Download the presentation (.pdf)


























































90 Responses to “2016 Romance Writers of America RWA PAN Presentation”

  1. Fantastic information, as usual, Data Guy! Thanks to you and Hugh for putting these numbers together and doing the crunching.

    I write a bit of romance under a couple of pen names, and this whole data set is an excellent resource for making decisions about the options available to indies.

  2. Suzan Harden says:

    Thanks for post your RWA slides, Data Guy! They show how much the market has shifted over the last five years.

  3. Samara says:

    I’d love to know the following data, which would help put things into different context:

    How many Big 5 authors & titles are there total?
    How many indie published authors & titledare there total?
    How many Amazon Imprint authors & titles total?

    HEAPS more insights into this data with those figures. Any idea how to get hold of them?

    • Data Guy says:

      Hi, Samara,

      Here they are… πŸ™‚

                      titles      authors     
      indie self-published 72,410 14,622
      small/medium publisher    43,856 9,227
      amazon imprint 2,040 752
      big five published 39,191 5,189
      uncategorized 33,780 9,540

      (To be clear, that’s the breakdown of Kindle Romance titles and authors in the data.)

      • Samara says:

        Wow! Thanks so much for that information. I really appreciate it.

        • Samara says:

          So I worked out (very roughly given I was reading off graphs and don’t have exact numbers) that approximately 7% of Big 5 authors make $10,000 per year or greater, compared to 11% of indie authors. That data surprises me, although I wonder how that would change given that hardcopy authors would primarily be Big 5/medium publishers.

          • Data Guy says:

            That’s an insightful way to break it down, Samara. Thanks! πŸ™‚

            I suspect the ratios would change a little when print sales are factored in, but not by a whole lot, given the small % of overall US Romance sales that are print editions (and most of those, inexpensive, lower-royalty-rate Mass Market Paperbacks).


  4. Really excellent, detailed data analysis of the romance market, and thanks for posting the slides here. Clearly, I should be writing a shifter/romantic suspense series with 11+ books and make the first one permafree. πŸ˜‰

  5. Diane Capri says:

    Fascinating info! Thanks for sharing. Any chance we could get similar info for Mystery/Thriller/Suspense (not romance) genres?

    • Data Guy says:

      Hi, Diane,

      I write Mystery/Thriller/Suspense myself… πŸ™‚

      For print Adult Mystery/Detective Fiction and Thriller/Suspense Fiction, the most reliable industry data on print sales tabulates a total of roughly 33.5 million hardcover, trade paperback, & MMPB units sold in the US in 2015. (That’s for adult fiction; for children’s books, they don’t break out M/T/S separately as a category). We can add maybe 5-6 million more print units for library sales, author-direct and publisher-direct sales, and other print outlets not covered by Bookscan. So rounding up, we can call it approximately 40 million US Mystery, Thriller, & Suspense print sales in 2015. (That makes M/T/S roughly 24% of Adult print fiction sales in the US, and 5.3% of *all* US trade print sales including trade nonfiction).

      For ebook Adult Mystery, Thriller & Suspense, our AE data shows a run rate of about 83 million sales a year on, making M/T/S roughly 21% of all US Kindle sales. 46 million of those sales are of M/T/S titles not in Kindle Unlimited, while 37 million are a mix of direct retail sales + KU full-read pagecount equivalents for M/T/S titles that are in KU. To that we can add another 29 million units or so at iBooks, Kobo, Nook, and Google, for a total of approximately 112 million US Mystery, Thriller, & Suspense ebook sales in 2015.

      So Mystery, Thriller & Suspense is a little less digitally-skewed than Romance… about 73% of US Mystery, Thriller & Suspense unit sales are ebooks, while 27% are print.

      When we hear traditional-industry e-versus-p ratios quoted, what skews those ratios so heavily toward print is the inclusion of nonfiction and children’s books. Nonfiction makes up more than half of all US print sales, but only a small percentage of US ebook sales even for traditional publishers. And of the remaining less-than-half of trade print sales that are fiction, more than half of *those* sales are books categorized as children’s books. And of course the industry’s inability to track and account for indie sales, which are overwhelmingly ebook fiction, skew the traditional stats even further toward print. But for adult fiction — particularly adult genre fiction — the majority of sales are digital today.

      Data Guy

      • Diane Capri says:

        Thanks for this, Data Guy. Does your data show income and revenue to indie/trad/hybrid authors in our genres similar to what you did here for romance?

        • Data Guy says:

          Hi, Diane,

          Here’s a breakdown by publisher type of Kindle Mystery/Thriller/Suspense sales β€” which make up 74% or so of all US Mystery/Thriller/Suspense ebook sales:

          Daily Unit Sales:
          β€” 32% Indie (Self-Published)
          β€” 12% Small/Medium Publisher
          β€” 17% Amazon Imprint (Thomas & Mercer, mostly)
          β€” 33% Big Five
          β€” 6% Uncategorized

          Daily Author $ Earnings:
          β€” 39% Indie (Self-Published)
          β€” 8% Small/Medium Publisher
          β€” 16% Amazon Imprint (Thomas & Mercer, mostly)
          β€” 34% Big Five
          β€” 3% Uncategorized

          Daily Gross Consumer $ Sales:
          β€” 18% Indie (Self-Published)
          β€” 10% Small/Medium Publisher
          β€” 13% Amazon Imprint (Thomas & Mercer, mostly)
          β€” 56% Big Five
          β€” 3% Uncategorized

          Hybrid is tricky — in the above numbers, the sales of hybrid authors are split between their different publisher categories, and counted separately in each. One of these days, I’m planning to do a deeper dive into they stats on hybrid.

          I’d also love to dig deeper breaking down the M/T/S data and put together a presentation similar to what I did for Romance at RWA 2016. RWA is a really diverse and progressive author organization, but ITW is pretty progressive, too. Maybe ITW would be interested in hosting a similar session at the next Thrillerfest… who knows? πŸ™‚

          Data Guy

          • Diane Capri says:

            Thanks for this. Really appreciate it. ITW might be interested in a presentation like that. I’m sure NINC would be.

          • I’m pretty sure they would! Drop me an email if you’d like me to follow up with them.

          • Cecilia Tan says:

            I’d love to see this kind of data crunching brought to the science fiction/fantasy authors community. Now that SFWA is admitting indie authors I think they’d be keenly interested to keep up with these developments. Maybe for the Nebulas Conference next year?

      • michael gorga says:

        Data Guy,

        This data is phenomenal! Thank you for sharing such a wealth of knowledge! Do you have any insight into the paranormal non-fiction market?

        Thanks again.

  6. Misha Hixon says:

    Thanks @Data Guy and @Hugh Howey for taking such an endeavor to bring among such a huge amount of transparency in concern with the Romance genre. Truly appreciated! Now, we can get a clear idea which the Big 5 had and self-published authors didn’t.

  7. Elizabeth Jennings says:

    Hi Misha–You know, I don’t think that the Big 5 have these numbers and even if they did, I don’t think they have the inhouse ability to crunch them like Data Guy does. They are surprisingly heavy on ‘managers’ but are understaffed in the editorial/marketing departments and don’t have the analytical ability to understand their own sales trends.

    That’s what I’ve observed, anyway.

  8. Jeff says:

    Hey DG, I know you are busy, but could we get a brief breakdown of Sci-fi? This is incredibly amazing and I am just stunned by the new info! Thank you for your hard work!

    • Data Guy says:

      Hi, Jeff,

      Total US print SF&F books (hardcover & paperback) sold each year, according to the most reliable source of industry data on print sales, add up to roughly 47 million units:
      – 34 million of that is categorized as Children’s Science Fiction/Fantasy/Magic;
      – 13 million of it is categorized as Adult Science Fiction or Adult Fantasy.
      Add the 15% or so non-Bookscan sales, and you get to 54 million SF&F print sales, mostly children’s titles.

      When it comes to ebooks, trad SF&F sales on are running at around 23M units for Kindle. Add in iBooks, Kobo, Nook, etc. and we’re talking about 33M or so total SF&F ebooks for trad.

      So that’s 87M US trad SF&F sales annually for adult+children’s print+ebook combined.

      OTOH, indie SF&F ebook sales on are at 32M units for Kindle, which, when you add in iBooks, Kobo, Nook, etc. will total out somewhere between 38M and 40M. Almost all of that is adult SF&F, not children’s (although Teen/YA makes the distinction murky).

      Which means that more than half of all SF&F ebooks sold in the US β€” and nearly a third of all SF&F books of any format sold in the US, including children’s books β€” are indie.

      Here are the sales breakdowns by publisher type for Kindle Science Fiction & Fantasy (which makes up roughly three quarters of the SF&F ebooks sold in the US):

      Daily Unit Sales:
      β€” 47% Indie (Self-Published)
      β€” 14% Small/Medium Publisher
      β€” 10% Amazon Imprint (Thomas & Mercer, mostly)
      β€” 20% Big Five
      β€” 9% Uncategorized

      Daily Author $ Earnings:
      β€” 54% Indie (Self-Published)
      β€” 13% Small/Medium Publisher
      β€” 8% Amazon Imprint
      β€” 18% Big Five
      β€” 7% Uncategorized

      Daily Gross Consumer $ Sales:
      β€” 29% Indie (Self-Published)
      β€” 18% Small/Medium Publisher
      β€” 8% Amazon Imprint
      β€” 34% Big Five
      β€” 11% Uncategorized

      What’s really remarkable here is that if even half of the “Uncategorized” are actually indies (and we believe most of them are), then that means US readers are already spending more of their bookbuying dollars on indie SF&F ebooks than on those published by the Big Five.


  9. Sela says:

    Hey DataGuy! Wonderful report. So useful for all of us indies in romance and those considering whether to go into it. Clearly, romance rules. πŸ™‚

    You know me and you know where I fit in your author earnings report. πŸ˜‰

    From your data, it looks like romance authors do better on Amazon when they are in KU. What about romance earnings on other retailers? Being in KU means exclusivity and thus no earnings on other venues. For example, I’ve had a few months where my income on Amazon has fallen quite low (by my standards) but my income on other retailers brings me up to a still 6-figure revenues for the year. When I had my recent free Bookbub, my revenues actually surpassed Amazon revenues. While Amazon is UUUUGE, there is a lot of money to make on other non-Amazon retailers for romance authors.

    To really understand whether KU is or is not better for indie romance authors, it would be great to track how much non-KU titles earn on non-Amazon retailers. I don’t know if you can do this with your methods. Amazon definitely dominates in the US market and is dominating in UK but there are still areas where iBooks and Kobo perform well. What about romance sales in non-US and non-Amazon markets? Would AE consider doing an analysis that looks at more than Amazon?

    I’m one of those authors with some books in and some books out of KU and I often struggle with whether I should go all in, hybrid or all out. It’s hard to know what to do and your data doesn’t make it any easier for me! πŸ™‚

    Keep up the great work!

    • Data Guy says:

      Hi, Sela,

      Author Earnings looked at ebook sales at the Apple iBookstore, B&N Nook, Kobo US & Google in our October 2015 report.

      Keeping 5 different custom-coded, site-specific spiders up to date was exhausting… πŸ˜›
      It required easily 5X the work of just spidering Amazon. And since the sales numbers for all 4 other sites put together added up to less than a third of Amazon’s alone, it was a pretty clear case of 80-20 rule and diminishing returns. Instead, I put the energy into revisiting our rank-to-sales conversion process, making it even more accurate.

      I do plan to eventually tune up the spiders for the other ebook retailers and revisit them, but it’ll probably be late in 2016 or even 2017 before I feel like tackling that. πŸ™‚

      Data Guy

      • It would be great – and I’m sure a ton more work – to see numbers that include the other retailers, because the data on Romance authors not in KU means that they’re elsewhere (since KU is exclusive) and making money and sales somewhere else. I’ve seen the numbers on the other retail sites increasing fairly significantly in the past 12 months. Hope at some point those numbers can be included in this data. I think it’s important to look at the overall market whenever possible when authors make conclusions about about how/where to sell. But really appreciate all your data/slides! So nice to see all that unreported income posted somewhere!

        • Awesome idea, Barbara! Amazon is only about 60% of my overall sales (with a 6-figure gross), so those are some significant numbers.

          But thanks for what you do have, Data Guy. I write long-ass (500-700 page) books and take forever doing it (one book a year in 2012-2016 so far) and am now going to think about whether I need to charge for the first in series that has been my permafree since June 2012. I’d still price it low ($.99-$2.99 for a 175k word set of two books), but I’m only giving away about 40 a day at Amazon nowadays, so it’s at least worth experimenting with. Maybe just one more BookBub freebie before I raise the price.


        • Data Guy says:

          Hi, Barbara,

          Exciting to hear you’ve seen an uptick at other retailers — let’s hope it becomes a trend. Back when I first started out, I was earning more money from Nook than Amazon; it’s been painful for me to watch Nook’s $60-million year-on-year annual declines, year after year… πŸ™

          I definitely intend for AuthorEarnings to revisit and report on sales at the other retailers; it’s just a matter of when. πŸ™‚


  10. Thanks for posting the slideshow. BTW, your live presentation was great–dynamic and engaging. Glad I had to opportunity to attend the session.

    • Data Guy says:

      Hi, Selene,

      Thanks for attending! It was really the audience that made things dynamic and engaging — y’all are the smartest, most entrepreneurial players in the industry and you know it. I loved all the great questions that came up; having the full hour worked out well, because it let us interleave the Q&A in real time. πŸ™‚

  11. Jackie Weger says:

    Data Guy! Just a stunning presentation. Using the figures you showed I crunched the numbers/sales/$$$ of my book list and found the averages are almost dead on for my romance ebooks not in promotion on May 5. I network with a small group of indie authors and track our books in KND ebook tracker. None of us are household names. Several publish wide. I discovered this: With few exceptions, titles in Amazon Select outperform titles outside of Amazon exclusivity, EXCEPT when in high-profile promotions i.e. Bookbub. One of the [wide] authors reports Kobo often out performs B&N. Your presentation is a plateful of food for thought.Nicely done.

  12. John Brown says:

    Wow. Fantastic data. Thank you so very much!

  13. Thanks–some great information here to think about. I see a lot of questions on forums about novel length so I found that slide of particular interest. I tend to write in the 70k word count range (although my work is mystery, and urban fantasy there is always a romantic subplot.)

    Appreciate the work that went into this. Thank you.

  14. ChloeG says:

    So… I’m boggling at this one that I just dug out of the numbers you published, and I feel like it’s gotta be wrong.

    It looks to me like the *average* romance title of 500 pages, enrolled in KU, is making $17.50 a day, or a bit more than $6,000 a year. I’ve looked at it every way I can think of, and that’s how the math keeps coming out. Is that really real, or have I missed a calculation? Is there something anomalous in the data that’s driving up that number higher than it should really be?

    • Data Guy says:

      Hi, Chloe,

      I double checked, and there’s no mistake in the data. The problem is that our intuition typically misleads us about what those *average* values imply, because of the common non-mathematical definition of *average*, which we tend to instinctively interpret as meaning *typical* or *most common*.

      Mathematically, though, the average earnings for titles in a particular category is simply their arithmetical mean value: which is calculated by summing up their total earnings and then dividing by the number of titles in that category.

      In other words, the average earnings per title can be very different from the earnings of the average title — which is how I think you might be interpreting what the graphs show.

      Because of the uneven way sales are typically distributed among book titles, it’s really only a small minority of titles in each category that typically account for the majority of sales, which means that the vast majority of titles in the category are earning significantly less than the category average.

      It’s like that old joke: E. L. James walks into a bar, and suddenly everyone in there is on average a millionaire. πŸ™‚

      In the case of Romance Kindle sales, only 10% – 12% of the titles in each category are earning the average value shown or more.

      Data Guy

      • Man says:

        Can you supply medians as opposed to means then? It seems like this might be more accurate in representing a “typical” author. Thanks.

        • Data Guy says:

          Hi, Man,

          Frustratingly, in this case median values would provide us with even less insight than averages do. That’s because the median income is basically $0.00 for every category.

          The reason that all median incomes come out zero is that a median value measures the minimum income achieved by at least half of all participants, whereas the majority of authors in each category we measured had negligible daily sales.

          Data Guy

          • Jens says:

            Wait, wait,
            This is kind of important information.
            It suggests that the data you are giving really is only relevant for the top 15% ?) of all self-publishing authors. For 85% of authors, they can safely not worry about free first titles in series, best possible price point, KU or non-KU, because they will be earning $0 or something close to Β£0. How frustrating!

            This does not in any way negate the value of your information, which I find insanely interesting and useful. But this is crucial context. Can you shed more light on the long tail of low earnings?

          • Data Guy says:

            Hi, Jens,

            What you say applies to *all* authors, not just self-published ones; in this data, traditionally published authors aren’t treated any differently. But I’d say it’s definitely still relevant to all authors.

            The reality is that there are far more folks trying (or hoping) to earn a living from writing than ever will. Indie self-publishing has enabled maybe 10x as many writers to do so than before. But it’s still only a few percent of writers who will ever manage to (as opposed to a fraction of a percent of writers in the traditional publishing world).

            Data Guy

  15. AJ Adams says:

    Thanks, Data Guy. This is solid gold!

  16. Thanks for the information. I know this is a lot of work that takes away from your own writing, so I appreciate your contribution to help authors.

  17. Fabulous number crunching, data guy! I have one question about the data on series with the first title free. When you calculated the average per title daily earnings for the series, did you include the free title? For instance, for a series of 2 to 3 novels, the chart shows that the average per book earnings is about two dollars if the first book is free but four dollars if the first book is not. Is that two dollars spread across the whole series or just the books that are not free?

  18. Erica Ridley says:

    This is ah-mazing, and everything my data nerd heart dreamed for. Thank you so much!!!

  19. Jen Crane says:

    This information is so very useful. Anecdotes and sharing among colleagues is all well and good, but these hard numbers will really help in my future decision making. Thanks so much for your time and effort spent doing it.

  20. Devon Royal says:

    Hey, Data Guy, thanks for the useful info! I’m wondering if your rank-to-sales conversion process accounts for the fact that KU books’ rankings are inflated?

    (I know this to be the case because I put one of my series in KU for a few months and those books’ rankings rose and remained consistently higher than my other rankings, despite the KU books selling less units and making far less revenue than the non-KU books.)

  21. This is great! As a hybrid author, definitely looking forward to any upcoming info on that. I would also love to see a breakdown of YA sometime in the future – esp. since many in the industry claim it mainly sells in print, not ebook. I think that is changing though.

  22. Nick R says:

    Curious about genres and sub-genres. Did you count a title in only one genre? Or is there cross pollution? You can choose 2 categories and keywords can put you in more, but a lot of the bestsellers are listed in almost every sub-genre category.

    • Data Guy says:

      Hi, Nick,

      When we look at breakdowns by genre or subgenre, there is a fair amount of cross-pollination, because of the multiple categories each title can be listed under (aside from a few mutually-exclusive pairings, like Fiction & Nonfiction, for instance).


  23. Nick R says:

    Another thought… Shouldn’t the uncategorized be lumped in with Indie? Would it make much of a difference?

    • Data Guy says:

      We leave “Uncategorized” authors/imprints separate, because they haven’t been individually checked and verified as indie.

      A quick scan of the top few hundred tells me that most of them *are* in fact indie, but a few non-indies are sprinked in there, too, so we don’t want to lump them all in with Indie and thus inaccurately bias our data in a pro-indie direction, even by a fraction of a percent.

      With that said, doing so wouldn’t make much difference in the numbers.


  24. I for one not only see the need for a guild but for an indie print and binding cooperative with non-profit options for disabled and underprivileged writers, youth and persons who would otherwise be totally ignored by the system.

    I hope to hear from people about this, I have enough land to build such a site, and would gladly go in with people on such a project, as I have experience with both digital and offset printing and binding.

    I think that the only real way we are going to get a fair deal in this industry, is to come together and create our own ‘union label’ if you will, or perhaps you may call it a confederacy, but still a way to mutually leverage our strength without giving up our freedoms and options.

    Yuri Futanari
    romance novelist

    • Pam Uphoff says:

      I suspect the Indie print issue is not the physical production of the book, but rather the near impossibility of getting the books into bookstores. Figure out distribution, and then printing books is useful.

  25. Dana Wright says:

    Hi. I was wondering what info you have on teen sales with print books vs. digital copies. πŸ™‚


    Dana Wright

    • Data Guy says:

      Hi, Dana,

      Print data on teen/YA is really blurry, because Bookscan has no Teen or YA categories; in the industry stats they publish, Teen/YA print titles get lumped in with “Juvenile Fiction”, a category that includes everything from Dr. Seuss and toddler’s board books to John Green’s Fault In Our Stars and Veronica Roth’s Divergent. But FWIW, about 36% of all US print sales are categorized as “Juvenile” (and 3/4 of that — or 27% of the US print total — is “Juvenile Fiction”). Of the remaining 64% of US print sales that are categorized as “Adult”, two thirds — or 41% of the US print total — is “Adult Nonfiction”, while roughly a third — or 23% of the US print total — is “Adult Fiction”.

      We have a much clearer picture of what % of ebook sales are Teen/YA, because those books are categorized on Amazon as “Teen & Young Adult”. Using the Amazon ratios, we see that roughly 6% of ebook sales are categorized as Teen & Young Adult, while 4.5% are categorized as Children’s. Assuming similar ratios hold at other ebook vendors, we get annual sales by format of roughly:
      – 170 million “Juvenile” print books a year (including toddler board books, early readers, middle grade, and Teen/YA)
      – 30 million Teen & Young Adult ebooks a year
      – 22 million Children’s ebooks a year

      It’s really hard to say what percentage of those 170 million “Juvenile” print books are really Teen/YA… I suspect it’s a fairly small percentage, but your guess is as good as mine. πŸ™‚

      Data Guy

  26. Hi Data Guy,
    Great presentation, really appreciated it. I’ve been spending some time explaining the “First-in-Series Free” slide to various folks. Here’s my stab at summarizing our discussion during the session, because the audience agreed with free-first being effective sooner than the slides illustrate.

    We kind of hashed it down this way (+ a few thoughts of my own):
    -At 2 in a series, you’re giving away 50% of your money, which would mean you’d need to double your sales of book #2 to get to break even. Similarly, #1 free in a 3-long series still has a 33% hit. (Both of these will aid the skew shown on his chart.)
    -The data can’t see if it was free before but is now back up to price. It can only see what is free on that snapshot day (whether briefly promoted or on perma-free). So, if you got mileage, then took it off free, it won’t show up. (Again, skewing up the “paid” bar.)
    -(This following was the audiences main conclusion): If you did a three-book series, made the first-in-series free and it was a success, what would you do next? You’d immediately follow it with more titles…thereby bumping the successful “first free” strategy up into the longer series categories. If it failed (poor writing or whatever), you’d be more likely to leave the 3-book series to die on the vine and go to more fruitful pastures.
    -Data guy was also the first to admit that he could make no sense of the 11+ inversion. Now that I think of it, I think that makes sense as well. If the series has been around long enough to reach 11 titles, most of them are founded back in purely traditional times, so the first book would never drop to free as long as it is in Trad. Pub’s hands. Newer, less established indie series might reach 11 titles, be doing well with First Free, etc, but not have the weight of the high earning titles of a long-running series that started in trad. and still continues, well established.

    Just thinkin’
    Thanks again,

    • Data Guy says:

      Hi, Matt,

      Thoughtful analysis. πŸ™‚

      I felt it made sense to include the “free” titles in the series-length counts. Otherwise we’d be comparing 2-book series against 3-book series (with one free), 3-book series against 4-book series (with one free), etc. Obviously an opinion call, but I don’t think it skews the data per-se, whereas I think the other way would. If you’re an author with a four-book series written, for example, those are the books you have to work with. You can choose whether to make 1 of your 4 books free or not, which is the comparison shown in the chart.

      Re: whether a few of the particular titles were only temporarily free, or were Bookbub-promoted that day, etc., it doesn’t tend to make much difference in our analysis. Because we have hundreds of thousands of individual Romance titles included in the data snapshot, each individual “bar” in the charts is based on thousands of individual series & books. That tends to average out most transient effects. πŸ™‚

      I bet you’re right about the weird inversion for 11+ title series. If we dug deeper, we’d probably find that the orange column are all decades-long-running traditionally-published titles (and only high-selling ones, because traditional publishers are less likely to let a low-selling series reach 11+ titles).


  27. Emma says:

    Hi I was wondering if for your graph on the number of authors per tax bracket there was a way to know what % of the category they represent. Indie are the most in absolute value but what percent of the total self published are they? 1, 10, 40? Same goes for the other they are less I absolute value but perhaps more in percentage…. It would be an interesting data

    • Data Guy says:

      Hi, Emma,

      See the answer to Samara’s question, earlier — she even worked out the percentages for us. πŸ™‚

      One thing to keep in mind is that the 30,000 Romance writers in our dataset don’t reflect all Romance writers — only those with a measurable degree of sales (i.e. at least a copy or two a month). There are probably another 80,000 Romance writers with published titles who aren’t selling even that much (plus a handful who, for whatever reason, don’t list their titles for sale on Granted, many of those 80,000 others have not published anything new in years or decades, and are no longer actively writing, so maybe it doesn’t make sense to factor them in.

      But whether or not we choose to include those 80,000 writers, too, any calculation of the odds of attaining a given level of success by following a given publication path would still unfairly leave out who-knows-how-many tens of thousands — or even hundreds of thousands — of still-querying traditional writers who already have completed manuscripts they want to see published, but who are still stuck in traditional publishing’s “slush pile”. For a truly accurate comparison of the odds, we’d need to find a way of counting and including those querying writers, too — after all, their indie counterparts who had stories they wanted to see published have already published them, and already being counted in the indie totals.

      The thing is, the odds of achieving financial success through writing, in absolute percentage terms, are always going to be daunting no matter what publication path one chooses. Indie writers have a significant odds advantage right now — that’s pretty clear from the numbers. But even so, there are far more writers hoping to one day earn a living from writing than the paying reader market can ever possibly support, no matter how those writers choose to publish.

      With that said, the data also shows that the odds of achieving financial success are far, far better for writers who build up sizeable catalogs of published titles, who publish more than once a year, who price their books correctly, and who target high-demand genres. When you look at only the authors who are doing those things, the odds don’t look quite so steep. πŸ™‚


  28. Traci Hall says:

    Thank you very much for putting this together and sharing real figures!

  29. Amy Jarecki says:

    I saw your presentation at RWA (thank you, it was awesome and incredibly informative). A few days prior I was thinking about metrics and comparisons (quite dangerous, I’ll say). True, romance is HUGE, but there are so many sub-genres. I write Scottish historical romance, some of which are time travels, and recently found a time travel sub-genre in fantasy/science-fiction. I am also plotting a thriller series that will be categorized as romance, but very heavy on the thriller side of the genre. It would be interesting to peel apart the romance genre by sub-genre and compare with the non-romance equivalent.

    For example, does romantic suspense sell more than suspense/thriller? Does historical romance sell more than straight historical? It might not always be apples to apples, but comparing romance as a whole to all other fiction genres can be a tad misleading. I suppose it is all in one’s subjective interpretation, but when you walk into an airport bookstore, there are usually only a few romances on the shelves. When you look at the USA Today, romance has a strong presence, but it doesn’t dominate. Hmm.

    • Data Guy says:

      Hi, Amy,

      We did peel apart Romance by by sub-genre in the presentation & slides, but didn’t compare against non-Romance equivalents. So the comparison you describe might be a very interesting one for the future. OTOH, Romance is so massively dominant in digital that I’d argue it probably does make sense to compare it to all other fiction genres put together.

      Airport bookstores may not carry much Romance, but airport bookstores only make up a tiny part of the (print) market: Hudson news and related outlets add up to around 6% or so. And that’s only print — remember that almost 90% of Romance sales are now digital.

      Comparisons against national “best seller” lists like the USA Today and NYT also have to be taken with a grain of salt. The weekly USA Today Best Seller List, and especially the New York Times Best Seller List, are no longer reliable indicators of which books are truly selling the most copies in the USA that week. Both lists systematically (and editorially) leave off quite a few books each week that are actually selling far more copies than the ones they do list.

      Why? Both lists curate their listings according to an ever-changing set of rules that automatically exclude:
      – any title by an Amazon imprint like Montlake, regardless of total weekly sales
      – any indie-published title that is exclusive to Amazon, regardless of total weekly sales
      – and even indie-published titles that are available at many retailers… but whose otherwise chart-topping sales that week happened to occur *mostly* on Amazon.

      I personally know authors who in one week verifiably sold 15K – 20K copies of a title that didn’t fall into any of the above automatically-excluded categories, and yet whose books failed to appear on the NYT List, the USAToday List, or both.
      Caveat emptor. πŸ™‚


  30. Mimi Barbour says:

    My brain’s fried trying to absorb all his fantastic info! Will reread it a number of times. I can add one thing and it’s that my numbers on KU keep crawling up every month. Not only in my Romantic Suspense books but the contemporary romance novels also. I’m thinking that means that the readers aren’t just reading one story or genre but they are searching and finding others I’ve written and I’m praying that means new steadfast fans.

  31. Kerry says:

    Hello Data Guy,

    Is there a recording of your presentation. If not, do you know if RWA plan to release a recording?

    Thanks for the data and taking the time to add even more data gold in your responses to the comments.

  32. Jami Gold says:

    Thank you for the awesome presentation, Data Guy! I was lucky enough to see it in person, and I gathered my thoughts of some of the highlights in a blog post today. Thanks again for all the information.

    (Also, I sat 3 chairs down from you in the Nielsen presentation and wondered how hard you were biting your cheek at some of their analysis. πŸ˜‰ )

    • Data Guy says:

      Hi, Jami,

      Nielsen’s doing the best they can with the information they *can* collect, and they publicly acknowledge that their data is missing a swath of the market that has grown huge. They just haven’t yet realized *how* huge…

      Nielsen’s RWA-presented guess of 25% (for the size of the self-published sector) comes from their consumer surveys. They readily acknowledge those results are non-scientific and largely directional in nature, not truly quantitative.

      Over time, Nielsen will be forced to add some new data collection techniques to their arsenal, because the fraction of the market that their current methods can actually measure is shrinking every quarter.


  33. Jami Gold says:

    Oh, one question-slash-comment… πŸ™‚

    Is there any way to compare earnings of a book 2/3 in a series WITHOUT a free first book to the earnings of a book 2/3 in a series WITH a free first book?

    In other words, as you mentioned in a comment above, the earnings for a free-first-book series are going to look lower because you’re dividing the number by more titles, but if you subtracted out the earnings (or lack of earnings) for book one in different series, I wonder how the sales earnings of later books are affected. We all know that a free first book is a long-term strategy, and it would be interesting to see what the results on a book 2/3, etc. sell-through actually is between those two categories. Thanks for everything you do! πŸ™‚

    • Data Guy says:

      Hi, Jami,

      I do think leaving the free first book out when calculating per-book earnings for same-length series makes the comparison a bit of an unfair apples-to-oranges one, per my comment to Matt, above.

      But it is possible to eyeball-estimate rough answers to how calculating that way might change the numbers.

      For 2-3 book series, it looks like per-paid-title earnings would end up roughly the same on average, whether you make the first one free or not. (But of course those earnings are on 2 titles instead of 3, so overall series earnings would be on average be lower if the first is made free, as the current graph shows).

      For the 4-5 book series, per paid-book earnings go up about 25% when you make the first free… but overall it ends up a wash on average, because you only get those 25%-better per-book earnings on 4 books now, rather than earning on all 5.

      As always, I feel like we’re just scratching the surface with this stuff — leaving a ton of questions unanswered. Perhaps next year we’ll do a deeper dive on first-in-series-free. πŸ™‚


  34. Fantastically useful stuff, DG. Thanks so much.

    Re the issue of averages, is it possible from your data to provide median author figures instead? For example, for author earnings? The high earners do seem to skew the averages so maybe the median could be more helpful in some cases. Just a thought.

    • Data Guy says:

      Hi, Joanna,

      I agree median figures would be interesting — or even quartiles. Perhaps in a future report we can include those.

      Before deciding to use averages to simplify the bar graphs, I tried a few different ways of presenting the data — including this one (apologies in advance for its horrific geekiness πŸ˜‰ ):

      The above log-log graph provides a complete picture of how many authors were earning how much from US Kindle ebook sales on May 5, 2016. It lets us measure medians, quartiles, percentiles, histograms, or whatever from it. But at the same time… it’s just too geeky to be readable by 99% of people. The nice thing about bar charts and averages is that they are self-explanatory, even though they don’t paint as detailed a picture. πŸ™‚


  35. Lou Cadle says:

    As always, I want to express my gratitude for your work and expenses incurred in doing this. DG, you’re the best. (And skilled in extemporaneous speaking. Since I’ve been able to hear you this past year, I’m even more impressed.)

    Earnings per title by months since publication date: Does that “books continue to sell a little something for 33 months” hold across all genres for indies? If so, it could be very useful to know. I wonder how the success of a book or series alters it. Surely the more popular the book, the longer of a tail it has…but that’s me guessing, and we know how useless guessing is! If you looked at books that sold 10,000 units in their first two months and 5,000 and 1,000 and so on, would the lines look appreciably different from one another?

    That guessing or knowing only our own data is worth nothing is why what you give us is so useful. We can’t effectively plan a career, or assess one, based on wishful thinking, intuition, and stuff we saw written on the bathroom wall in 2012 (in effect. I’ve never actually seen indie business advice on a bathroom wall.)

  36. Rinelle Grey says:

    Hi Data Guy. Some great insights here. Wish I could have been at the actual talk, I suspect that would have been very intersting.

    I’m curious, with the results you’re seeing from first in series free, does that account for the fact that most books that have the first in series free aren’t in KU? The results have clearly shown that books in KU are going to perform better on Amazon, and I’m wondering if this could be skewing the results a little.

    It would be really interesting to see a breakdown of books not in KU (first book permafree vs not) and books out of KU (first book permafree and first not), and see if this changes what we’re seeing here.

    I’m struggling with this one, because it just doesn’t match my experience at all, whereas most of the other data does!

  37. Terry Mixon says:

    Great work, Guys! Is there any chance seeing a presentation like this tested toward science fiction and fantasy such as those in SFWA or Mystery for the MBA? I’m tempted to assume some trends carry across, but I’m sure I’ll be wrong.


    Terry Mixon

    • Terry Mixon says:

      I should never post on a phone that early in the morning. The typos. Shakes head.

      To be a little clearer, I loved the breakdowns you had for romance. Particularly the subgenre breakdowns and earning levels/number of authors in the genre.

      I’d love to see a similarly granular look at science fiction. Not even as part of science fiction and fantasy combined. That’s huge and I think should be separate. Just my opinion.

      I’m betting a lot of people in various genres would love to see how they stack up inside their genre, in average earnings, with a breakdown of the income levels inside the genre, and which subgenres are the most popular and even how the total earnings break down that way.

      I’d love to see you guys hit each of the major genres this way going forward.

      Thanks for doing this. We all appreciate the hard work you do.

  38. Jenny Brown says:

    Data Guy,

    Very interesting charts, but as someone who used to sell charts and data to industry I have to question why you chart Averages when Medians are so much more informative. My income, your income and Bill Gates’ income averages out to many millions, but that is a very misleading figure. With some authors earning huge amounts and so many others earning almost nothing, I suspect that the averages here can be very misleading.

    Using a median with a standard deviation would be even more informative. I know that the public isn’t all that educated about statistics, but medians would give authors a far better idea of what they are likely to earn, and the standard deviations, explained properly, would show them where in the spread most earnings concentrated.

    • Data Guy says:

      Hi, Jenny,

      Your point is well taken. However, in this case the median value would unfortunately tell us absolutely nothing — see my answer to Man above. The median daily income in every single category works out to $0.00.

      Statistically, author incomes are Pareto-distributed, rather than following a normal (Gaussian) bell curve distribution. Thus standard deviations won’t necessarily correspond to our intuitive understanding of what one-sigma, two-sigma, three-sigma, etc. variations imply, either.

      Full log-log plots of rank-ordered daily author earnings, like this one in my answer to Joanna above, can provide a complete picture of how total income gets allocated between authors. But they are too technical a depiction for a presentation of this sort.


  39. Mary says:

    Hello! Is there any relationship between where a romance story is set and how it sells? We have a theory that stories set in the US sell better, but no actual way to validate.

  40. Tsilum says:

    Great presentation, really enjoyed reading it – thank you for sharing!

  41. Banot says:

    Awesome presentation that I enjoyed reading – thanks for sharing it.

  42. Bulli says:

    Hi Data Guy!

    Awesome post! Thank you so much for all this data.
    Why do Amazon Imprint titles sales show such a different (and higher) sales pattern so long after initial Publication?

    • Data Guy says:

      Hi, Bulli,

      I think Amazon imprints aim to optimize each of their titles’ sales over a much longer lifetime than most traditional publishers do. Many traditional publishers remain trapped in a brick-and-mortar print mindset. For a lot of the ones I’ve spoken with, their approach to marketing is still a holdover from the olden days when fixed initial print runs and limited bookstore shelf space meant that they jammed all of their marketing effort for a title into its first few launch weeks, and then–unless that title happened to be one of their very top few megasellers–they essentially withdraw all marketing support from it to instead concentrate on their next batch of brand-new-title launches.

      It’s a totally different business model.

      Traditional publishing is something of a “fresh produce” model (as Kris Rusch so aptly described it). It’s 100% focused on this season’s “frontlist,” with “backlist” titles getting basically ignored from a marketing standpoint. Each new book gets a brief run in the sun, then it’s taken off the shelf and set aside to make way for the next new hot thing.

      Some savvier traditional publishers are starting to realize the limitations of that model in an online, digital world where no title ever truly goes “out of print.” Those publishers are retooling their marketing spend to heavily emphasize backlist as well. But shockingly few publishers have made the shift; most still operate as if a title’s launch window sales were the only thing that matters.

      OTOH, indie publishing is about building long term revenue streams that grow steadily over time as an author’s catalog of titles grows. Backlist? Frontlist? Who cares… it’s all fresh to a reader who hasn’t seen it before.

      The marketing philosophy of Amazon imprints lands somewhere in between, which is why Amazon imprint titles have sales “tails” that last far longer than most traditionally published titles. πŸ™‚


  43. No words for your hard work. I just Love it.

  44. Justin says:

    Hi DG.

    You shared a cool breakdown of title and author counts by publisher type for romance. Do you have something similar for the other major categories/genres?


  45. neha says:

    Hey, Jami,

    I do think leaving the free first book out while ascertaining per-book profit for same-length arrangement makes the correlation a touch of an out of line one type to a totally different type one, per my remark to Matt, above.

    Yet, it is conceivable to eyeball-evaluate harsh responses to how figuring that way may change the numbers.

    For 2-3 book arrangement, it would appear that per-paid-title income would wind up generally the same overall, regardless of whether you make the first free or not. (Obviously those income are on 2 titles rather than 3, so general arrangement profit would be by and large be lower if the first is made free, as the present diagram appears).

    For the 4-5 book arrangement, per paid-book profit go up around 25% when you make the main free… however general it winds up a wash by and large, since you just show signs of improvement per-book income on 4 books now, as opposed to procuring on every one of the 5.

    As usual, I feel like we’re simply touching the most superficial layer with this stuff β€” leaving a huge amount of inquiries unanswered. Maybe one year from now we’ll complete a more profound plunge on first-in-arrangement free. πŸ™‚

    Here’s to you,


Leave a Reply