Some very bright folks pointed out that the data used for the survey on the 26th was actually pulled from Amazon's Popularity index, rather than their Bestseller index. What's the difference? Popularity is the default index readers see when they just hop into a genre and browse. That makes it an excellent snapshot of what's in front of readers at any given time. However, Amazon has set up the popularity algorithm to tend to favor books in the Select program, by "carrying over" a certain level of the popularity earned when a book goes free. Book goes off free, it lands on the paid list again, and as soon as it sells a little the carried over popularity catapults it up to the top of the list. That makes the Popularity listing volatile, a constantly shifting array of books.
Which is awesome for Amazon, from a sales perspective. You want your readers to see a new batch of books every time they visit. Churn is smart. But it's not quite as good if one is collecting raw data on the state of sales - as we'll see.
Amazon also has a Bestseller list, for ebooks, by genre. This is a much more static listing, made by compiling raw sales. It's updated hourly, but books that get there tend to stay there for a while. So what might the difference be, if we compared Bestseller status with the constantly changing Popularity index?
To answer that question, I've compiled the same set of data on Amazon's ebook bestseller list, for Science Fiction. As an added bonus, I've also compiled the same data (bestsellers) for fantasy. There are some interesting differences between the two data sets, as we'll see.
As per the previous survey: EVERY attempt was made to ensure the data was as accurate as possible. Publishing companies owned by the author were counted as self publishing. Publishing companies which publish any submitted book for a fee were likewise counted as self publishing. Publishing companies which in any way vet incoming books or have a submission process were counted as traditional publishers. Whenever a question existed whether a publisher was trad or indie, I counted it as trad.
Science Fiction Top 99
When I collected this data, there was a glitched duplicate copy of "Dance of Dragons" on the top 100. The listing said unavailable, but was still showing in the top 100. Yes, we know Mr. Martin sells a lot of books. You don't need to show us two copies of the same book in the top 100 to prove it. ;) Anyway, I simply counted the other 99 listings, rather than count the same book twice.
The breakdown was 38 (38.4%) traditionally published books and 61 (61.6%) self published books. Of the 38 traditionally published books, 22 were first published ten or more years ago; only 16 were "recent" releases.
Pricing for SF
I've broken out pricing again by price, and by indie/trad. I've wrapped prices +/- 21 cents into the corresponding $x.99 category, for simplicity.
$0.99 - Indie 22, Trad 4, Total 26 (26.3%)
$1.99 - Indie 2, Trad 0, Total 2 (2%)
$2.99 - Indie 22, Trad 2, Total 24 (24.2%)
$3.99 - Indie 5, Trad 1, Total 6 (6.1%)
$4.99 - Indie 9, Trad 5, Total 14 (14.1%)
$5.99 - Indie 1, Trad 3, Total 4 (4%)
$6.99 - Indie 0, Trad 2, Total 2 (2%)
$7.99 - Indie 0, Trad 8, Total 8 (8.1%)
$8.99 - Indie 0, Trad 2, Total 2 (2%)
$9.99 - Indie 0, Trad 2, Total 2 (2%)
$10.99 Total 0
$11.99 Indie 0, Trad 2, Total 2 (2%)
$12.99 Indie 0, Trad 2, Total 2 (2%)
$13.99 Indie 0, Trad 4, Total 4 (4%)
$14.99 Indie 0, Trad 1, Total 1 (1%)
For a visual reference:
Once again, we're seeing solid prices from many indie publishers. The 99 cent bracket was set for only 36% of indie books. About 25% had priced their books above $2.99. It will be interesting to watch how that pricing trend continues into the future.
Overall, while indies maintained a solid lead in the bestseller list (61 of 99 books), they lacked the complete dominance seen in the Popularity index - likely due to decreased influence of the Select free periods. Good information to have.
Fantasy Top 100
The fantasy results were substantially different from science fiction - surprising on the surface, given that the two genres are often merged in bookstores, and are part of the same category on Amazon. But fantasy readers showed a marked lack of attention for older works of fantasy. Of the 47 traditionally published titles, only four were first published over ten years ago. Over 90% of trad pub fantasy bestsellers were relatively recent books.
Overall, self published books represented only 53 of the top 100 books, substantially less than we saw in science fiction.
As we'll see, fantasy readers also showed substantially more tolerance for higher prices, both from indies and from traditional publishers. The lack of indie writer dominance in this genre is probably related to this price tolerance.
I've broken out pricing again by price, and by indie/trad. I've wrapped prices +/- 21 cents into the corresponding $x.99 category, for simplicity. Additionally, one $2.51 book was counted as $2.99 (indie) and one $7.39 book was counted as $6.99 (trad).
$0.99 - Indie 13, Trad 0, Total 13 (13%)
$1.99 - Indie 2, Trad 0, Total 2 (2%)
$2.99 - Indie 23, Trad 0, Total 23 (23%)
$3.99 - Indie 12, Trad 0, Total 12 (12%)
$4.99 - Indie 2, Trad 1, Total 3 (3%)
$5.99 - Indie 0, Trad 0, Total 0 (0%)
$6.99 - Indie 0, Trad 1, Total 1 (1%)
$7.99 - Indie 0, Trad 20, Total 20 (20%)
$8.99 - Indie 1, Trad 11, Total 11 (11%)
$9.99 - Indie 0, Trad 3, Total 3 (3%)
$10.99 Total 0
$11.99 Indie 0, Trad 1, Total 1 (1%)
$12.99 Indie 0, Trad 6, Total 6 (6%)
$13.99 Indie 0, Trad 0, Total 0 (0%)
$14.99 Indie 0, Trad 3, Total 3 (3%)
$29.99 Indie 0, Trad 1, Total 1 (1%) (Martin boxed set)
For a visual reference:
This is a really interesting graph. What we see is a clear divide down the middle, with almost no overlap. On the one side, indie writers publishing mostly at the 99 cent, $2.99, and $3.99 price points. On the other, traditionally published books dominated by $7.99 and $8.99 pricing, with a scattering of higher numbers in sufficient density to be statistically significant.
Clearly there's a higher tolerance for steeper prices in this genre. It's likely that indies could raise prices somewhat and still be effective sellers in fantasy. The steep valley sitting between the blue and red cliffs represents a potential area of additional profit for fantasy writers.
I intend to do more of these surveys down the road. Watching the changes over time will be interesting, and I believe it will help empower writers to make better, more informed decisions about their publishing options and pricing of their products.
It's clear that there is substantial variance from genre to genre. Even in two genres typically seen as related, even overlapping, the data showed a large difference in both indie market share and in popular price points. Other genres will likely have their own secrets to reveal.
Lastly, while it appears that bestsellers are not dominated by indies to quite the degree the Popularity index suggested, self published books still have a majority of the bestseller lists in both genres surveyed, especially in science fiction. This would have been impossible a year ago - unthinkable two years ago! It's fascinating to watch this sort of dramatic change unfold.
Thanks for the comments on the last survey. I hope this data proves as useful to everyone as the other, and hope you'll continue to add your comments here! I enjoy hearing from folks who've gotten some value from what I'm doing here.