Friday, June 12, 2015

Let's Talk Numbers: Reader Retention Rates Across a Series

So today, completely of his own volition (and because he is a giant data geek), my husband, business partner, and far more sensible half, Travis Bach, is here to share some publishing numbers! Today, we're looking at reader retention rates across series, how many people who buy the first book in a series can be expected to buy the others.

Now, of course, this number will vary wildly between writers and even between series by the same author, but (as Travis is about to explain) looking at your book sales from a top down, data driven perspective can enable you to make some pretty sharp best guesses, or at least ballpark. If nothing else, it's an interesting topic to think about and I thought you guys would enjoy getting a peek behind the curtain for how we make our business decisions here at the Aaron/Bach book factory.

But first, a disclaimer.

This is a numbers post! There will be lots of graphs, tables, and postulations about the math of self-publishing. If that's not your thing, that's totally cool, but you'll probably hate this blog post. I'm sorry. Here is a picture of a Toothless to make it up to you.

And a link to something else to read!

Still with me? Okay, then I'm going to assume you like numbers. Yay! That said, it also should be mentioned that this is a post about my specific numbers. As in sales.

Posting numbers on the internet is always a super awkward thing to do. Talking about your sales in public in any fashion has always been one of those Things You Don't Do as an author. Personally, I think that's silly. Sure we're artists, but plenty of us are also trying to make a living here, and numbers are what make business go 'round. Withholding information about the financial realities of a business, any business, only helps those who are trying to take advantage of the ignorant. Also, I personally LOVE reading numbers posts from other authors, and as they say, you should write (or get your husband to write) what you love :).

All of the above is a fancy way of saying that I'm not putting these numbers up to brag. (There are tens of thousands of self-published authors out there who sell better, worse, and about the same as I do) Nor is it meant to seem cocky or disrespectful to the readers who so generously made these numbers possible. (Seriously, if you bought my book, I love you. You are the reason I am living my dream as an author, and there are not words to express the depth of my gratitude for that. Thank you!!) Everything posted below is offered purely out of the spirit of openness and sharing information that is one of the things we love most about the Indie Author community. The more we share knowledge, the more we all know. Also, one of the stated goals of this blog is to show you the life of a working writer (me), and looking at sales data is a huge part of that.

So, with all that covered...take it away, Travis!

Let's Talk Numbers!

Thanks my love! Hi everyone, I've been wanting to do another big numbers post since there was such a great response to the Kindle Big Deal vs Bookbub article. There's so much to talk about that my potential posts threaten to become books, not blogs, and so today I'd like to focus on a piece of the whole I've been looking at, which is how book consumers move through a series, and how we use that information.

A quick disclaimer though - I can only talk about the data from Rachel's books as that's all we have to work with. That's not much mathematically. This information is completely anecdotal and is not statistically sound.

What is Reader Retention Rate?

How many people are likely to read book #2 (or 3,4,5,6) of a series? This is a question that I ponder a LOT. It's useful for answering questions like:
  • How much revenue do we expect to have this year?
  • Should we write [book x]?*
  • Should we promote book 1 (or book n-1) more?
 *TBH though, commercial value is usually a tiny factor in determining what gets written next. Rachel's enthusiasm + existing series commitments are the primary determining factors.

In seeking the answers to these questions, I've been looking at what I call the Reader Retention Rate (cause the term sell-through is already taken). Reader Retention Rate is the percent of people who read "the next" book.

Generally, this is computed simply by dividing net sales+borrows for the sequel by the same for Book 1. So if 20,000 people bought the first book in a series, and 10,000 people bought the second, then I'd say that there's a 50% retention rate from Book 1 to Book 2.

What Does It Look Like?

Without further to do, here's the retention rate numbers for Rachel's series as of the end of 2014. (Click to enlarge.)

The reader retention rate for Eli is a little weird because the first half of the series was compiled into a 3-book omnibus. I include its sales when calculating book 4, the 34% you see above.

We have plenty of asterisks here since this aggregate ignores things like specific marketing efforts, sales, and so on. However, we've also never done a serious, prolonged free promotion of any kind for any of Rachel's titles. I've been told that those have very poor follow through on readers since many people who snap up free eBooks when they go on sale never actually get around to reading the book. None of our books have been there though, so this assumption (true or false) is not artificially dragging anything down. (Not to say that free isn't a valid tactic, this just isn't that conversation)

Another factor in our favor here is that all books in all series shown are very well reviewed (4 stars+), which means we are at least comparing books of similarly received quality. Fortunately, Rachel hasn't written a book that's tanked (knock on blessed wood then burn as offering), so I don't have any data for how a bad book (or at least one that readers hated) would affect reader retention rates. (Rachel note: probably very badly. Note to self: never write a bad book!)

What I've Learned From This

When I first made this table, I was pretty astounded. Coming from the marketing world where a 5%  conversion rate is amazing, these numbers appeared astronomically high. First, I'm baffled by the 133% for Spirit's End, it has somehow outsold Spirit War considerably. All I can think of is that lots of free sources like person-to-person borrowing, libraries, and piracy carried readers en mass to this final book. Second, and more importantly, I'm pretty wow'ed by the retention rates in general. I was expecting something less than 25%, so these are high numbers and that changes our game plan in some important ways.

I think that the Eli book 3 to book 4 retention rate is artificially low because of the omnibus. At $9.99 for three titles, the omnibus is basically $3.33 per book, a price that appeals to price-conscious consumers, while the the fourth book, The Spirit War, is $9.99 by itself. That's a 3x price cliff between book 3 and book 4, and I'm sure we've lost a lot of potential readers right there because of sticker shock. Even with that gap, though, 40% of readers who started The Legend of Eli Monpress finished the series as a whole (compared to 50% for the Paradox Trilogy).

The third thing I've learned is that retention rates shoot up towards 90%+ after book 2. In both our 3-book and 5-book series, retention rates for book 2 to 3 to 4 to 5... are all generally high.

A well written series is likely a strong factor here, but the real takeaway from these numbers is that they show the importance of getting people to read book #2 in a series. It also helps illustrate why publishers and book sellers love series in general, particularly long ones.

The common wisdom is to get as many people to read book 1 as possible since sale for a series take the shape of a pyramid. Because new readers have to start with book 1, that book will always be the best selling title of the series. So, the logic goes, the more you sell book 1, the better everything else will do (broaden the base of the pyramid you can make the whole thing taller).

That's still a good strategy, of course. Going forward though, we're definitely going to place increased importance on getting people to book #2 in any series as well, especially for series with more than 3 books since those readers will stick around in much higher numbers from then on.

How I Use Reader Retention Rate

When we found out Nice Dragons Finish Last was selected to be part of the Kindle Big Deal (still on sale for $0.99 until the end of June!), we started scrambling to update the back matter with sample chapters and a call-to-action for pre-ordering One Good Dragon Deserves Another. We did this specifically because we'd seen how selling readers on the second book hooks them in for the whole series.

(Rachel note: I've been saying this without proof for years! "The first book sells the second, but the second sells the series" is one of those annoying things I say to people all the time. Nice to finally have some data to back up my previously unfounded suspicions!)

In a more round about way, I also use these retention rate figures to peer into our future.
Guessing future sales has always been an invaluable tool for me when making our income projections. I plan our financial situation out 18 to 24 months in advance. Estimating book sales as accurately as I can is absolutely vital.

Let's use the upcoming One Good Dragon Deserves Another (available for pre-order now!) as the prime example.

Before I explain this picture, I have a confession to make. I'm scared as heck to post these numbers. This is straight from my central planning sheet, the realest deal I have. One thing that Rachel and I both love about the self-pub industry is how open everyone is with their numbers. So I'm putting my trust in you all.

What you see here,

  • Total Nice Dragons Finish Last (NDFL) - this is how many copies sold or borrowed of Heartstrikers #1 that I estimate will have sold by August 1st. 
  • Adoption Rate - aka Reader Retention Rate from book 1 to book 2... I probably need to change that for consistency sake. More on how this is used below. You'll notice I have it set to a "cynical" estimate of 50%. I feel that the data from Eli and Paradox suggests that we should will see something more like 60% to 70%, but I have a policy of always low-balling my guestimates.
  • Sales Schedule - I have noticed that monthly book sales, when graphed, all seem to have the same shape of curve. I'm always working on this, but I have an experimental method of estimating this curve for any book release. As seen above, this spreads out the estimated sales for book 2 across a timeline.
What you see in the sales schedule is the estimated 10,000 copies of One Good Dragon spread out on my experimental book sales curve (slope really). It's only 9 months of sales because I haven't figured out how to estimate long tail yet. Over 2 years, approximately 80% of a book's earnings appear to happen in the first 9 months it's out, so I focus on that for now.

This curve is assigned actual dates on my income schedule sheet based on when actual royalty payments will start arriving. Add a few more books with this same method, say Rachel's next 2 years of writing projects as well as some rough assumptions of performance for older titles, and that's how I figure out how we'll be doing.

For example, since people who read book 2 are very likely to read book 3, I assume a 90% retention rate and estimate book 3 using the same curve as above.

I'm not going to talk about when this book might come out (Rachel note: because it's not written yet!), but it does have a place on my income schedule. Despite planning Rachel's writing so extensively, that plan shifts and changes in minor and major ways depending on countless factors and events. We only announce release dates when we are fairly sure that they are true.

Does this mean I update my sheet a lot? Yup.

"But I don't have a sales curve thingy that I can estimate with!"

I'd love to do a post about common sales behaviors I see in books. Hopefully sooner than later there'll be a Let's Talk Numbers post about that. In the meantime, though, this retention rate is still useful.

This post isn't about, "You should do things my way," but is instead all about an interesting metric and some ways that we put it to use.

For another example....

Should I Write Novel X?

This retention rate and estimation method is pretty useful for evaluating new book and sequel potential. I need to add a disclaimer to this section though - we don't use sales projections to determine if a series is going to keep going. That is a purely authorial decision based on telling a good story. We might make a lot more money from making a series 5 books instead of 3 books, but if the story in question cannot be told well in 5 books, then it's not gonna be 5 books. (Rachel note: damn straight. The story will be what it'll be!)

Lately, though, one question we've been wondering about is - should Rachel write another writing book? Would [enough] people want to read it?

The gut answer is yes, because the writing posts on this blog are very popular and 2k to 10k is popular. We'd like to have some semblance of data though, too, because writing any book is a big decision that carries opportunity costs. Time spent on a writing book is time Rachel isn't spending on fiction, and it's nice to have some solid logic to use when we're trying to decide if that's a good idea or not.

Let's say that 25% of people who've read 2k to 10k will try out another Rachel Aaron writing book. That's a low rate given our data so far, but this is a different animal than a true sequel. Even so, we see that about 30,000 people have bought or read 2k to 10k already, meaning that 7500 people might want to read another one. That's a comforting guestimate that says that Rachel's time spent writing it would not be wasted.

Is there going to be another Rachel Aaron writing book? It's not been started yet, so I cannot confirm anything. All the other decision making factors (time, interest, quality, value to reader) have to be there for a book to happen. This is all here to illustrate some practical uses and applications I've found while studying Reader Retention Rate numbers.

All in all...

I hope you've found this post informative, or at least interesting. It's hard to get good data in the publishing world (or life in general), so we do the best we can with what we have. My predictions here are certainly going to be wrong compared to the real world numbers, but that's not important. What's important is how wrong they are. I'd be really happy if I called it even remotely close; like within 20%. That'd be amazing - or not given how far down my estimates are adjusted. If I'm way off, oh well. I'll keep striving to get better.

If this whole post feels like counting eggs before they hatch, you are probably right. It always makes me feel like I'm making plans with a side of hubris. Those plans are super essential though. Projections like the ones above are what I use to answer questions like "can we afford to fix the car?" or "can our son go to summer camp?" You know, the important stuff.

When you're supporting a family on writing, having the information to answer questions like this is a vital part of life. So I try to support that nagging internal voice, the one that says I'm making plans with guesses, with fairly pessimistic overall performance estimations. If a number feels even slightly too good to be true, I move it closer to the worse scenarios. As I said, I'm not super accurate, nor can I be since no one can predict a book's future, but I've been consistently 'accurate enough' that we've been able to run this place for years without any real fiscal crises; and that's nothing to sneeze at.

Lastly, I'd love to ask if anyone out there has some Reader Retention Rate info they'd like to share with us to help us see the larger picture. I only get to see Rachel's info after all! If you don't want to leave a comment on the blog with your info, please fire off a message via the contact form - Rachel doesn't mind forwarding the info over to me at all. (Thank you honey!)

That's all, thanks for reading,

Rachel here again! 

Isn't Trav the greatest?! Best career move I ever made was marrying a person who loved business and spreadsheets! For real, if you're serious about making writing your job, try to find a Significant Other who's on board with that and excited to participate. They don't have to be a numbers person like Travis or anything specific. You just need someone who will support your ambitions, because the writing life is truly a whole household affair, especially when deadline time rolls around!

Thank you again to Travis for putting all of this together, and thank all of YOU for reading! I hope you enjoyed this look at how the sausage of self-publishing gets made. If you have any questions for Travis about his process or if you have any numbers of your own you'd like to share with us so we can get a better picture of how things work and maybe do a more comprehensive post that doesn't just focus on one data set (my books), please leave it in the comments below or shoot us an email through the contact form. I'll make sure he gets it!

Happy Friday, and happy writing!

Yours always,


Zsquared said...

I love your numbers posts! Keep 'em coming!

One factor you didn't mention -- until recently, only the first three books in the Eli Monpress series were available as audiobooks. I know because that's why I waited to buy book 4! I hate switching between audiobook/print book formats. Have you tried breaking your sales down by format to see if that contributed to the drop-off after book 3?

Anonymous said...

As a reader, I don't think your reasoning on book 2 selling the series is accurate. I read a lot of books and always start at the first in a series. If I like it, I'm going to read the next book and every book after that, I've already made up my mind after the first book. I can count on one hand the number of times I've decided to read the second book without knowing if I wanted to read the rest of the series.

On the other hand, when I read the first book and don't like it, I'm just not going to read any sequels. No amount of advertising or action by the author is going to change that. I still show up as a sale for the first book, but I won't be getting the others.

The only time I can see reading book 2 selling me on the rest of the series is if it diverges into a side character's series. If you have a first book that's good, then the protagonist goes on in sequels and I read those, I might not read a new series that starts after the first book and goes on with a different protagonist. Getting me to read that book might make a difference. Ronald Weasley and the Chamber of Secrets has to work much harder than Harry Potter does.

Travis Bach said...


Looks like my earlier reply was lost. I just wanted to say that audio book sales aren't included here because we don't get unit counts in the royalty statements.

The numbers in that regard are very small though, so I don't think much would change here if included.

Travis Bach said...

Fair enough. I'm not saying that the first book isn't super important though. I'm saying that the second book is much more important than I ever previously thought.

Travis Bach said...
This comment has been removed by the author.
D.E. Mitchel said...

I wonder if the "reader retention rate" (R3) is different for horror or romance genres? When I worked in a library, devoted readers followed the author, waiting for the author's next book, be it Steven King or Barbara Cartland. Mystery readers tended to follow the author if it had the same "detective" or lead character, and felt disappointment if the same lead character was not in the next book. Different profile for SF/F readers? Daniel

Zsquared said...

Thanks for the response! I'm surprised the audiobooks don't make more of an impact (just because I'm biased, I guess, and think Rachel's books make for particularly kick-ass audiobooks), but numbers don't lie...

Nicole Montgomery said...

This is an awesome breakdown! Grateful, grateful, grateful!

I read both fiction series and non-fiction and, if it helps any, here's how I decide:

Fiction - If I read book one of a connected series (i.e. same characters, one long basic story arc), I will keep reading until the author stops writing them, unless they do something really, really weird to lose me. I actually can't think of a series like that that I bought book 2 and stopped there. I'm a super fast reader, so I love series and I love long books. When a series is originally a trilogy, and then they keep going after that, sometimes the second set isn't as good and I may lose interest (although I'll always go back and reread the first. The only other exception is Robert Jordan's Wheel of Time series. I stopped at the beginning of Book 11, because he was introducing new characters and I just couldn't take it any more. Sigh. And the Honor Harrington books - eventually they connect to another separate set of characters in a way that sort of lost me, and I kind of petered out on them, but those are exceptions.

Other series, like JD Robb's "In Death" books, I will also generally read each new one as they come out, but since the stories themselves are more stand-alone, I feel less urgency. There's probably a technical term for those but I don't know it.

Non-Fiction - these days those are writing books, whether craft or marketing, or whatever. If I read one, such as yours Rachel, and you write another one, I will snap it up immediately, even if I'm not currently working on that part of my writing. I just may sit on it until I'm in that mode, then read it. But once I like a writer's style or topics, I tend very much to have "brand loyalty," as they say, and unless the topic varies wildly from what I'm interested in, I'll buy it. I read a lot of non-fiction history (I teach) and once I find an author I like, I'll stick to him or her like reader-glue.

Yes, that's a plug for "please, please more writing books!" :-D

One caveat: I'm not a typical person, being one of those oddballs who reads over 100 books a year.

But thanks for this -- this kind of breakdown is immensely helpful!!

MrFester said...

Interesting read, as a full Audible listener. I am also wondering about the Audio side of the books and the retention rates based on if the VA is any good or one VA is changed during a series.

For example, I was listening to a authors book series that had 12 books in that series and at book 6 she or someone decided to change the VA and that VA was freaking terrible and everyone hated the VA and most stopped listening to the books, at least the rest of them in the series. Now that doesnt mean the rest of the books were bad, its just pure agony for a audio listener to listen to a bad VA.

Travis Bach said...

I would love to know as well. Mainly cause I'm curious, but it would be useful to know for planning things other than straight up sff series

Travis Bach said...

@MrFester I'm curious as well. Our audiobook data is effectively years behind that of our data on print and eBook. Almost nothing to work with for now.

Loredena said...

I'm strictly a reader not a writer, and I'll speak first to my bit in your numbers ;) I bought the Eli trio book because it kept popping up in 'you might like' and posts from friends. But I have yet to read much of it, it didn't grab me at the time I looked at it (I read prolifically, and I get into genre-moods) and I've not gone back to it. So I'm part of the drop off there but there's always the chance that I'll pick it up and get into it a year from now too (one advantage of ebooks, the entire series is always available! With physical books if a series is several books in when I discover it I'll probably never read it due to the lack of availability of the earlier books.)

One Good Dragon is another that kept popping up on my radar and in January I purchased it (at full price!) and then didn't read it for another 4 or 5 months. Once I did though I read straight through, and found your blogs for info on the next, which I've since preordered. So, the fact that it had been on various sales probably contributed to my noticing it, but I didn't hold out for a sale to buy. Having preordered the 2nd, if I read and enjoy it as much as the first, I'll continue preordering the series (assuming it becomes one) until I get disappointed.

In general, as someone who reads quickly, I love series and long books. I really enjoy the opportunity to get to a know a world and its people in depth. If you can get me to read the first couple I'll likely keep reading. But there are series that I've dropped -- either because I got older/ceased to be the target audience (though note that I still read some YA books at the age of 50), lost interest in the genre, or (and this is the death point) lost interest in the series. That's most likely to happen with a complete shift in time and character focus, but that's not the only trigger.

The trick with me is getting me past the first 50 pages ;) I'm a lot more likely to stop reading a book when irritated by something than I used to be, because I have so many more choices now. Jordan's endless exposition killed any shot at my reading the series for instance - 50 pages in, still in the first locale, and I didn't like anyone is just death. with as many yet-to-be-read ebooks on my Kindle as I have, if I lose patience with a character or with the world building early on, I'm more likely to stop reading than I was when I had to hit the local bookstore in forlorn hope of something else to read.

MadScientist said...

Let's see .. I bought the first Monpress trilogy an a ebook from Amazon UK in 2012, paying pounds 5.99, and enjoyed it greatly. After that you were on my radar but I hadn't heard there was anything more from you until I saw an omnibus of 4 and 5 on an Orbit giveaway table at a convention.

I next heard of you when the Fortune's Pawn trilogy started getting a lot of mentions in feminist social media, and realised you were the same person despite the different name. That was enough to convince me to buy them (again, as ebooks).

At this point you're on my automatic buy list, and I've already pre-ordered Heartstriker 2.