The Secret Anatomy of KDP Select

KDP-Select_smallI’ve mentioned before the all-powerful shadow of Amazon algorithms, looming nigh incomprehensibly over any publishing endeavour. That this was not just a great hyperbole, I hope to show in this post, using the experience of four consecutive KDP Select promos in as many months.

1. What’s the Deal?

First, a bit of a background. If you’ve never tried to self-publish with Amazon’s KDP, you may not be aware of the Select program. It’s a program of exclusivity with Amazon’s platform; for 90 days you may not distribute your e-book in any other way. In exchange, your book can be borrowed for free by users of Amazon Prime (which grants a sometime hefty additional income, if you’re lucky) but more importantly, you get 5 days to give your book away for free.

Why would you want to give the work of your life for free? Well, because that’s the best possible way to game Amazon’s algorithms. “Game” is a harsh word here – “utilize” is better, since Amazon seems to have deliberately built their system around Select.

In fact, if you believe hearsay (and there is no other data than hearsay here. No company would ever divulge anything about their search algorithm) Amazon realized Select has too great a power over its market and worked to stem its influence a little. If you read self-publishing blogs from two, three, four quarters 😉 ago, a common theme is the reducing effectiveness of Select promos year after year. Perhaps we will see the last throes of the system this year – or perhaps not. Either way, contrary to what many authors report, the promo still works. It just doesn’t work miracles.

Select Chart 1
Sales by day since September 2012. Stars indicate promo days.
2. The Books Have to Move

I used to work in a brick&mortar bookstore many years ago. One of the first rules I’ve learned is that books have to move around. Even the greatest best-seller must one day be removed from the prime slot, to make way for a newcomer. There are many reasons for it, but the rule is sound, and all good stores use it. Somebody recently explained it in a perfect way – I can’t remember the source of the quote, please comment if you know:

“If you have a book like the Lord of the Rings, which will always sell 20 copies a day no matter what, and a new book which might sell 15 copies a day, there’s no point in keeping the Lord of the Rings on the best-seller display. You move it to the back, and let the readers discover the new book, instead.”

This is what Amazon’s algorithms do: they change the discoverability of the books around. They are the equivalent of the bookstore clerk. And just like the bookstore clerk needs to be told what books to put on which shelf, so do the algorithms need to learn about your book. This is where KDP Select comes in. A successful promo tells the algorithm: “this book has potential. Put it on display instead of that other one. Let the people see it.”

Select Chart 2
Cumulative Sales by Day. Stars indicate promo days.
3. What is a “Successful” KDP Promo?

Let me tell you first what a not successful one looks like. You will notice in the charts that my third promo – third star – had almost no impact on the sales. This was my benchmark: a giveaway lasting only one day, one that I almost didn’t mention anywhere on the social networks, and didn’t buy any ads for, just let it run its course. It was an abject failure, by any measure.

A successful giveaway must count in thousands. Three thousand is a good start. Ten thousand is better. Breaking into the best seller charts for free books is a must; breaking into Top 10 in your genre is great; breaking into Top 100 total is a guarantee of a long-lasting success.

As you can see from the charts, each giveaway resulted in bigger spike in sales than the previous one. But also, each time I gave away more books than the last time. But there is something else you can read from the charts, something that’s very interesting and tells you a lot about the power of the algorithm:

Regardless of how big the initial spike was, each bump eventually drops down to pre-spike levels. And fast.

selectchart_3
Sales by Week since September 2012
4. We Know Major Tom’s a Junkie

The drops in sales are automatic, regular, unstoppable and easy to predict after a while. Once the algorithm asserts that your time in the spotlight is up, that’s it. The sales can halve overnight, without any apparent reason. And because your success was too quick to build any loyal following (see below) it eventually fizzles out without a trace.

This is a brilliant strategy – for Amazon. The Select quickly becomes addictive. The spike in sales is like a heroin rush, and the drop is like a withdrawal downer – with the promise of another rush as soon as you succumb to another exclusivity period. And of course, the strategy would not work if the program wasn’t so damn effective.

It’s all in the scale. The algorithm shows your book to millions of readers; there is no ad that reaches more people, no social network campaign. And yes, most of them will not be interested in it; others will just download it for the heck of it, and never read it. But a tiny percentage here equals a whole lot of people.  And as this last chart shows, this tiny percentage of readers will likely move on to your next book, and then the next; and the algorithms will pick up your other books and present them to other readers, and so on – the wheel will keep turning, slower and slower, until it grinds to a halt eventually, unless you go for another promo. But before it does, you will have sold more books than you could have imagined.

selectchart_4
Sales of “The Year of the Dragon” saga, by volume. Volume 1 is the only one that’s ever been in KDP Select.
5. So What’s the Bad News?

So I suppose the only question that remains is: what are the negatives of using KDP Select?

There is one distinct disadvantage of this system. It misses the target. If your book’s demographics is broad enough, this may not be a problem; then again, if it’s broad enough, you may not need to use Select at all: your book is likely to sell on its own merit. 

But if you had a specific target in mind, then using Select is the equivalent of trying to shoot at ants with a double-barreled shotgun. Sure, you will hit a few people you wanted to read your book, but in the process you will reach hundreds who couldn’t care less about what you wrote.

If you want to build a loyal following, if you want to reach fans, Select is not the way. You have to do it slowly, in the old-fashioned style.

PS: There are at least two other ways to do what Select does, without exclusivity: use your friends and fans to “bum-rush” the charts, or pay for an expensive ad on one of the few remaining sites that reach thousands of eager readers. I haven’t tried any of these methods yet, but I’ve seen both of them work well.

All the above caveats remain, however: a flash-in-the-pan success is never a good way to build a stable following. It may only serve as a foundation for the real hard work.

You can trace the success of my last Select promo yourself, as between February 4th and 6th “The Shadow of Black Wings” is once again FREE on Amazon.

E-publishing: solid numbers

publishers-weekly-logo_smallI was looking for some numbers on e-book sales per device and store, and found this article on Publisher’s Weekly.

Here are some highlights:

“According to the Book Industry Study Group’s fourth volume in its “Consumer Attitudes Toward E-Book Reading” survey series, 73% of e-book buyers bought (or got an e-book for free) from Amazon, with 21% getting their e-books from B&N (…) Apple’s content stores were only used by 10% of e-book users, but that is expected to change as device sales pick up.”

That’s just for the US, though. I expect B&N was non-existent anywhere else in 2012 – it only just started expanding into Europe at the end of the year. Kobo probably takes B&N’s place in other markets.

I’m not sure why everyone expects Apple sales to pick up, it’s not like they haven’t been selling their devices in droves before – or is everyone pinning their hopes on iPad Mini? If that’s the case, why doesn’t anyone mention Google Play, which is just as expansive globally, and available on more devices? Especially in the light of this:
“iPad is the leading tablet used on Books-OnBoard, representing 63% of all tablet downloads, but Android tablets are growing share rapidly. A year ago, iPad completely dominated this, with 93% of tablet downloads.”

Curious bit about power buyers:

“According to the BISG’s consumer reading survey, “power buyers” (those who purchase e-books weekly) show an increased preference for reading on tablets, with more than 38% indicating so, compared to 19% a year ago.”

“The Kindle Store, for example, accounted for 46% of the e-book purchases of Galaxy users compared to 83% of Fire owners, while “other” outlets represented 19% of e-book purchases of Galaxy owners compared to 5% for Fire owners.”
I wonder what are the most popular “other” people use?

And here’s the bit I was looking for:

“50% of the respondents to a Diesel survey reported that they use dedicated e-readers, 20% use a personal computer, and 16% use tablets, but the tablet component is growing.” So that’s only 20% PC purchases. I presume most of them are still using branded bookstores instead of author websites. That doesn’t bode well for the author-vendor model.

“LiVolsi also pointed to a “migration of about 34% of our readers over the last 18 months” to sub-$100 Nooks and Kindles that don’t support content from other retailers.”

It will be interesting to see where we are next year…