. Read your Kindle Unlimited books on any Amazon device, or free Kindle reading app. (Terms)

Scribd reinvented itself from a pirate site reader-centric sharing site (Irony ALERT!) into a subscription service. For $8.99 a month. They paid all authors/publishers the same as a sale.

If you’re a reader, that’s a pretty sweet deal, assuming the books you want to read are in the program.

If you’re an author, deciding whether to have a book in KU is a business decision, and not everyone’s business needs and goals are the same. Everyone’s reasons for being in or out are different. Last year when KU debuted, I blogged about it here. Here’s what I said then about how that would be profitable:

If you are paying authors/publishers a percentage of price, then for your business to be viable, that payout amount per month HAS to be less than 9.99 * (number of users subscribed).

This means a profitable user will read a number of books N per month where the payment due to vendors is less than 9.99. The more books they read, the less the wholesale price has to be (obviously), and, at 9.99 per month, the wholesale price has to be less than 4.99 for 2 books per month, 3.99 for 3 books, etc.

Not long after that post, it turned out the payment terms for traditionally published books in KU were different than for self-published books. Traditionally published books receive the same payment as if the book had been bought — that is 70% of the purchase price. Further, certain self-published authors were given those or similar terms in order to convince them to put their books in the program.

Self-published authors can only participate in KU if they put their books in Kindle Select — that is, have those books exclusively at Amazon. Scribd does not require exclusivity. For some authors, Kindle Select makes sense. But for others, it doesn’t. Doing well at other vendors or wanting to avoid the risk of having a business depend on a single vendor are good reasons not to be in Select and therefore, not in KU.

Traditionally published books need not be exclusive. Because, as Amazon recognized, that would be a non-starter.

Arithmetic

What the Romance community knew, and what I suspect Amazon knew (because DATA!) and what Scribd apparently did not know (Because why would anyone pay attention to what goes on with those books women read?) is that Romance readers are the Great White Sharks of the reading world. They are the 80 in the 80/20 rule. They are the power in a power law.

Solving for X

Remember my ruminations over profit, book prices and that monthly subscription rate? Amazon had the data that would have told them everything they needed to know about those Power Readers (before KU debuted). Amazon solved the math problem with deep pockets but also by offering self-publishers a substantially worse deal. The KU reimbursement rates started decently, then took a swift dive until the reimbursement fell to around $1.34. Why? Well, either you sustain losses because of the Power Readers or you find a way to compensate for that. Falling KU reimbursement rates were exactly that, that is, KU’s “flexible” reimbursement rates to self-published authors was their hedge.

As KU continued, Amazon kept talking about how much money they were putting into the monthly fixed KU pool to be distributed to the self-pubbed authors, but reimbursement rates from that pool continued to fall. Because the hedge was needed. (So I speculate.) Scribd had no such hedge in its business model. (To my knowledge, anyway.)

How did Scribd solve for X? They didn’t. It’s hard to understand why Scribd thought $8.99 for all readers was viable even in the medium term. If they knew about Power Readers then they either didn’t know enough or they thought the same thing most of the traditional world thinks about products for women. How could they possibly matter when they were up against REAL books and REAL readers?

$8.99 is a brilliant strategy for competing for potential KU subscribers. It’s not a brilliant strategy for paying authors/publishers in an environment that includes Power Readers. The rational solution after the short to medium term is to introduce tiered subscription rates. It’s blazingly obvious that in an environment that includes Power Readers you must also have a bazillion 1-2 book a month readers or you have to charge Power Readers more. Or you have to pay authors/publishers less. Scribd did a great job going after traditional publishers, and they probably had a better selection of books than Amazon. And, by the way, the word is lots of Power Readers (those sharks!!) had subscriptions to both services. Because the pool of books was different.

But if they charged those readers more, then KU looks more attractive… It’s a tough situation.

Solving for Y by Killing X

Scribd’s solution was to remove 80-90% of Romances from their service.

Sure. Of course. Now they will be paying out less to authors and publishers because the books people women actually want to read are gone. Now that they’ve basically told the Power Readers they are unwelcome with all their womanly reading of THOSE books—who the hell knew they read that much???—what they have left are the 1-2 book a month readers.

This makes a certain sense. Because maybe what will happen is the Power Readers keep their subscriptions to both Scribd and KU, but now only borrow 1-2 books from Scribd and things are sustainable for a bit longer for them. Yes, an FU to romance readers, but Scribd maybe wasn’t in a position to feed the sharks.

If I were a Romance publisher ::cough::Harlequin/Avon::cough:: who just put substantial backlist into Scribd only to have their reader base told to fuck off, I think I’d be pretty pissed off.

The more established self-publishers, the ones who cannot afford Amazon exclusivity financially or at the cost of reader-relations will likely move to Oyster in order to have some presence in a subscription system. I wonder if Oyster knows what’s coming their way?

Cue the theme from Jaws….. LOOK OUT OYSTER!!!

Segue

Early on, long before KU, I put one book into Select into order to have data on the program. I asked my newsletter subscribers to tell me what they thought about my decision. Their answer? The non-Amazon readers were angry. Rightly so. That was enough for me. My experiment was done after the first angry letter. (After 90 days, you can elect not to re-enroll in Select.) If it had been possible, I would have ended it immediately, but I had to wait out the 90 days. I sent a copy of that book to every single reader who let me know how they felt.

Amazon’s Adjustment

The initial structure of KU with its fixed reimbursement pool meant that a longer book that make $2-4.00 for a sale, made $1.34 in KU. Shorter books, on the other hand, that would be sold in the $0.99-1.99 range and thus net the author a dollar or less, made $1.34 in KU. In other words, a book priced at $0.99 made $1.34 in KU. Anyone with half a brain can see that this meant shorter books were way more profitable and that longer books were way less profitable.

The adjustment Amazon made was to address that disparity. Instead of paying the same amount per borrow regardless of length, authors are now paid based on pages read. “Pages” read, actually. Basically, Amazon had to normalize what a page means for a digital book when displays are reflowable and resizable across different sized devices. A “Kindle Page” is the same for all devices regardless of settings. (Presumably, of course.)

To me, that’s fair enough. Authors who write shorter books make up the difference by writing more books. I should think that’s obvious, though apparently not. Category authors tend to write more books than single title authors. Three 30K word books will make you the same as one 90 page book, assuming the books are read all the way through.

I have to shake my head at the suggestions from some that readers should make sure to page through shorter books, because otherwise those authors are screwed.

No they’re not. They’re only screwed if readers never actually finish the books, and if readers aren’t finishing their books, well, maybe those authors should worry about why that is. There absolutely is a market for shorter books and short stories. Just like there’s a market for longer ones. I have short stories, novellas, and novels on sale. They achieve different goals for me. I’m quite sure that readers have different goals and preferences for reading works of varying lengths.

Final Thoughts

I don’t have any books in KU. I did have books in Scribd, but I assume the only thing left is Scandal, which is currently free and so would not have been removed. I’ll probably go pull Scandal because I’m vindictive that way.

But now I’m kind of wishing I did have something in KU because at last at LONG LAST Amazon is giving authors data about how much of their books get read, but the only way to get it is to be in KU. I had this idea that authors could put a book in KU, let it sit for 90 days and watch the data about pages read. You’d rewrite if no one gets past Chapter 10. ::snort:: Mostly I’m kidding.

[Update: MelJean Brook pointed out that Amazon is NOT providing meaningful page read metrics so my plan would not work. There is no way to tell from the data provided if 2000 Kindle pages read is 2000 people reading one page or one person blowing through 2000 pages of an author’s work.]

I Lied. This is the Final Thought

I was talking to a friend the other night about why Amazon didn’t fix their issue sooner since they surely had the data about the problem of shorter works no later than 6 months in. Assuming that’s true, that gives them 6 months to develop, test, and QA and then prepare the PR for the Kindle Normalized Pages scheme. This is aggressive but doable. You’d have to test a lot of scenarios and then make absolutely sure all the calculations are correct and reach consistency.

Maybe the schema changes weren’t as big a deal as they would be in a traditional SQL Server or Oracle environment, but NoSQL solutions have different challenges, and one of them is hidden errors because of eventual consistency or problems with “schemaless” documents. (It’s only schemaless if you never hired a data architect, and if you didn’t sooner or later you’re fucked. *)

I’m thinking of Wattpad and its problem with user comments attributed to the wrong account. That’s a total NoSQL error that a good OLTP-trained data architect could have said, hold on a sec here… What happens if…. And then all the developers stick their fingers in their ears and sing LahLahLahLahLah because the architect just added 3 months to the delivery date. And nine months later your data is untrustable. There are scores of developers out there who got burned by thinking schemaless means never having to think about data consistency across transactions.

Eventually, your financial data has to be in a transactionally consistent state and stay that way and it can never ever revert to a previously inconsistent state. Or you can’t pay people correctly. So, you know, 6 months seems like a decent guess for how long it would take to roll it out and be certain it works for paying people reliably. The concept isn’t hard. The execution is.

Interesting.

* OMG. I actually made a database joke in a writing blog! More than one, actually. This is very strange.

Note: Regarding NoSQL, it’s a very very fast way of scaling data. Although UC Berkeley had one of the early such databases, Amazon more or less put the concept into widespread use, followed by the original developers at You Tube who had to massively scale MySQL. Those guys needed to ramp up fast and on a scale that traditional transactional database could not then achieve. When I say “documents” in the sense of a NoSQL database, I don’t mean a Word document. I mean a collection of information of related items where Item 1 may not have the same information as Item 2 in the same set of related information. In that sense, there is no “schema” (that is a definition of what information is contained in related data. In a transactional database, all objects of a defined type have the same structure, even where elements of the structure are NULL.)

The NSA, by the way, collects your information in Hadoop, a NoSQL database backed up with some Postgres SQL functionality for the sorts of transactions that MUST be consistent.

This is a laughably high level explanation. It’s way more complicated. I’m a SQL Server DBA and Data Architect, but I’ve done some Mongo DB where we needed to address some shortcomings with our SQL Server applications without spending a fortune. For anyone who cares, Microsoft’s SQL Server 2014 changed the query optimization engine in significant ways — and I suspect it’s a direct response to NoSQL. For example my current employer had ugly queries that were taking 2 minutes (on completely under resourced SQL 2008 servers and for data that SHOULD have been in a datawarehouse but wasn’t, so I’m sorry, but the situation is long and convoluted and no one here cares, just know that 2 minutes for a query result is beyond embarrassing) that went down to 45 seconds when run on a SQL 2014 install.

Basically, the point is that the situation is considerably more complicated than, hey, let’s do it THIS way instead. Amazon is not just a company that sells stuff. They INVENTED the technology they needed to massively scale because no one else was doing that, and then they open sourced it. So when we talk about Amazon having advantages, the advantages are even bigger than most realize. Amazon IS data. I don’t think they do anything without knowing what the data says, and they have more data than anyone.

It’s why we’re seeing such an upheaval in publishing. It’s why Romance matters more and it’s why companies and analysts who dismiss Romance are in big trouble. Amazon knew about Power Readers. The usual gendered biases very likely got exploded by the facts. Traditional publishers need to lose the bias. Companies who want to compete in this space need to fire anyone who talks about REAL books and REAL readers.

The Romance Sharks will eat their lunch.

Share