Two months ago, I took a look at how many of the bestselling Kindle genre titles were self-published. I had two purposes in mind: first, to see whether there were any differences in the success of self-publishing between the big four genres (Romance, Mystery/Thrillers/Suspense, Sci-Fi, and Fantasy).
The second purpose was to provide some more data for the initial Author Earnings report. The report indicated that self-publishers were doing incredibly well within genre ebooks, but one of the more widespread criticisms was that the report was just a snapshot that might not represent anything more than that moment in time.
I thought that was a valid critique, but I also suspected it would prove false — Amazon is amazingly consistent from day to day and month to month, and the AE report looked at a substantial chunk of data. I was betting that later studies would show similar results.
Among the report’s conclusions was that genre fiction accounted for about 70% of all Kindle ebook sales, and that self-published titles accounted for roughly half of that. I used a different methodology, and a worse sample size, but when I checked in February, self-publishing’s share of the bestselling Kindle titles was as follows:
- Romance: 49%
- Mysteries/Thrillers/Suspense: 11%
- Science Fiction: 56%
- Fantasy: 49%
Three of the four genres were roughly 50% self-published, with the glaring exception of the thriller market. Meanwhile, here was each genre’s overall share of the Kindle market (methodology explained in the original post):
- Romance: 40%
- Mysteries/Thrillers/Suspense: 20%
- Science Fiction: 5%
- Fantasy: 6.33%
This added up to 71.33% of all Kindle ebook sales. I pulled my numbers a few weeks after the first Author Earnings report collected its data, yet my conclusions mirrored theirs: about 70% of all Kindle sales were in these four genres, and of those sales, close to half were of self-published titles.
It’s been two months since then. How do things look today? First, here are the four genres broken down by method of publication — self-published; through a small or medium press; Amazon publishing imprints; and by the Big 5, which includes major genre houses like Harlequin and Baen, where appropriate.
ROMANCE
- Self-published – 59%
- Small/medium – 3%
- Amazon – 12%
- Big 5 – 26%
MYSTERY/THRILLER/SUSPENSE
- Self-published – 26%
- Small/medium – 1%
- Amazon – 15%
- Big 5 – 58%
SCIENCE FICTION
- Self-published – 53%
- Small/medium – 7%
- Amazon – 12%
- Big 5 – 29%
FANTASY
- Self-published – 45%
- Small/medium – 6%
- Amazon – 8%
- Big 5 – 41%
There are a few differences between the first grab and this one. The percentage of bestselling self-published romance titles is up by a good percentage. Thrillers are way up, more than double the initial look. Meanwhile, self-published sci-fi and fantasy titles are slightly fewer. Amazon’s publishing imprints are up, representing just under 12% of the total, compared to a little over 9% the first time.
I wouldn’t draw too much from any of these changes, though. You can hardly conjure a pattern out of two whole samples drawn from a methodology that’s prone to variance. What’s most interesting to me here is how little is different: in three of the four major genres, self-published titles still represent about 50% of the bestsellers. Thrillers continues to lag behind, but this month’s look suggests it’s not quite as tough for self-published titles to compete as the original breakdown suggested.
Okay, so what about the genres’ overall market share? Here’s how it breaks down this time:
- Romance – 35.2%
- Thrillers – 26%
- Science Fiction – 5.4%
- Fantasy – 6.4%
This adds up to 73% of overall Kindle ebook sales. Crazy.
Compared to February, sci-fi and fantasy are essentially the same. Romance is somewhat smaller, but thrillers are up by a decent percentage. As before, however, I wouldn’t try to read patterns in the differences — I’m not at all sure that romance sales are actually down. The sample sizes involved make this part of the data prone to a fair amount of variance.
Again, what’s most interesting to me isn’t the differences. It’s how similar these numbers are a full two months later — these four genres continue to comprise ~70% of Amazon’s ebook sales, and roughly half of those sales are of self-published books.
From the department of “It’s About Damn Time,” I’m happy to announce I’ve set a new Breakers book helpless into the world. Go catch it! Quick, before it escapes!
______________
AVAILABLE AT:
______________
I’m happy to have added Google Play to my list of distributors. For those of you here for the publishing-related stuff, I’ll try to get up a post about them before too long.
Until then, buy early and buy often!
Yesterday, Dear Author blogger Sunita raised the idea that self-published genre fiction is creating a market for lemons–an environment where readers have no easy way to identify good books from bad books. If true, the author argues, this would be a very bad thing: if readers have no way to tell good from bad, many will simply quit reading altogether, turning to other media instead.
The argument goes like this: on Amazon, the chief ways to determine whether a book might be good are a) price and b) reviews. Yet both are highly flawed. In other markets, higher prices are usually an indicator of better quality. But with ebooks, you’ll often find a New York-published bestseller priced the exact same as a completely unknown self-published title. Thus price tells us nothing about whether a book is likely to be any good.
Reviews are no better. As evidence of this, Sunita points out that bestselling genre fiction typically has higher ratings than literary classics like The Great Gatsby. Self-published bestsellers have even higher ratings than the classics within their genres. Hugh Howey’s Wool, for instance, is shown to have better ratings than works like Ender’s Game, Cryptonomicon, or Neuromancer. Since the author can’t believe Wool might actually be better, Amazon’s reviews clearly aren’t useful for helping readers find good books, either.
It seems to me the discrepancy in ratings is evidence of a much simpler possibility: there is no problem at all. The system is working perfectly.
If the reviews are better on popular, bestselling genre fiction than on the classics, maybe what that means is.. genre fans enjoy genre fiction more than the general populace enjoys the classics. Classics which, incidentally, are largely recommended through word of mouth and trusted sources like reviewers and critics–who Sunita states are the best ways to discover new writers. Yet reviews are better on self-published bestsellers, whose initial popularity is generated almost entirely through Amazon’s recommendation system. Wouldn’t that mean that Amazon’s recommendation system is better than word of mouth or “trusted sources”?
Well, no. Not for her, anyway. Because she’s making two big mistakes. The first is assuming that her consumer habits are commonplace. I.e., the way she uses reviews doesn’t work well for her, therefore they must not be working for any customers. Yet the amount of people participating in the review system indicates that’s far from universal.
The second mistake is one she actually approaches in the article–and then immediately dismisses: “It’s entirely possible that readers of the Ward and Howey books were more satisfied with their reading experience than readers of the Tartt, Gibson, etc. … I have more trouble with the idea that the Ward and Howey books are better books.”
What is the difference between a “better” book and a book that readers are more satisfied with?
I think that, to many if not most readers, that’s two ways of saying the same thing. For Sunita, however, there is clearly a distinction. That’s because she only seems to recognize one area of quality: a book’s artistic or literary quality. What she’s leaving out is a book’s commercial or entertainment quality. These aren’t exclusionary. I like both. My personal favorite books are the ones that combine literary flair with strong and active plots (including many of the SF titles Sunita listed).
But I think it is beyond clear that most readers care far more about being entertained than being arted at.
Since more people are reading for entertainment than literature, Amazon’s reviews reflect those interests. Since Sunita values the opposite, it’s no wonder the system doesn’t work for her.
You know who it does seem to be working for? The readers. Who choose genre fiction 70% of the time they enter the Kindle store. And who, within those genres, choose self-published fiction as much as 50% of the time. And who leave higher ratings for both genre fiction and self-published titles.
If we’re lobbing lemons into the market, they must taste pretty god damn good.
ETA: Some cool stuff in the comments, particularly from Courtney Milan, who says smart things about the Amazon review system and the way indies interact with it. (At this point, I feel like using “smart” in conjunction with Courtney is getting redundant.) Some interesting replies from Sunita, too.
I’ll say this: it’s weird and somewhat counterintuitive that indie books average higher ratings than trad-published titles. (The main reason for this, as Courtney mentions, is probably that we push more actively for them.) Obviously, that could have implications on reader purchasing behavior–but even so, that would only matter if the books weren’t actually all that good, right? Which ought to result in more negative reviews, which would then balance things out. I’m still confused by the thrust of this post, and think its conclusions are overstated.
Inspired by the Author Earnings report, I’ve taken a quick whack at looking at what percentage of Kindle ebook sales self-publishers represent by genre. To get there, I simply look at the top 100 bestsellers in each genre—romance, mystery/thriller/suspense, science fiction, and fantasy—and split them up by method of publication. Note that, unlike the Author Earnings study, this is merely a breakdown of the raw number of self-published titles on the bestseller lists, not the number of total book sales within each genre.
Also, instead of five categories of publisher, I use four: self-published, small/medium press, Amazon Publishing, and Big 5 (including, where appropriate, major genre publishers like Harlequin and Baen). For books where the publishing method was unclear, I did a search of the house. If the house published only a single author’s works, I listed it as self-published. If the house published multiple authors, even if it was obviously an author collective, I listed it as small/medium.
Okay! Without further ado, the numbers:
ROMANCE
Self-published: 49%
Small/medium: 11%
Amazon: 9%
Big 5/Harlequin: 30%
MYSTERY/THRILLER/SUSPENSE
Self-published: 11%
Small/medium: 5%
Amazon: 16%
Big 5: 68%
SCIENCE FICTION
Self-published: 56%
Small/medium: 9%
Amazon: 5%
Big 5 (plus Baen): 30%
FANTASY
Self-published: 49%
Small/medium: 7%
Amazon: 7%
Big 5: 37%
One of these things is not like the other! At an immediate glance, one thing is clear: if you’re publishing in romance or SF/F, self-publishing is an extremely viable method. Roughly half of all the bestselling books in each of these genre is self-published. That’s pretty remarkable.
For mysteries and thrillers, however, it’s a different story. Of course you don’t have to be a bestseller to make a living as an independent author, but it’s equally remarkable that just 11% of the top 100 mysteries and thrillers are self-published. That suggests two things. If you’re a thriller author, you may want to keep querying agents. Or that there’s a market inefficiency in thrillers, where there aren’t enough good indie titles to meet demand. It’s also possible that both of those things are true! I couldn’t say.
Also, it should be said that this is just a look at the top 100 in each genre out of hundreds of thousands of total books. It’s quite possible, perhaps even likely, that a broader look at the data would present different trends. However, it does match up well with the Author Earnings study of these genres combined, so I’m not sure a bigger sample would be that much different.
Of course, there’s one more big factor here: each genre’s total share of the Kindle market. Fortunately, that’s really easy to ballpark. By looking at the #100th-ranked book in each genre and dividing that by its overall Kindle rank, we get an estimate of what percentage of the entire Kindle market each genre represents. For instance, if the #100 book in Romance were #1000 in the Kindle store, we could figure that 1 in 10 sales, or 10%, are of romance books.
Here’s how it shakes out:
Romance: 40%
Mysteries/Thrillers: 20%
Fantasy: 6.33%
Sci-Fi: 5%
You’ll note that adds up to 71.33%. Hugh Howey’s much bigger and better sample suggested these four genres comprise 69% of total Kindle sales (though it didn’t break it down by genre). To me, this means the above numbers should be pretty accurate, despite the crude methodology used to determine them.
Obviously, romance is the runaway winner. There is a huge market for it and self-publishers do very well there. Fantasy and science fiction are about neck and neck: fantasy is a little bigger, market-wise, but self-publishers have more share of the science fiction market. Mysteries and thrillers have a very big overall market—half as much as romance, and a fifth of all Kindle sales—but taking advantage of the size of that market appears to be a challenge for self-publishers.
Also, if the Author Earnings report didn’t already make this perfectly clear—holy shit self-publishers sell a lot of books. I knew we’d taken over a big part of the market. I didn’t know that, within three of the four most popular genres, we’d taken half of it.
Quick edit: I should make it perfectly clear that these percentages are very preliminary. Where the Author Earnings report samples nearly 7000 books, including about 2600 of the top 7000 titles in the Kindle store, I’m only sampling the top 100 in each genre. In a sample that modest, even a small variance from the norm might throw things out of balance. For instance, if just five of the books in fantasy were switched from self-published to Big 5, the numbers of each would be nearly equal. I will try to remember to run this again in another month or so and then again later in the year to see whether the results hold.
That said—there are several signposts the data’s pretty accurate. For one thing, among three genres, the percentages are pretty similar across the board. For another, although I divide things up differently, and am only measuring number of titles instead of number of sales, my results are pretty close to those of the Author Earnings survey—which was taken, to my knowledge, 2-3 weeks ago. The lists I looked at today were certainly comprised of many different titles, yet the number of self-published titles on both studies is pretty close. This makes it less likely that either study is an anomaly.
Ultimately, though, time will tell.
THE EXPERIMENT
Early this year, after getting excited by what the Self-Publishing Podcast crew was up to, and after seeing a friend have great success with it, I decided to try my hand at a serialized novel. Serials were clearly working for a lot of people and it looked like fun on both the writing side (new format!) and the publishing side (a new release every week!).
So I set to work, and by April, I was ready to fling mine out into the world. How did it go?
Well, for the TL:DR version, and my all-time favorite post on the matter of selling serials vs. novels, see Susan Kaye Quinn. The slightly longer version is this: there are advantages to writing serials, but they don’t sell themselves any more than novels do. So if your new release strategies are based on, say, advertising novel-length works, releasing story/novella-length episodes might present you with a challenge.
Anyway, back to my results. I wrote a time travel thriller called The Cutting Room. I decided to write 6 episodes, each one running between 12,000-16,000 words and 84,000 in total, with a TV-style arc. I found a pre-made cover from James at the excellent Go On Write and, for a few bucks more, got him to set me up with six distinct looks as well as a full-length version (a 3D box set version, and a 2D version for Apple, which won’t take 3D covers). Individual episodes looked like this:
Not an ideal nailing of the genre, but suggestive of it, and perfect for the mood. In any event, enjoy the pictures now, because a wall of text is about to follow.
OUT INTO THE WORLD!
The first episode went live April 22, 2013. I alerted my Facebook page, then sent to my mailing list the next day. Neither was huge at that time—my FB page was probably around 100 Likes, as I recall, and my mailing list around 300—but that and some advertising had done quite well for the third book in my Breakers series two months earlier.
Excitement! The first day, I sold.. 4 copies. By the end of the first week, I was sitting pretty at 31. (Amazon.com numbers only—Amazon UK and B&N probably put that around 40, but I didn’t keep records for them.)
Well.
Don’t get me wrong, that’s not bad, given the modest size of my lists and the fact I was offering them a new series in a different format. But by comparison, Breakers #3, augmented by some serious ads, had moved 767 copies on .com in its first week. By contrast, this was looking like a bust.
But the advantage of serialization is you don’t get one release, you get a bunch. Six, in my case. With so many books hanging out as new releases, they should pull each other up the charts. Ideally.
Mine didn’t. To cut to the chase, each episode performed about the same. 25-30 copies sold its first week, about twice that in its first month. In an attempt to kick things up a notch, I made the first episode permafree about three weeks in. That helped a little, but with no way to advertise it on the freebie sites (too short), there was no significant bump.
Here is a chart of my first few weeks. It is mostly made of sad.
This is how each episode fared over its first ten days. Again, Amazon US only. Sales are cumulative; i.e., by day 3, episode #1 had sold 23 copies. Each episode was released exactly a week after the first. So in this chart, Day 1 for episode #2 happened on Day 8 after #1 was released. According to my records, #1 went free the day #4 went live. Also, you’ll note these numbers don’t perfectly match up to the ones I quoted above. That’s because I didn’t start pushing the episodes until the day after they went live, so that’s where I started counting for the chart.
Anyway, not a lot to see here. Every week was about the same as the one before it. At least the few people who got into it stuck with it!
Mostly, the lackluster results were because none of my launches was ever significant enough to start getting the books recommended to other readers. I think that if my first couple days of sales had been 30-60 rather than 10-15, I would have seen growth from episode to episode. Without hitting high enough to garner an internal push from Amazon, I was selling to the same group of saps each week (my readers). (That’s a joke, my readers are the best because they read my books, QED.)
So was it a bust? Well, I’d sold a few hundred copies of the episodes, which was better than a sharp stick in the anything. But my serial didn’t really expand my audience—my primary commercial reason for this experiment—so it certainly felt like a failure at the time. So much so that, before the final episode went live, I altered its ending to be a little more self-contained, so the book could better function as a standalone. (I had ideas for at least one more book if it took off.) Rewriting to audience response (or lack of it) was a fun experience, one you could never pull off in a novel. So, there was that. Overall, however, I was disappointed.
THE COMEBACK
But. I had yet to release the full book. Emboldened by my critical failure as a serialist, and with no momentum on the individual episodes, I decided to go all-out with the complete novel, releasing at $0.99 backed by whatever ads I could scrape together. Here was my cover:
I was in no hurry, and it took about a month to schedule everything, leap through Apple’s hoops, etc. Once it went live into the world, I discovered something funny: a lot of my readers hadn’t been interested in the serialized version, but they were plenty happy to pick up the full novel. With the individual episodes, my readers on FB and my mailing list were good for about 10 Amazon US sales in the first two days. With the full book, over an equal period, they were good for 54, and crossed 100 the day after that.
Then the ads kicked in. Which I could run, because this was a full-length novel, not a 15,000-word short. (Serializing gave me one advantage there, however: since some of my readers had already read the full thing, they were ready to review it right away. It was sort of like ARCs. That I made them pay for. Hahaha.)
With the initial push from my readers, the book became embedded in Amazon’s recommendation algorithms, which the ads helped amplify. Within a week, it had sold 575 copies there. I switched it to $2.99 a couple days after that. By the end of its first month, its Amazon US sales were about 1150, with another 150-200 on the other sites as well. Compare that to 50-60 sales of each episode over a similar timeframe.
Hooray for me! Wait, that’s not what this post is about; this post is about cold-blooded dissection. Where did I leave my scalpel?
LESSONS LEARNED
The first, and the biggest, is that serials aren’t a magic bullet. I guess that should be obvious. Nothing is! Earlier this year, however, it sort of felt like they were; at the very least, it seemed like serialization was a sure-fire way to expand your audience through the boost given to each new release.
For me, it didn’t (except maybe a little bit at Kobo). It could be the book or some part of its presentation hampered it, but whatever the cause, my episodes never gained enough momentum for the algos to take them off to the races.
Know what though, we can break this down. Here’s the main cause of my failure to launch: a) I was starting a new series my readers weren’t familiar with b) in a format they weren’t used to buying (serial rather than novel) c) with a limited fanbase to begin with (~400-500 potential readers on my lists) and d) with no outside sources to augment that potential readership; the episodes were too short to advertise in the venues I was familiar with, and I wasn’t creative enough to find alternate ways to reach people.
So basically, the only people buying the episodes were my core, core readers. The people who would buy and read the Kleenex I just sneezed into. If you’re looking at serializing purely for the benefit of multiple new releases, take a long hard look at your audience and understand that most of them aren’t going to follow your experiment right away.
Genre is part of this equation, too. Serials work better in some genres because those readers are actively searching for new content. Romance, definitely. Erotica/erom, for sure. Zombies, I think so. Time travel special ops? I.. no. No, there’s no rabid readership waiting for the next one of those to drop.
ON THE UPSIDE…
I’m talkin’ all mercenary here, but this experience was a ton of fun. Publishing a new episode every week was a blast. I would love to do that again.
Now, back to mercenary sales talk! Additionally, the format of serials provides you with many opportunities you don’t have publishing full-length novels. After the tepid response to the initial episodes, I was able to adjust my promotional tactics on the fly, permafreeing the first episode before the last was out. Not only that, but I was able to change the last episode itself based on this (lack of) response—since it looked like the season was a failure, sales-wise, I revised the ending to let the book function as more of a standalone story that would, hopefully, be more satisfying and self-contained. ‘Cause I sure as hell wasn’t gonna write a sequel to something nobody appeared to want!
There are obvious dangers with making changes like that, but being able to adjust and adapt to reader response is an incredible option to have in your back pocket.
Also, now that the full book is out there, I still have episode one free pointing to the whole thing. It doesn’t give away copies in the volume that a full book does, but it’s a nice little long-term funnel.
HOW TO DO IT BETTER
First: stick with it. My first season didn’t see any growth from episode to episode, but quite a few people wound up picking up the full novel. I think that, if I were to do more seasons, I would do a lot better. Mostly because my lists are much bigger these days. But also because I’ll have created a readership for The Cutting Room and that readership will be more used to serialization, meaning more of them would pick it up right off the bat.
Along similar lines, it would help lots to serialize something in a series/world where you’ve already got readers. Those people are already waiting for the next installment, whatever it is. That’s going to reduce a lot of their resistance to purchase a different format.
Note that I’m not saying everyone should serialize the next novel in their popular series. Just that, if you are interested in trying a serial, it’s going to help if your readers are already into the world. You could do a spinoff, say; pick up a secondary character or storyline and branch out into that in a serialized format. Now I’d better quit exploring this idea before I convince myself to do it.
Another area to explore with serials is pricing. When I released mine, I screwed up royally. Since $0.99 is the lowest you can charge for an ebook, those faithful readers who picked up The Cutting Room episode by episode paid $5.94. Then when I released the full book, I kicked it out the door at $0.99. That was due to circumstances forcing my hand, but.. that is not how you want to treat your most loyal readers, haha.
So, here’s my wonkiest idea of all: use inverted pricing. Price your episodes so buying them all will cost less than the full book. If you have 4 episodes, buying them will cost a minimum of $3.96; thus, sell your episodes at $0.99, and let your readers know that if they wait to buy the full book, well, it’s gonna cost $4.99. If you’ve got 6 eps, buying them one by one will run them $5.94, but the collection is going to be set at $7.99.
MADNESS!
Yes. Madness. A higher price will make the full-length book less appealing to readers who stumble onto it later. But that price doesn’t have to be permanent; when you get to season two, you could cut a couple bucks off the price of the complete season one. Either way, season one will still have a permafree entry point going for it. You might even package the first two episodes into a double-length pilot, the way a lot of TV shows do, and set that free to help people choose whether to plunk down for the full book. Size matters, gentlemen. If that double-length pilot is up around the 40K word range, you might have an easier time advertising it.
In any event, the point of inverted pricing isn’t to make money here and now on the full-length novels. It’s to take advantage of the perks of multiple new releases, reaching new readers episode by episode, expanding your reach each time. It’s a short-term hit for a long-term gain, Amazon-style.
GOOD LORD THIS IS AS LONG AS A KKR BLOG!
This post has largely banged on about sales, but serializing a novel was a really, really fun experience. I don’t want that to get lost in all the numbers-talk. Serializing challenged me to think about story structure in a new way, and publishing a new episode every single week was tremendously enjoyable. Despite the difficulties, I’d love to try it again some time.
It also taught me a lot about why books sell. Much of what I learned is very basic—people are more likely to buy what they already know and like, be that novel-length fiction or a world they’re already familiar with—but the fact it’s simple means it’s all that more valuable to understand.
The other very simple thing it taught me: episodes aren’t novels. Trying to sell serialized fiction is a much different world than trying to sell full-length books, complete with different advantages and different challenges. If you’re going to try a serial, I would examine those challenges ahead of time and do your best to nullify them.
Maybe that’s just a matter of sticking with it.
This is the first in a sporadic series of posts examining the publishing market, specifically as it relates to self-publishing. To start things off, I want to look at pricing, and present an ass-backwards case: that it may make more sense to price your oldest, least popular books the highest—and your new books the lowest.
Traditionally, a book is priced highest when it’s brand new. Part of this is due to the formats of a new release itself, which begin with expensive hardcovers and later move on to less expensive paperbacks, but this is frequently how ebooks are treated, too.
For instance, Stephen King’s most recent book Doctor Sleep is currently $10.99. Take a look at his backlist, however, and it’s pretty much all priced lower—mostly $5.99-7.99, with a more recent and enduring book (11/22/63) leading the way at $9.99, and running all the way down to $3.99 (The Shining, currently price-dropped to promote the new sequel Doctor Sleep).
There’s no difference in production costs for an ebook sold today and a copy of that same ebook sold on the day the title came out. So why does it cost less now than when it was new?
The answer is demand. When an author puts out a new book, there is a high level of fan demand for that title. Everyone knows you’ll be able to get it for cheaper later. But so long as the price is semi-reasonable, readers don’t care about shelling out a few extra dollars in exchange for having the specific book they want now. And publishers are happy to take advantage of that demand by setting initial prices higher.
Why reduce price of that same ebook format later? I don’t get invited to sit around a lot of Big Publishing House marketing meetings, but the probable answer is obvious. As demand wanes, you cut price, hoping to lure in new readers. Especially the segments of the market that are more price-sensitive. Over time, as a book steps down the pricing staircase, it drops below the purchase threshold of several different markets, garnering new purchases at every step of the way. Eventually, once the higher-paying markets are exhausted and demand settles down to a trickle, you slot the book into a low price—one that minimizes readers’ resistance to purchase, but isn’t so low that it devalues the author and/or the publisher’s entire catalogue.
That’s the traditional model. It makes a lot of sense. Maximize initial revenues by taking advantage of pent-up demand, then lower prices to draw new readers into the backlist (and, hopefully, convert them into fans who’ll then go on to buy frontlist).
The self-publishing/indie approach is way more fragmented, with different people trying all kinds of different things, but I think the general approach is similar: release a new book at full price, or at a slight discount, let it fade into backlist, then run sales on backlist titles to goose new releases and/or the rest of the backlist.
This makes sense, too. I don’t think it’s a bad approach in the slightest. But I think there is a stark difference between the pre-digital publishing market and the current market. Back in the day, you counted on audience growth through word-of-mouth, right? People still treat word-of-mouth like your #1 weapon for growth from book to book.
Well, I think that’s been replaced. Overshadowed, at least. By visibility.
Yay, a buzzword! That’s how you know this must be smart. To explain this particular use of “visibility,” it’s time to dissect Amazon’s recommendation engine.
When a new book is released on Amazon—and the other stores, too, but Amazon is the one I’m most familiar with, and the one I believe does this best—its resulting sales aren’t just about the pent-up demand waiting for it. Instead, Amazon’s bots process its initial sales, then actively promotes that title to other customers likely to purchase it.
This is largely black box stuff accomplished through emails and on-site recommendations, so it’s hard to capture hard data on these processes. But I’m guessing Amazon’s system is a lot like Netflix, where people who watched/enjoyed Movie A, B, and C will be recommended Movie D, which other people who enjoyed A-C also watched and liked. Amazon has their own version of this front and center, the “alsobot” recommendations on every single product page.
So a new book comes out. Its fans buy it. And then the mighty Algorithm kicks in, recommending the book to potential customers. The people it recommends the book to are (probably) determined by taste-constellations, i.e. the people Amazon’s system thinks are most likely to buy it. And the volume of those recommendations is determined by the volume of outside sales—anything not initially generated by Amazon itself.
The recommendation algorithms are highly adaptive. Both the targeting and the volume of that targeting are influenced by the book’s performance when Amazon puts it in front of potential customers. The better a book converts potential customers to actual purchases, the more the algos ramp up the juice. The worse it does, the more the juice dries up. (For the record, I make no judgments here about a book’s literary quality—just its commercial potential, as predicted and then tested by Amazon.)
For a brief foray into concrete examples, last month, I released a new book. I was able to push about 250 sales through my own devices. Over the course of the month, it saw “organic” algo-driven growth, and finished September with about 1950 sales. Thanks to Amazon’s marketing, it sold 8x as many copies in its first month than I conjured up on my own.
For anyone who spends any time on Amazon, this general process is obvious. But look at what’s happening here: Amazon’s sophisticated recommendation system is identifying potential customers and then pushing something they’re likely to buy right in front of them.
In other words, they’ve slashed a huge shortcut through word-of-mouth. They have replicated an organic process that used to take weeks, months, or years to really kick in, and condensed it to a matter of days.
I’ve got one bit of data to help approach this, though. In my observation and experience—anecdote alert!—given equal visibility, a book priced at $0.99 will sell roughly 2x as many copies as one priced at $2.99. This is in SF/F; in particularly price-sensitive genres, like New Adult and possibly Romance, the difference may be higher, maybe as high as 3x.
Working from a 2x figure, though, let’s crunch some numbers. Say you have the option of 1000 sales in your first month at $2.99 or 2000 sales at $0.99. $2.99 earns 70% royalties—$2100, in this hypothetical. $0.99 earns 35%—$700. That is a huge difference, one that I have no doubt Amazon intentionally laid out to create a soft floor for ebook prices.
But let’s say you can convert 1 out of every 20 purchases into core fans, people who’ll happily sign up for your mailing list/Facebook page/blog so you can reach them directly when the next book is out. At $2.99, you’ve added 50 core readers. At $0.99, you’ve added 100. Your next book will have twice as many purchases to impress the recommendation algos with than if you’d launched the previous book at $2.99. Multiply by as many books as you intend to write in the series.
Caveats here, of course. A few of the people who signed up for $0.99 books may not buy a sequel at $2.99. And conversion rate naturally degrades the deeper you get into a series. And a 20:1 purchase:signup ratio is generally too high (though it doesn’t really matter what the absolute ratio is, it’s more about the relative ratio of sales between $2.99 and $0.99).
So you can downplay this idea a little bit. Even so, the long-term implications of launching cheap are pretty interesting. And obviously there are more ways to look at this besides $0.99 vs. $2.99. Such as this:
This is from Smashwords’ yearly survey on self-publishing trends. The first thing you’ll note is that it directly contradicts what I claimed above about a title selling twice as much at $0.99 vs. when it’s priced at $2.99. So it’s kind of hilarious that I’m going to use other parts of the data to make a point. But whatever, apply the salt grains as necessary. This Smashwords survey isn’t an apples-to-apples comparison. It’s looking at sales across its entire catalogue, not just at the difference in sales volume when a specific book with constant, equal visibility is switched between different prices.
Anyway, that’s outside what I want to look at here, which is that big ol’ hump in unit sales for books priced $2.99-3.99, the slope between $4.99-6.99, and that floor at $7.99+.
If this data is trustworthy, you can use it to explore the various mini-markets of the overall ebook market. For many people, up to $3.99 is a bargain, an impulse purchase. Another smaller segment doesn’t think much of paying up to $6.99 for an ebook. Move to $7.99+, and you’re at frontlist prices. There’s no bargain at all.
Thus I think there’s a pretty crazy case for punting new books out the door at $3.99, max. Even if you’re an indie and you don’t charge much to begin with—few of us go higher than $4.99—just a small initial discount could pay off.
Whatever the case, applying traditional frontlist prices to new releases maximizes earnings from preexisting fans, but drastically reduces your ability to create new fans. Not just over the long-term, but during the short-term recommendation-driven visibility Amazon will grant you as a new release.
That by itself is a strong argument for inverting the traditional pricing structure. On top of that, you’re actually rewarding your existing fans for their loyalty by charging them less than the book’s eventual list price. Treat your readers well, and they’re more likely to stick around for the future. Yay, a lasting career.
I know there are a lot of good arguments against this general premise. Particularly in series, where you can put out new releases at whatever price you want and use the visibility of the new release to push potential readers toward the (bargain-priced!) beginning of the series. Even so, I think it’s worth thinking about. Particularly for standalones and the beginning of a new series, launching at a lower price than your “list” might be the best move.
A couple days ago, I rejoined the Self-Publishing Podcast once again. This time, we’re talking launch strategies, box sets, sales trends, and how to give a big boost to later releases in your series, among other things. This is either the third or fourth time I’ve been on the show—apparently I’m now the Justin Timberlake to the SPP’s SNL—but I had a particularly great time on this episode, and feel we covered an awful lot of ground.
9/23 Update: Someone contacted Fiverr about a breach of privacy regarding the “quotes” featured in the accusation (which were the only real piece of “evidence” in the entire report). Fiverr confirmed the accuser never worked for their company.
I’m loath to give it any traffic, but you may have heard of a so-called scandal making the rounds: the “Fiverr Report,” which claims to blow the lid off a number of bestselling indie authors who’ve bought fake reviews.
Let’s just get this out of the way: it’s total nonsense.
It’s easy to believe a lot of the big indies got where they are through fraud. Hell, early success John Locke admitted to buying reviews, and if you followed the rabbit trail on his fakes, it led to a lot of other self-published authors. Of course, few of the authors indicated in that scandal were big-time or even midlist indies. Most of the people who appeared to have used the same service as Locke fell into a far different category of author: people who had published one or two or three books, typically in paperback, and had bought a handful of reviews (1-10) to try to increase their nonexistent sales.
This was the pattern because authors who are already selling tons of books don’t need to purchase reviews. The authors fingered in this “report” sell thousands—if not tens of thousands—of books every single month. When you’re moving that many books, your reviews arrive in a steady flow. (For reference, I find I receive about 1 review for every 100 sales. Reviews come in faster when a book has few or none, and start to slow down once you’ve got a whole bunch.)
Sure, these authors could have bought their reviews at the beginning, using them to launch themselves into outer space; in that case, the fake reviews would since be buried/diluted by the real ones that came with all those new sales. But there are at least a couple problems here.
First, I am a crazy author-stalker. I look at dozens of Kindle titles a day. I watch the bestseller lists, the popularity lists, and I keep up with the books and new releases of countless sci-fi and fantasy authors I find interesting. I’m pretty familiar with the careers of at least eight of the authors on this list, and while I can’t say with 100% certainty that none of them has ever bought a review, I know at least some of them haven’t. I’ve watched one of these authors from the very beginning of his career. Literally days after he first published. A flood of fake reviews simply never showed up. If some of the accusations are false, it casts doubt on the whole thing.
Second, the report claims these authors have bought at least 500+ reviews. Several of these authors hardly have 500 reviews across their titles. The math just doesn’t add up. Not given how many copies they’ve sold.
Last, the (probable) author of the report has been outed on KBoards. He isn’t mentioned by name, and I don’t know the evidence to support the accusation, but the author in question is infamous in certain circles. He actually has left hundreds of fake reviews on his books. They’re not there anymore—because Amazon stripped them all, and actually made it impossible for people to leave new reviews on the books—but let’s just say it’s plausible.
In summary, these people wouldn’t need to leave fake reviews, because they’re already tremendously successful. It’s hypothetically possible they could have bought batches of reviews early on when they weren’t selling well, but I know for a fact that isn’t true of some of them. Meanwhile, there’s indications the accusations are being made by another indie with an axe to grind.
No actual evidence is presented. Instead, the report presents several “quotes,” without sources, and then seems to have picked a bunch of popular authors with highly-rated books. And the averages on some of those books could look kind of funny. How is it that a book like the Wool Omnibus can have a 4.7 rating on 6235 reviews while a mega-popular author like Stephen King only has a 4.0 rating across 3382 reviews on Under the Dome?
Part of the answer is the Sequel Effect. The Wool Omnibus may have a 4.7 average, but the main lead-in to the series—Wool the short story—only has a 4.4 average. The difference may seem slight, but a 4.4 is actually far more achievable than a 4.7. My novel Breakers has a 4.4 average on Amazon, for instance. What’s happening here, then, is a lot of people are picking up Wool, but only the ones who like it are then moving on to the Omnibus, whereas people who didn’t like Wool, and are more likely to give it a poor rating, never move any further into the series. And later books are buffered even more. This is why a book like A Modern Witch, first in the series, has a 4.3 average, while the last book in the series has an incredible 4.9 on nearly 500 reviews.
The Sequel Effect isn’t the only factor in why indie authors can pull great ratings compared to big-timers like Stephen King. This is more speculative, but I think self-published authors are often far more engaged with their fans. We speak directly to our readers on Facebook, through email, on forums. This leads to a more active, enthusiastic fanbase, one that may be more likely to have positive feelings about the authors’ books. There’s nothing sleazy about this. It’s just a natural outgrowth of authors who are touched about the fact there are actually people out there—lots and lots of them!—who want to read our silly books.
I think the pace of indie publication may have something to do with it, too. I think there’s an unquantifiable amount of goodwill generated when you put out a new book in your readers’ favorite series every 1-4 months. Thought experiment: Currently, A Feast for Crows has a 3.5 average on about 2500 reviews. A Dance with Dragons sits at 3.6 and 4500.
But many of the reviews flat-out state they’re frustrated with having to wait so long between books. AFFC was published in 2005, five years after the previous novel. ADWD came out six years after that. How much long-simmering resentment toward these books would there be if they’d been published every 5-6 months instead of every 5-6 years? I think a great many readers would be far more forgiving of George R.R. Martin flying off track and spinning his wheels if they knew there would at least be a new sequel later in the year.
Anyway, much like A Song of Ice and Fire, this post has gotten far off-course. The fact is that, while I don’t know with 100% certainty that none of the accused authors are guilty of having bought reviews at some point in the past, the report is clearly a sham. If any of the accused actually has bought reviews, their presence on this list is purely coincidental. It is best to assume that none of the named authors has done anything wrong.
It’s easy to believe accusations like this. Self-published authors have bought reviews in the past. Heck, somebody’s probably buying some right now. Meanwhile, indie book ratings can defy belief.
But a successful indie has less motivation to buy reviews than the author you’ve never heard of. And there are multiple reasons that self-published books wind up with better ratings than traditionally published beloved bestsellers. In the case of accusations like this, you can’t go on the basis that a book’s rating just “feels” wrong. You need evidence. Documentation. Screenshots. A trail of behavior.
This report has none of that. It’s a few undocumented quotes followed by a bunch of completely unsupported claims. It’s not worth the time of day.
Now let’s see how long it takes before Salon runs a report on the rampant cheating in the indie world.
