The Futility of E-Book Completion Data for Trade Publishers

Introduction

Today’s technology makes it possible to track people’s actions to an extent that used to only be restricted to fiction. Unlike in George Orwell’s Nineteen Eighty-Four, however, people today are being watched by numerous companies and agencies, not just one Big Brother. Nearly all of one’s actions are constantly being tracked and measured in some way by one entity or another. In the last couple of years, the manufacturers of popular e-readers and apps have added reading habits to the long list of behaviours on which data is being collected. All of the major e-reading devices – including Amazon’s Kindle, Apple’s iPad, Barnes and Noble’s Nook, and the Kobo eReader – are capable of tracking which books users open, how long they spend reading them, and how much of each book they actually finish. [1, 2, 3, 8] The Kindle’s data servers are even known to store users’ highlights, bookmarks, notes, annotations, and the last page they read. [3] Presumably, every e-reader is capable of doing this by design. [5] In its 2014 report “Publishing in the Era of Big Data,” Kobo attempts to educate publishers on the value of analysing such data as the completion rate of titles.  It states: “We now have the opportunity to act on engagement, not just sales, which over time should lead to a stronger, more viable list overall.” [1] Surely, there is an argument to be made there, but this assertion should not be accepted as fact — or at least not yet.

This paper will argue that today’s trade publishers have little editorial insight to gain from knowing the completion rates of their nonfiction and general fiction titles because, firstly, there is often a lack of correlation between such data and other measures of a book’s success; and more importantly, current e-book reading data offers a very narrow understanding of the general reading habits of present-day book buyers.

Proposed value of reading data analysis

There are many who share the belief that analysing e-reading data can prove to be a useful tool for publishers. Micah Bowers, founder and CEO of e-reading services provider Bluefire Productions, sees great potential for the applications of reader data analysis: “[it] can have an impact on decisions across a publisher’s business . . . whether its search engine optimization, or content structure, marketing or who you give that next big advance to.” [2] Acting upon this potential, Kobo’s 2014 report aims to provide publishers with “a practical guide to data analysis [by laying] out some key methodologies and [explaining] which conclusions to draw from them.” [4] One highlight from the report is Kobo’s assertion that hidden equity and bestselling authors can be found among books with low sales, but high completion rates. It also states that e-reader data can help guide marketing and publicity strategies, and provide publishers with useful insight for negotiating advances and multi-book deals. Furthermore, it concludes with the statement: “Publishers have the power to both create and define culture; the books they create are the stories we tell ourselves about our time, history and the future. Big Data is a window into what truly resonates.” [1] This is a powerful yet highly questionable statement because what can be viewed as ‘catering to the masses’ may help generate fads, but not classics. Noah Genner, CEO of BookNet Canada, supports the idea that “the ability to collect data about how people read has the potential to revolutionize the industry by helping publishers tailor books to their audiences.” [9] Nevertheless, he agrees that it’s also important to publishing classics that resonate with readers for “50 years, 100 years, and those successes are harder to predict.” [9] Length is one factor that Kobo has attributed to low completion rates. [7] Publisher Jonathan Galassi told The Wall Street Journal that, “We’re not going to shorten ‘War and Peace’ because someone didn’t finish it.” [3] However, if some publishers start taking e-reading data too seriously, there should be a real concern that it could compromise the quality of literature in the future. 

More valuable measures of a book’s success

At a Digital Book World 2015 conference panel, Kobo’s president and Chief Content officer, Michael Tamblyn, echoed what is arguably the core message of Kobo’s 2014 report, stating: “Before you had basically two ways to evaluate success. There is the critical response, and unit sales. [Data] brings the reader into the picture.” [2] This statement seems to fail to recognize that direct reader response and feedback have actually been a measure of a book’s success for a long time through online reader reviews, surveys, and fans’ interacting with authors. Tamblyn points to the two traditional measures of a book’s success: sales and critical reception, which includes literary awards. These should undeniably remain the most important factors for trade publishers to track, since one determines their financial success and the other strengthens their reputation. If analysing e-book completion data can help authors and editors create bestselling titles, then today’s most popular books should have the highest reader engagement. However, the e-reader data released to date has shown that the titles with the highest unit sales have actually had low completion rates. According to Kobo’s 2014 data, the majority of readers didn’t finish some of the year’s bestselling titles. For instance, The Guardian reports that 44.4% of Kobo’s British readers finished Donna Tartt’s Pulitzer-wining bestseller The Goldfinch and only 28.2% made it to the end of Solomon Northup’s megahit Twelve Years a Slave.” [7] Since e-reader data isn’t matching up with other clear measures of success, it’s logical to question its accuracy.

Tamblyn himself concedes that readers “may not always read one book from start to finish before jumping into the next great story.” In addition, as he says, they “may wait days, months, or even until the following year to finish certain titles.” [7] This undoubtedly affects e-reader completion data, as does the fact that some people jump between reading the same book in electronic and print formats. As writer Francine Prose proves through her own experience, there are still those out there who prefer physical books and have a habit of reading the e-version of a book out of convenience when they’re on the go, and continuing where they left off in their print copy when they get home. [8] These potential causes of e-reader data inaccuracy show that sales and reception continue to be the most effective ways to evaluate a book’s success.

More accurate perspectives on reading habits

Analysing information from a single e-reading platform offers a narrow perspective on how consumers interact with e-books, since there’s evidence that a number of users are regularly reading on more than one device. According to Jared Friedman, cofounder and CTO of Scribd, the company has found that “10% of Scribd users read on three devices or more in any given month.” [2] Meanwhile, David Burleigh, the director of marketing and communication at library e-book lending facilitator OverDrive, has stated that 40% of their users read on more than one device. [2] These statistics indicate that analysing data recorded on one e-reading platform, such as a Kobo, can lead to inaccurate analysis, since it doesn’t provide a complete overview of how readers engage with e-books. Even more importantly, it currently provides a very narrow look into general reading habits due to the continued popularity of print books. According to BookNet Canada’s market research, print books are still significantly dominating the Canadian book-buying market. BookNet reports that e-books only made up 15% of all book purchases in 2012 and plateaued at 17% in 2013. [10] Assuming this means that over 80% of titles are still being read in print, how could publishers let e-reader data affect their editorial decisions? It’s clear that they shouldn’t – at least not in the case of nonfiction and general fiction. Genre fiction provides a unique case in which publishers may find it worthwhile to take e-reader data into greater consideration, but this is a topic that’s too large to be discussed within the scope of this paper.

Conclusion

Although publishers and authors may be curious to see how readers interact with their books, they would be wise not to attribute too much importance to the completion rates of their e-books. Since this data is not as useful as it may seem, e-reader manufacturers should not be looking over the shoulders of often unknowing readers and tracking their behaviour simply because they have the technology that allows them to do so. This constitutes an unnecessary invasion of privacy for e-reader users, many of whom may not even realize that their behaviour is being tracked. [5, 9] As author Francine Prose writes, “solitude is and has always been an essential component of reading.” [8] People should have the right to keep it that way. Although she recognizes that data is currently being aggregated to disclose patterns of group behaviour, she expresses the rational concern that “Today ‘the information’ is anonymous; tomorrow it may well be just about us.” [8] Although it is possible to disable the tracker on Kobo eReaders, few users know this because the option is not overtly presented to them. [9] If publishers and the major players in e-reading technology want to experiment with analysing reader data, all users of e-reading platforms should at least be given a clear and direct choice of whether or not they want their reading behaviour to be tracked.

References

  1. “Publishing in the Era of Big Data,” Kobo, posted December 25, 2014. http://www.pdflibrary.org/pdf/publishing-in-the-era-of-big-data-home-kobo-cafe.html
  2. Andrew Albanese, “DBW Panel : Can Publishers Take Advantage of Reader Data?”, Publisher’s Weekly, January 15, 2015. http://www.publishersweekly.com/pw/by-topic/digital/conferences/article/65290-dbw-can-publishers-take-advantage-of-reader-data.html
  3. Alexandra Alter, “Your E-Book Is Reading You,” The Wall Street Journal, July 19, 2012. http://www.wsj.com/articles/SB10001424052702304870304577490950051438304
  4. Rich Bellis, “Kobo Gives a Data Lesson,” Digital Book World, October 13, 2014. http://www.digitalbookworld.com/2014/kobo-gives-a-data-lesson/
  5. James Bridle, “Digital reading: not so discreet after all…” The Guardian, August 11, 2013. http://www.theguardian.com/books/2013/aug/11/digital-reading-privacy-problems?et_mid=632149&rid=234308373
  6. Alison Flood, “Big e-reader is watching you,” The Guardian, July 4, 2012. http://www.theguardian.com/books/2012/jul/04/big-ereader-is-watching-you
  7. Alison Flood, “Ebooks can tell which novels you didn’t finish,” The Guardian, December 10, 2014. http://www.theguardian.com/books/2014/dec/10/kobo-survey-books-readers-finish-donna-tartt
  8. Francine Prose, “They’re Watching Your Read,” The New York Review of Books, January 13, 2015. http://www.nybooks.com/blogs/nyrblog/2015/jan/13/reading-whos-watching/
  9. Mika Rekai, “Your e-reader is watching,” Maclean’s, November 26, 2012. http://www.macleans.ca/culture/books/your-e-reader-is-watching/
  10. Kayla Smith, “Ebook Sales and Pricing Trends,” Booknet Canada, March 27, 2014. http://www.booknetcanada.ca/blog/2014/3/27/ebook-sales-and-pricing-trends.html#.VQHVWYF4qFV

One Reply to “The Futility of E-Book Completion Data for Trade Publishers”

  1. I enjoyed this well-written and researched essay. It succeeds at raising some flags and for publisher’s use of reader data by opening up useful questions. After reading the essay, publishers should question whether pandering to the data will maximize sales in the long term, and whether the data can be trusted to be accurate at all.

    The essay moves along gracefully, with interspersed quotes, expert testimony, and the author’s own points of view. This all reads smoothly and convincingly, despite some questionable claims throughout the essay (or at least claims to which I may take exception), until the last paragraph. The concluding section presents an abrupt pivot in the argument being made from one of practical implications of using the data and the data’s validity, to an argument over the ethical implications of collecting the data in the first place. This pivot, right before the end, leaves the reader partially disoriented.

    Still, a worthwhile read for all the advocates of using data for all decision making in publishing and elsewhere.

Comments are closed.