That's not the point. You suggested they were engaged in tax avoidance by underreporting their sales. I was telling you that's not really a thing.
That 2.2m figure is for every e-book on Amazon, not just self-published ones. So while I imagine there were titles that didn't sell a single unit, we can't say that at least 75% of self-published ones didn't shift any. The minimum proportion will be lower.
I'm not great at maths but are you sure? I mean, all 448k sales could be for three titles for all we know, in which case 99.9% of self-published books didn't sell anything. Obviously it won't be that extreme but the point is, we can't make any meaningful deductions can we?
The 75% came, I believe, from assuming that each of those 448k self-pubbed sales was from 448k books making one sale each, out of a total of 2.2 million. Obviously, that's not realistic - but if it was true, it would mean about 1.75 million of those eBooks didn't sell a single copy, because the sales came from the others. 1.75m out of 2.2m is about 75%. But the 2.2 million is all eBooks, not just self-pubbed ones. Therefore there's less than 2.2 million self-pubbed eBooks out there, so those 448k self-pubbed sales are spread over a lower total. This means the minimum possible proportion of self-pubbed books that didn't shift any copies is lower than 75%. Which is a long way of saying: yes, I'm sure. I think you took 'minimum proportion' to mean minimum proportion of books that DID shift some units, rather than the other way round. All that said, no, you can't really make any meaningful deductions from those figures because they're representing two different things - one's just relating to self-pubbed books, and one isn't.
Unrelated question. On the KDP Reports page, there is a graph showing Kindle Edition Normalized Pages (KENP) Read, and I have a spike of 1170 right after E&D hit the stands. What is this? Clicking "what is this" on the page didn't tell me anything. Spearate from the downloads graph
Hi, KENP - Kindle Edition Normalised Pages. Yes that's the number of pages (as defined by the kindle program not your word file) that were read. Most of my books run between 6 or 7 hundred pages and a thousand. But say it's a thousand and a thousand pages were read this may mean a thousand books were picked up and one page was read before the book was discarded, or one book was read completely or anything in between. Now Kindle Unlimited has been a mixed blessing to authors. Some like me find it wonderful as it greatly increases our income. Some don't. Those who don't usually do what they call "going wide" meaning they put their book out on as many different publishing sites as they can. This means all their sales are not purely Amazon which underestimates the figures given. Some have reported their non-Amazon sales are much larger than their Amazon sales. Meanwhile for me using KU, I don't have this option since its a condition of KU that I don't publish with any others. But having said that I normally find that KU income doubles my income. Ie I get as much moola from it as I do from sales. Next to throw into the mix - a great number of authors write crap (yes I said it and I'm an indie - mostly) and their books having piss poor covers, bad blurbs and being badly written and edited never sell a single copy. Thisis why self publishing gets such a bad rap. Having said that many more authors are writing books that they know won't sell - like family genealogy etc. What that means is that these 'overall" sorts of figures don't really mean a lot. If most authors are writing crap and you're being professional, than you have a much better chances of making money from your work through sales. Last, and this is important, you don't write just one book. To be successful unless you're extremely lucky, you have to keep writing, putting out book after book. Now say you sell ten copies a month of your book, and you have one book out which you sell at $2.99 and get roughly two bucks back in royalties. You've made twenty bucks. But if you have ten books out - that's $200. In short you can make a successful living as an indie author even with modest sales, simply by being productive and professional. Cheers, Greg.
The post from Sack-a-doo had the sales for January by type (re-linked for convenience). Maybe a quarter way down the page, there's an underlined segment called "Amazon's daily ebook unit sales" where the total of the "indie published" (including kindle unlimited and regular retail) adds to 448k units. Exactly, and you're right to point out the flaw in my logic. The only assumption we can make is (looking at the total sales of 1,064k units in January) that at least 52% of all e-books, including both trad-pub and self-pub titles did not make a single sale in January. (1-1064k/2200k) As you say, based on the absurd but statistically possible, if improbable premise that each sale was of a unique title. Someone in my book needs to say "statistically possible, if improbable..." preferably while adjusting his glasses...
Thanks, Greg! @psychotick ! I found that report from @Sack-a-Doo interesting, but I wish the author had defined the terms better. I read that from beginning to end but they were not. Big Five and Small to Medium publisher is understood, though not defined. but the remaining three are not: Indie, Amazon and uncategorized. Is an Indie an independent publisher or an independent author? If independent publisher, what is the difference between an indie publisher and a small to medium one? Is Amazon a book that is only available on Amazon? And what is uncategorized author? And what, by these undefined terms, am I? My book was published on Create Space and only available on Amazon in paper or Kindle.
I have no idea what you were thinking, I was only going by the words you typed, which were: "And considering ego and tax avoidance, some will lie and say they do while others will lie and say they don't." You SAID tax avoidance.
Good questions all. did you find the excel at the bottom? Clicky Linky! I did some investigating and: - realised the figures I was talking before are daily unit sales, not monthly. 1m books a day, 30m per month. (more hope for us, yay!) The Big 5 seem to be Hachette, Harper Collins, Macmillan, Simon & Schuster and Penguin. Unclear how the rest were defined, it's all been anonomysed. It doesn't answer your questions, actually if anything it raises more, but: - they've used some fancy pants math to backwards-forecast the unit sales based on the rankings. - 32 titles sold more than 1000 units per day, - 893 titles sold more than 100 units per day (which is about the limit you described for making a living) - 2007 titles sold more than 50 units per day - 14k titles sold more than 5 units a day - 56k titles sold more than 1 unit per day. It also seems like the number of sales per day isn't based on actual sales, but on an extrapolation based on total sales units, and the title's position in the ranked book sales.
I don't understand math enough to understand the criticisms, but some of you guys seem more involved, so let me link to http://www.slhuang.com/blog/blog/2014/02/16/links-to-analysis-regarding-authorearnings-com/, which is a critique of the Author Earnings conclusions coming from a mathematician (and self-published author) with lots of links to other criticisms. The post is a couple years old now, so it was of an earlier version of Author Earnings - not sure if the criticisms still apply or not. Maybe you guys can sort through it?
I only read through one linked article in BayView's link (the topmost). I am not a proper statistician though I have learned the basics and written a few articles in scientific journals of my own, so here goes: From the main critique points alone: - (1) This is Amazon-only data. - (2) This is one day’s worth of data, one 24-hour period of sales and rankings. I say that the report as stated back then cannot be used as an overall trend and any and all conclusions which have been drawn from this data are invalid. Take the 24-hour period alone: one can't extrapolate from 24 hours to the lifecycle of a book. Sales within 24 hours are dependent on where the book is in its own respective lifecycle.
As far as I can tell, in this report they've fixed some of the big things that were criticised in that article - the model for calculating author earnings is different and I can't find any batshit insane inferences for author income based on a single day's sales in the article. What gives me most pause about the data is the formula they've used to turn the sales rank into the number of units sold: 10^(3.64959191*log(sales rank)-0.07791303*log(sales rank)^2)/2 Mostly because I have no idea where it came from. It might be accurate, but with the data we've got there's not really any way to tell, and pretty much everything money-related we can try to infer rests on it. They've also used blanket figures for what publishers pay their authors (25%) and assume every book has the average number of page-reads for Kindle Unlimited money. Not unreasonable assumptions given what they're trying to do, but worth bearing in mind if you're trying to draw any conclusions from it. On the points @Lifeline brought up, the data is still Amazon-only, and it is still just from one 24-hour period, so while you could make some judgements about trends based on this data and the equivalent set from last year, it'd be a bit dodgy. Anyhow, I find it hard to argue with any of the conclusions in that report because they don't really make any. It's more just 'based on our data of people spending money on eBooks, people are totes spending money on eBooks. Self-publishing's fucking amazing, you should do it.' Even that's a bit of a leap - people may be spending more money on self-published eBooks than ever before, but the number of eBooks available has gone up as well, so it doesn't mean the average author earnings are higher.
As an engineer, I would say that is an empirical curve-fitting formula to turn scattered noisy data into a single trendline, derived from a the data, notnything analytical. Finnegan's finagling factor: the number you add to the data you got, to get the answer you want.
Yeah, it was more where that data came from. They're using that formula to turn the Amazon sales rank into an estimate of books sold, presumably with the constants being some function of the total books sold on that day - but if I was using an equation like that, it'd be to find the relationship between two existing variables. That is - if I had a data set of sales rank to books sold on one day, I might use it to get a regression line like that to predict the books sold from the sales rank on a different day. But if you can get hold of that data to begin with, I'm not sure why you'd need the model for these purposes. Just take your snapshot as the period you've got hard numbers for. Possibly I'm missing something - it has been some years since I had to use this stuff in anger - but the model they're using feels a bit lacking in foundations.
I think the rule is that if you are in the 100,000 ranking you are selling about 1 per day, 500,000 a couple a month, and higher than that just once in a while. And I know precisely how many I am selling from my CreateSpace dashboard and KDP reports page, though I have some problems with that page
Yeah, the authors themselves should have no problem knowing how many copies they're selling on Amazon, but it's pretty hard for anyone else to get hard numbers. Another issue would be trying to figure out how many books are attributed to each author. I probably have 30 books on Amazon, but they're spread between different pseudonyms, etc. They're also at different prices. In terms of being able to make a living at something... If I have 30 books at $4.99 selling three copies a day, I'm not making great money, but I'd be above the poverty line. But if each of those books was at $.99 and each book was from a different author, it'd be a whole different story. Also important to remember that we tend to be looking only at Amazon.com, as I understand it. I'd say I get about 50% of my Amazon sales from .com, with the other 50% divided up amongst the other domains (.ca, .de, .uk, etc.) And while most of my sales (probably 70-80%?) come from the Amazon family, I've heard of other authors who get less than 50% from Amazon. Way too many variables to come up with anything more than vague impressions, I'd say.
Hi Guys, Just to comment I can tell you that the formula given to convert sales rank to sales is innacurate. It has two major problems. The first is that the Amazon sales rank also includes the pages read (KENP) of KU. (Another eason why adding your book into KU is beneficial since it'll rise higher in sales rank hopefully if it is, and the higher the sales rank the more likely it is that Amazon's own algorythm will give it a promotional boost.) So any sales rank is comprised of both pages read and sales and the formula doesn't seem to allow for that. Second and more troublesome, the Amazon sales rank is also based on the price of the book sold. If you sell a book for $9.99 your sales rank rises more than if you sell that book for 0.99c. Again the formula doesn't seem to allow for that either. Cheers, Greg.
Do you have a firm source for the price-related part of that? I've read other sources saying that the rank is based solely on KENP plus sales numbers. If there's proof of the opposite, I'd love to see it! There are always new ways to game the Amazon system, of course. Isn't there something about a transition period between the "free" rank and the regular rank? Like, if you have your book for free and it gets ranked in the top 100 of the free list, what rank does the book have when it goes back to the "paid" rankings? And how are the KENP rankings included? If someone reads half the book, does it count as half a sale? Or is it every download that counts as a whole sale? I've noticed that since the latest incarnation of KU started, the Amazon bestseller lists are DOMINATED by KU titles, suggesting that the KENP numbers have a really significant impact... or that free days have an impact? The formula I heard for Amazon rankings was something about taking the previous day's sale, or rank, or something, and giving it half-value, and then adding the current sales on top of it... or something like that? The conclusion was that steady sales rather than a single-day spike will have the biggest impact on the sales rank. You can make yourself crazy trying to keep track of it all. I think some people enjoy the puzzle, but I just find it a distraction, mostly.
