Reading Fast and Slow – Observing Book Readers in Their Natural Habitat

Expert publishing blog opinions are solely those of the blogger and not necessarily endorsed by DBW.

Audience + InsightIn last week’s post, I introduced the notion of the “Internet of Bookish Things,” and how ebooks were now nodes on the Internet that could record how books are being read. In this week’s post, I will begin the exploration of what we can learn using this new “superpower.”

One of the memes that has made the rounds at publishing conferences recently is the notion that books are competing with Facebook, YouTube and Angry Birds for readers’ mindshare. Personally, I find the idea that I will not buy a certain book when browsing Daunts Books on Marylebone High Street in London, Green Books in San Francisco, or on Amazon’s website ludicrous. A typical shopper does not think, “I am not going to buy this book, because I am going to spend those six hours on Facebook instead.” He or she might pick Girl on the Train (the girl at the office kept taking about it) instead of another title, but shoppers are not consciously choosing between books and social media as a form of entertainment, and especially not at the point of purchase.

However, what are readers doing subconsciously? How are they actually spending their free time? How much time do they devote to reading compared to other activities? Reading a tweet, for example, is a form of reading—a very social one and a guilty indulgence for many who work in publishing. So are people reading for hours at a time, similar to sitting down for a three-course dinner (slow food), or are they consuming books like snacks (fast food, in other words)?

I borrowed the title for today’s post from psychologist’s Daniel Kahneman’s excellent book Thinking Fast and Slow, which details how the human brain makes decisions. The question I am interested in getting at is whether reading a book is similar to consuming fast food or more along the lines of slow food? Put simply, are we reading fast or are we reading slow? Let’s have a look.

GraphTo the left is the graph of a reader who is devouring a book from start to finish in just two days (Monday, November 2nd and Tuesday November 3rd, 2015). The graph is basically a calendar showing eight days (Nov. 2nd–9th) with every block of reading shown in blue.

Two days to finish a book? That’s fast reading! This is reading as “fast food.” It will not come as a huge surprise that this book was a pretty saucy romance novel. Romance readers are among the most voracious readers, and they read fast. “Fast” here means not the number of words per minute read, but rather the velocity with which the reader completes a book from start to finish, with all the breaks and interruptions in between, including sleep (no reading between 10pm and 5am for this reader).

What one also notices are the frequent interruptions. The book was not consumed in one continuous session. Rather, there were seven sessions that lasted more than 20 minutes without interruption (none was more than 45 minutes) and many shorter sessions that lasted from five to 20 minutes. In total, there were more than a dozen reading sessions that were just five to 10 minutes in length, in addition to many more in which the book was picked up but quickly abandoned again, as the user switched to email, Facebook or Twitter.

More than half the readers we observed reading this book completed it in less than three days, so the above reader is pretty “typical.” It was a “high velocity” book, a term we at Jellybooks prefer to “fast read.” This measure is a good indication of a book that is keeping the reader glued to the story.

But it is not just romance where this phenomenon is observed.

In a recent study with a Big Five publisher, we recorded people reading one of America’s most beloved authors whose writing deals with complex topics, family dramas and much more. Yet her style of writing is so spellbinding that readers cannot help but return to reading her books for several hours every day until they have completed the novel.

GraphIt took the median reader about seven days to finish this book, which still makes it a “high velocity” title. The reading pattern of one of those readers is shown on the right. This is also the kind of book for which we observe sky-high completion rates.

The typical book struggles to retain the attention of more than half its readers, meaning the majority of them gives up on the book and never finish it. However, a high velocity book is often completed by 70 percent or more of readers. This is typically a solid indicator of a book that is strongly engaging the reader.

By observing readers in their natural habitat—reading at home, on the commute or on the beach—we are measuring the level of emotional engagement a book is stimulating in the reader. A great book pulls the reader back despite all the interruptions and distractions of modern life. A great book stands its ground against Angry Birds and adorable cat videos, and we can quantifiably measure what kind of book achieves this by watching how readers engage with the book.

A good editor has a nose for what a great book may be, but ultimately it’s only the act of observing readers that can confirm if the editorial instinct was on the mark or not. Until now, we have used sales data as a proxy, but buying a book is not the same as reading it, and only actual reading tells us how well a book resonates with the reader.

Sometimes, we don’t have the time to read, though, so below is the graph for the same book but read by a low velocity reader who could only spare a bit of time late at night.


It took this reader more than a month to finish this book (started on Saturday, December 6th, 2015 and finished on Sunday, January 18th, 2016). What is noteworthy is that the reader came back most nights to the book.

Having looked at high velocity books, what does a low velocity book look like?

To answer this question, we turn to a study of the audience for a novel that quite recently won the Man Booker prize. It took the median reader almost a month to finish it. Below, however, is a visualization of a reader whom it took just three weeks to read the highly acclaimed novel, starting on October 21st and finishing on the November 9th, 2015; this person was one of the faster readers for this book. This is not a time-constrained, but an attention-constrained reader, as we observed that this reader also read two other books at the same time.


The reader is quite typical for this book’s audience, as it took the median reader a month to finish it. It was a demanding novel dealing with a difficult subject and written in a literary style. This is the equivalent of a five-course dinner at a Michelin-starred restaurant. It may represent the epitome of English literature, but it is decidedly not mass market. Only 12 percent of readers managed to complete the book. The rest gave up—not because the writing was poor or the book “bad,” but because they were in search of more digestible entertainment, and because the distractions of modern life lured them away.

We have only done a brief walk through the treasure trove of reading data this week, with more to come over the coming weeks. Reader analytics is allowing trade publishers to better understand how strongly a particular title is engaging readers and what the actual audience for a book is (in many cases, the data was recorded through focus groups prior to publication date). This allows publishers to better position a book, to better plan marketing approaches, or in general, to be more data-smart about the “discoverability” properties of a book. We are moving from digital being a format—a “container” of a book—to a tool that allows us to better understand readers and be data-smart publishers.

In future posts, I will further explore insights gleaned from the data, including how demographic factors like, age, gender and reading device shape the reading experience (in the above graph, for example, medium blue represented reading on an iPad and light blue on an iPhone).

Note: All the data reported in this post was collected in studies financed by Innovate UK. EPUB3 files modified with candy.js by Jellybooks were used in the studies. These ebooks were capable of recording and storing offline the user’s reading behavior when using iBooks, Adobe Digital Editions (ADE) and selected Android reading applications. The data stored within the ebook file was extracted when the user clicked a “sync” button in the book. Most participants in the study received their ebooks for free in exchange for pledging to sync their reading data with the study organizers. While a certain “free ebook” effect may have happened, we conclude from interviews with readers that the effect is small. Most readers pretty much forgot that their reading was being recorded. The primary influence on a user’s reading behavior seems to be the content, the quality of writing and the demands on the user’s time, not the monetary cost of the book to the reader.

We will also be holding a workshop on data-smart book publishing at the upcoming Digital Book World Conference in New York City. The workshop takes place on Monday March 7th from 2pm to 5pm, just prior to the main DBW conference. We will look at the challenges publishers face in collecting data, making sense of data and applying it so they can publish smarter, more efficiently and more profitably. The workshop will include speakers from Elsevier, Bonnier and others sharing their experience of turning themselves into data-smart publishers.

To get all the ebook and digital publishing news you need every day in your inbox at 8:00 AM, sign up for the DBW Daily today!

3 thoughts on “Reading Fast and Slow – Observing Book Readers in Their Natural Habitat

  1. Palessa

    It is so funny that you mentioned the Man Booker prize because I had a good friend and fellow author really go in a rant about a Pulitzer winner. It was 800 pages and just unreadable, according to her, but it won the Pulitzer. She is a really great writer herself so I took her word for it. I personally don’t think I would ever win either that, the Pulitzer or the Nobel because, I’m one of those, “twisty-turny, must be consumable” readers and that’s how I tend to write. I’m all about drawing the reader into the story as best as I can. If someone takes more than a week to read my book, they don’t like it. But I like this type of data for sure. Helps reinforce my evil plan of making readers addicted to my stories and toying with their emotions. 🙂

  2. Bruce Watson

    I love the analytics but isn’t there a contradiction here? On one hand you say that no one stands in a bookstore and compares a purchase to the time spent on social media. But then you write of the Booker novel: “The rest gave up—not because the writing was poor or the book “bad,” but because they were in search of more digestible entertainment, and because the distractions of modern life lured them away.” Exactly. Study after study suggest that attention spans are shortening, novels are in fact competing for time, and losing. And a generation raised on social media will probably never make it to that bookstore.b

  3. Andrew Rhomberg

    Hi Bruce,

    I don’t really see the contradiction. What I am trying to describe is the psychological inconsistency in the human mind:

    yes, social media distracts from book reading (observable)

    no, we don’t think “I will spend that time on Twitter rather than reading a book” when picking a book in a book store… (in fact a lot of book buying seems to be done on emotional impulse, rather than based on Vulcan-style logic…)



Your email address will not be published. Required fields are marked *