How Books Are Made

Measuring reading for smart marketing

Arthur speaks to Andrew Rhomberg, the founder of Jellybooks, about how publishers use smart ebooks to measure what readers think of a new publication, and to figure out whether it could be a bestseller.

It is one of the marvellously crazy things about publishing that most books are published long before you have any idea whether they’ll be popular. Publishers will spend small fortunes on advances, editing, design, digitization, printing, and marketing before knowing whether a book will sell more than a few hundred copies. Perhaps early reader data can help solve that problem – something Andrew and his team have been working on for ten years.

Andrew has a broad, global perspective, coming to publishing from chemistry, telecoms, and music, and having lived and worked in Denmark, Austria, Italy, the USA, the Netherlands, Russia, and the UK.

Links from the show:

This episode was published on 2 November 2020.
Supported by Electric Book Works: publishing reinvented for the digital age.


Arthur Attwell 0:05
Welcome to How Books Are Made, a podcast about the art and science of making books. I’m Arthur Attwell.

Arthur Attwell 0:11
It is one of the marvelously crazy things about publishing that most books are published long before anyone tests whether they’ll be popular. Publishers will spend small fortunes on advances, editing, design, digitisation, printing and marketing before finding out if a book will sell more than a few hundred copies. I have never been a good judge of popular tastes. That lack of judgment has led to some very expensive publishing decisions in my past. I suppose what would have been useful is some kind of window into what people really enjoy reading, some kind of magical prototype book that will ask readers what they think of it and let me know.

Arthur Attwell 1:08
Of course, that’s long been a holy grail of ebook publishing: lower early production costs, and an automated way to track what people read and what they enjoy. For the last 10 years, Andrew Rhomberg has been working on that and his company Jellybooks. He’s learned more about how to measure what people enjoy reading than anyone I know. Coming to publishing after a career in chemistry, telecoms, and music, having lived and worked in Denmark, Austria, Italy, the USA, the Netherlands, Russia and the UK, he has a broad, global perspective that I find really valuable. I’ve been looking forward to asking him all about it. Andrew, thank you so much for joining me here today on the podcast. It’s a pleasure to speak to you.

Andrew Rhomberg 1:58
It’s my pleasure, always fun to talk to Cape Town from London.

Arthur Attwell 2:01
Last time we met, we were in a coffee shop in Cape Town about 10 years ago. Our entrepreneurial journeys have covered a lot of ground in that time.

Andrew Rhomberg 2:09
Yes, January, it will be, in just about three months, it’ll be the 10-year anniversary of that meeting.

Arthur Attwell 2:15
That’s amazing. Which makes it pretty much somewhere around the 10th anniversary of Jellybooks, is that right?

Andrew Rhomberg 2:20
That is correct. I met you I think two weeks after we incorporated the company.

Arthur Attwell 2:24
That’s amazing.

Andrew Rhomberg 2:25
A long time it has been.

Arthur Attwell 2:27
It has been a very long time. Before that, you worked for big companies like Shell and Skype after a PhD in chemistry from MIT. What has kept you in book publishing all this long, all this time, because it’s clearly not the money?

Andrew Rhomberg 2:43
Obviously not the money, that’s for sure. Book publishing is a way of getting poorer, not richer in this day. It’s an interesting origin story how I even drifted into publishing. Before I joined Skype, I worked with a company called gate5 in Berlin. We did location-based services and maps. Their Chief Financial Officer started a company called Texter who did one of the first eink readers and the first Adobe, IMSDK-compatible app. So an ebook app that could handle certain types of DRM. That’s how I drifted into publishing. After a while, I decided to do something different and strike on my own and overstayed there. Somehow, publishing has never ever left me out of its clutches again.

Arthur Attwell 3:32
It does that to people, it seems.

Andrew Rhomberg 3:33
Indeed, we are both members of that club, I think.

Arthur Attwell 3:36
Absolutely. Now, for me, it’s, there’s always it always feels creative, and the people are lovely. I can, I can set the money aside because those are rewards in themselves. Hopefully, it’s something like that for you.

Andrew Rhomberg 3:47
Indeed, indeed, if compared to the music industry where I also used to work at Receiver, I find publishing much more collegial, much more handshake-based, much more friendly in a way, and has in some ways had a much more stable business and economic environment, which is also the biggest curse because it has had the least pressure to really, really innovate.

Arthur Attwell 4:14
Really interesting. Since we did have a chance to catch up recently, I want to see if I can describe what Jellybooks does. Then you could tell me if I’m close and what I’ve missed? The way I understand it, a publisher will come to you with a book before it’s been released, before that book is published. Then you’ll give that book to people for free as an ebook. In return, they let that ebook measure how they read it. They … how far they get, how fast as they read and so on. That data that you collect from all these people reading these, these books is really useful to the publisher. I know that sounds like a simplistic summary of what I’m sure is much more complex operation but how close am I?

Andrew Rhomberg 4:55
Quite close, although I have to say it was our second take of the company. We started that particular element in 2014, so year three of our existence, in our project we did for Penguin Random House and the UK Government. Now it’s one of five services we do, but it’s certainly the best-known. It was our start in how to record reading behaviour and really watch how people read, kind of stare over the shoulder while they’re reading a book, so to speak, and understand … are they starting the book, are they finishing the book, how do they react. It’s first commercial application for that technology was testing books before they are published. Sometimes we send out books with different covers to different people. So we A/B test covers, and see how the reaction of people differs, especially books where maybe the publisher was a bit unsure what is the audience reaction going to be, but also many times, who is the audience for this book? Really understanding how to match a book to the audience, and that has become much, much more important. That was our start in data-smart services.

Andrew Rhomberg 6:08
From there, we branched out to do some other applications also that are very data and analytics heavy. So now also been around, how do we create reviews and create feedbacks from people? We will be announcing, at the beginning of November, a large project we are doing here in the UK with one of the big retailers to build a cross-industry platform for ebook excerpts and audiobooks snippets. There, we will also collect data and try to understand, are books correctly classified? What kind of books are read by the same people? By creating a national infrastructure and seeing how people discover books, how they sample them, we try to understand more about the general trends, but also making it easier for people to sample books. UK retailers have no equivalent to the look-inside of Amazon, for example.

Arthur Attwell 7:02
Wow. What exactly do you measure when someone is reading a book? I know that there are a number of axes that you like to get some metrics against, like the speed of reading and so on, and the velocity, which is an interesting term. Could you explain a little bit for me?

Andrew Rhomberg 7:17
Certainly. The so-called objective criteria, observational criteria, we collect. One of them is, do you actually finish the book? It’s quite surprising how many people will give up on a book because their attention just drifts to something else, or to another book. The completion rate for a book is quite critical. In every category, or genre, you have books that may have very low completion rates, and books that have very high completion rates. If there’s any general tendency, it is that maybe nonfiction books are a bit more likely not to be finished. But within fiction, it’s not to say that literary fiction or romance has higher or lower, it’s really from book to book, huge variations. Velocity is something different we look at. So we look at the first day you picked up the book, until the day you actually finished. So with all the interruptions included, so this is not how fast do you turn the pages, but are you glued to the book? Are you going to read through the night so to speak, does … does this book make you return every day? So say a crime novel that has a velocity of less than seven days, so you start and finish it in less than seven days, it’s a real page turner. Is a book, has it such a strong narrative arc that it’s genuinely measurably a page turner? Yes, that can be measured. Whereas say a biography where you maybe in between go to a few other books and dip in and dip out, and it takes you five weeks to finish. Yep, that’s not exactly the same page turner. If you have a crime novel, say, where it takes people three weeks to finish it on average, that is a big, big warning sign. Whereas for biographies, it’s normal.

Arthur Attwell 9:13
That’s amazing. I’m interested in this concept of cross-media testing that I’ve seen mentioned, how does cross-media testing work? For instance, asking people what they watch on Netflix.

Andrew Rhomberg 9:25
Our service has, the classical way the service has evolved, with every month and year you always come up with new ways of probing how people read and how to understand more about the audience. One of the first things we did is, how often do you read in the genre, and see if … how the data is influenced by your genre affinity. So um, true mass-market titles should be very independent of whether this is your preferred genre or not. If something is the true niche title, you’ll see much higher completion rates and KPIs among genre fans and it will perform very poorly about people outside. The next such question we did is, have you read the author before or not? To measure, is there a fan effect of fans more likely to finish the book, versus newcomers or not? So is the author’s platform growing? Is he attracting new readers? Or is he just living off old glory? Those were the first questions we did.

Andrew Rhomberg 10:24
Then we started, when the big debate happened – is Netflix distracting people from reading? – we thought, let’s ask people about their Netflix viewing habits or their TV serials habits and their film habits. In particular, we asked about serials because if you watch a series, that means you’re probably a fan. Whereas having seen one movie once off doesn’t mean that you necessarily liked it. We very much focus on series behaviour like Netflix, but not just Netflix, we also look at TV. We ask people, what do you watch? Then we take that data and see how do completion rates, how do velocity, how do KPIs such as how much I like the book, or would recommend it to others, depend on what your cross-media affiliation is.

Andrew Rhomberg 11:14
We discovered that some movies are very good proxies for genres. When we test in Germany, a long-running TV series called Tatort, it’s the perfect predictor for you liking crime. Both in Germany, UK, and US, if you have watched one of the Harry Potter movies, you are likely to be someone who likes young adult in general. If you have seen Lord of the Rings, or The Hobbit, you’re a fantasy fan, and we see that 80, 90% of all people who read in those genres will have seen those movies, so they become very good predictors. Then there’s the extra dimension of asking about all sorts of other, more specialised movies. So in crime we might put in things like Babylon Berlin, a series set in 20’s Berlin, so a very historic setting. You could ask about Narcos, a very brutal kind of TV show, right. You can see that how people read can vary greatly depending on which genre they really enjoy. So we have seen that an author we tested was much more on the brutal side of like Narcos, Miami Vice and these kinds of films, whereas certain authors might have gone heavily in the direction of say, Sherlock Holmes, very introvert.

Andrew Rhomberg 12:37
We had one book which was pitched as being Scandi Noir, so we thought this should correlate to The Bridge or Wallander or The Girl with the Dragon Tattoo but it didn’t. It correlated with Shetland, a very different kind of crime series. An interesting beta crime series is based on the novels of Ann Cleeves. So we knew that people who like this book would correlate with Ann Cleeves’ novels, which is the classical comparable, the comp of the industry. Now you might wonder, why don’t we ask people about other books? What we consistently find is that readership is so diverse, that if we ask about specific books, the overlap is quite small. All, everyone in the industry always thinks about other books. But we find it’s much better to define it by TV habits to really get meaningful data and good overlap. Defining audiences by their taste in movies and TV shows gives us a much better insight into how to segment audience than asking them about other books, which runs completely contrary to how people inside publishing look at books.

Arthur Attwell 13:51
What decisions do publishers then make with this data? They’ve got this data about correlation between reading habits and preferences and so on. What are the kinds of decisions that they make with it?

Andrew Rhomberg 14:01
For some books, it’s the decision, do we go ahead and this is the lead title and we throw a big marketing budget after it? Or is it maybe not going to be such a big head and we give it a much more modest support, both in terms of marketing and PR support? This can go up and this can go down. Sometimes we A/B test and then it’s about, how do we get the right kind of cover that not only makes the book be picked off the shelf, but also makes people expect the right kind of novel behind it or story, and above all covers that also encourage people to recommend the book to others. This has been one of the surprising things we found, everyone thinks about covers as what seduces you. But the even more important aspect of covers is, does this encourage you to recommend a book to others? We never consciously think about a cover when recommend a book to others, but we have found time and time again that it has an immense influence on people. You’re just not as likely to recommend a book where you find the cover too ugly or misleading, or suggest something other than what’s inside, as it may really dramatically reduces how people recommend book. It’s immensely important.

Arthur Attwell 15:18
I could so see that in myself, I wouldn’t want to have a book out on my coffee table when someone visits that made me feel embarrassed to have it out there. The cover is everything in that case, I suppose the same as when you’re recommending a book.

Andrew Rhomberg 15:32
It’s also interesting, we ask people, how much did you enjoy the book? After they have finished. We only use the data for people we can verify have finished the book and ask them, would you recommend this book? For some books, these numbers are the same. Whereas for other books, the recommendation factor is much lower than the satisfaction index. So those are the guilty rates, so to speak, very dominant in romance, even more dominant in erotica, as you might imagine. But for example, you also see it in the crime genre, the more brutal, the more sort of masochistic or gothic it is, the more it is a guilty pleasure, what Benedict Evans used to call the books for my bedroom, as opposed to the books for my livingroom. The reverse effect can also happen where people said, I didn’t enjoy that book that much, but I will recommend it.

Arthur Attwell 16:27
That’s interesting.

Andrew Rhomberg 16:28
That happens to certain prize winners and classical books where every, people think they ought to recommend it, but personally, they didn’t actually enjoy them at all.

Arthur Attwell 16:37
[LAUGHS] There’s, so many of those on my shelf that I have, because I think I’m supposed to have them.

Andrew Rhomberg 16:42
Well, actually, we did, two years ago, a few campaigns on our own initiative, testing some of the classics. On the one hand, we verified what everybody thought. So the completion weights for Anna Karenina were like 3%, but Call of the Wild, a book quite old by now, almost 100 years, and still gets 60, 70% completion rates at the same level like Gone Girl, basically.

Arthur Attwell 17:11
That’s amazing.

Andrew Rhomberg 17:12
Yeah, Moby Dick to take one out now which has also a reputation of being difficult … 10% completion rate.

Arthur Attwell 17:21
That one doesn’t surprise me, it’s a tough book. What’s a bit of a breakthrough for me is realising that this collection of data is a really good way to determine your marketing budget and where you’re going to spend that marketing budget. Because as a commissioning editor publishing person, I tend to think the data must have to inform the publishing decision. But often the publishing decision is done. Publishers like to trust their gut on publishing decisions. But this is a marketing tool. That’s, that’s really interesting to me.

Andrew Rhomberg 17:50
It was the obvious way how we could break into the market because it was also where people were more willing to pay attention to data. Marketers are much more data-oriented. In the early days, we had huge resistance from editors, because they feared it as something that would interfere in the editorial decisions. Though the tool doesn’t tell you anything about, you should rewrite this. Yes, we can say this ending was poor, because the recommendation factor dropped like a stone because the ending was too open. Very common correlation. But the books don’t tend to get rewritten at that stage.

Andrew Rhomberg 18:29
After a few runs, we find at most publishers that the editors come around to it because it’s kind of the equivalent of the applause in a theatre or cinema, you get a feedback loop of how well it was received and above all, why it was well-received. So editors take it as a cue sometimes to say, will I continue with this offer, and go deeper, or do I drop them? But also, how do I make the next acquisition decision? Where should I focus? What were the elements that made this book really a success? Because sometimes we miss understand why something is a success. Acquisition editors often saw it as a very good tool to better understand the decisions they have made in the past so they would be better informed in the future. Data is something that informs you how to make better decisions. It does not make the decisions for you.

Arthur Attwell 19:27
Really, really makes so much sense. What kind of correlations, and maybe you’ll have some examples, have you found between the patterns you see in the data and whether a book is going to be a bestseller?

Andrew Rhomberg 19:40
We actually use a colour code internally for … blue is weak and yellow is good and green very good and purple, excellent. We don’t have bad books, only weak books. If a book has a certain constellation of these parameters, we can with some confidence predict that if it is in the hands of a good marketing team and with enough support, it will hit the top 10. So both in Germany and the US and UK, we have consistently seen that happen. The criteria for fiction is completion rates have to be over 50%. A book that doesn’t hit that will not take off, generally speaking. You need a satisfaction index, which is a form of the net promoter score, over 50 points. So at least half the readers must say, I really, absolutely enjoyed the book. The recommendation factor has to be such that at least 40% of people said, I will definitely recommend this to others with not too many detractors. And you need a cover match factor of over 30 points. It has to be at least a reasonable match. You can have books where everything else looks great. You can ruin a bestseller with the cover.

Arthur Attwell 21:02
Wow. Do you A/B test the covers?

Andrew Rhomberg 21:04
Not all the books. We do A/B test, but we don’t A/B test small differences. We A/B test more very conceptual differences, different images, you know, sometimes radically different background colours. Think about it more as the mood of a cover. Is this a dark mood, cheery mood … colours convey a lot of information. You don’t always want the cover that gets the most attention, you know, the equivalent of posting kittens on Instagram. You know, you get cat lovers, you don’t get necessarily book readers. You have to have an appropriate cover. That can be tested very well. We can have covers that really get people excited and click through, then afterwards, they said, what did I get here? I expected something totally different. Covers should not be misleading. That’s one of the mistakes we sometimes see in the industry.

Arthur Attwell 22:02
I can imagine that for many publishers, gathering and analysing data like this might seem expensive, given the kind of money they’re used to spending on data gathering, if any. How do you think about the costs and benefits of doing this? How do publishers choose which books to subject to this kind of research?

Andrew Rhomberg 22:23
The costs are not that high. Of course, what is not high for a big publisher may be high for a tiny, tiny, tiny publisher. I mean, when we are testing with 500 to 800 readers, the cost runs depending on the sophistication of the test, between about £500, £950 at the high end. So it’s not the world, especially a test of £500 or £100, or £1,000 relative to a £50,000, £60,000 marketing budget. It’s tiny. But not every book comes in for testing. Because sometimes they are absolutely sure this will be a bestseller and they don’t bother.

Andrew Rhomberg 23:06
We tend to get books where there is division in the publishers. So the typical Monday meeting, editorial, marketing can’t agree or everyone has a different opinion and the like. It comes in for testing and find it out, or they’re not quite sure who the right audience is. Or they have published a few books, it’s going okay, but should actually go, be doing better, and they ask themselves, who exactly is the audience? Can you give us a bit more information? Who is reading this book and who we should be targeting? Or if the book sales are not ideal, is it the content? Are we doing something wrong? Or is there something we could improve?

Andrew Rhomberg 23:47
I think it’s not necessarily the monetary cost that’s the biggest barrier. People require time to read. A test of this nature does take five, six, seven weeks. It has to be done early enough so that you can still act on the data. If your manuscript is very late, or just comes in very shortly before publication date, if you don’t have the time to react, the test doesn’t make sense unfortunately. I wish we could change that. But we can’t force readers, 600 readers to all read in 48 hours, it’s unrealistic.

Arthur Attwell 24:26
It would it would not be a fair assessment of their interest in the book, I suppose. I really like that your readers get to make a conscious choice to share their reading data with you in return for the free box. It seems to me that if our data is that valuable that we can get books for it, we should be able to use it like currency deliberately rather than have it hoovered up and used without our knowledge. How do you think about the ethics of the data gathering?

Andrew Rhomberg 24:54
We started, ironically, in Germany, despite being a British company. Germany has some of the toughest rules, but also most skeptical users. People said it wouldn’t work and it worked quite well. One item is we were very open and were transparent about what we were doing. We made it very easy, including when you submitted the data, it said, thank you, and the readers thought it was kind of like the book saying thank you to them. We very accidentally gamified it a little bit without having designed it that way. We discovered, readers really liked publishers and authors listening to them and paying attention. This kind of feedback loop was important. But we also gave readers the access to their own reading data. It is eye-opening to see how you actually read a book, sometimes how many more pauses and interruptions, there are a little bit of the Fitbit for ebook element there.

Andrew Rhomberg 25:51
Above all, you use the data only to inform certain decisions, and you don’t take it and sell it now to third parties, and there’s no use to marketing to you totally unrelated things. I think that is the really creepy element of data use. In our case, because we’re always commissioned by one publisher, it means the competitor couldn’t get the data, which surprised a lot of publishers who are used to the Nielsen model that they can look up all the competitor data. That’s not how it works in our case. We do market studies for one publisher, at least in the classical settings.

Andrew Rhomberg 26:28
With the samples, it’s a little bit more like you can also see market trends and more general stuff. But we’ve always had a philosophy, the data we collect is immediately within this constrained use. We do not broker data into other industries or through brokers and similar because you’re really totally dependent on the trust of the readers. If you’re consistent with that, and transparent, then I think readers are also willing to share the data with you and see the benefit. But you really have to be I think, consistent and transparent about how the data is used, and what is the benefit to the end user, too.

Arthur Attwell 27:13
Yeah, sure. When we talk about data and Amazon collecting presumably a vast amount of data through Kindle that no one has access to at all, and they have their own publishing imprints as well, which must give those publishing imprints an enormous advantage. Do you see yourselves as a way to help publishers compete with Amazon’s growing power in publishing?

Andrew Rhomberg 27:39
Well, yes and no. What shocks a lot of publishers is how good our relationship with Amazon is. We have in the background often shared some of our findings, and they’ve told us if we were right or not, or said look here closely, so it helped us validate some of the trends we’ve seen. If you sit inside Amazon, you have a very strong infosec department, there are some things you can’t do, because the colleagues on the other side of the world will not let you. There is not so much data access inside Amazon as some people might imagine. Amazon does pay attention to privacy, even inside their own company. Think about Kindle Unlimited where they determine book finished or not based on last position read. So they used a shortcut, or very easy to obtain signal, which is something also we do in our cloud reader. Where did you, what was the last position you read? But that’s not a good proxy for how much you read or watched, you have gone through. That was massively abused by some spammers, when they found it out.

Andrew Rhomberg 28:45
Now, in its own publishing, Amazon does look very much at certain kinds of data. Amazon’s equivalent to our velocity data is something they look at very, very heavily. Also, completion rate is something they look at a lot. They don’t use our surveys, so they don’t have the kind of survey data we have. They’re reliant on the same completion rate and velocity as key parameters. Those are the most critical observational parameters you can collect, and they use that to assess who really gets high engagements or not. How high are the sales? Or, how high are the downloads? Are people really engaging with that offer? They’re using that data to inform their own decisions, what to publish, but when they take on an offer, they also think differently to how trade publishers think. They’re not thinking, how many books would they sell across the whole industry? Because it’s difficult for them to sell into Barnes and Noble or similar. They say, who is the audience? Who is the typical user I have, on Amazon, who regularly buys from me, who buys from me ebooks, and how many units will I be able to sell to this captive audience I have?

Andrew Rhomberg 30:03
So they are quite modest sometimes with their advances and they base all the calculations on kind of their own captive audience. Will this sell to this audience? They’re incredibly audience oriented. It’s interesting to observe that a lot of the industry, the trade industry is now moving in a similar direction. Who is the audience I as a publisher can reach? Where do I have a contact? Where do I have a mailing list? Where I have to contact through booksellers? Can I reach that audience and what should I be publishing towards that audience? But they don’t have the same feedback loops, yet. They only have sales, whereas Amazon really looks at engagement data. There we are trying to help publishers move more closely towards that goal. But there is also some resistance, because publishing has grown up with as long as it sells, that’s all that matters.

Arthur Attwell 30:59
As we begin to wrap up, you’ve, you had to evolve Jellybooks, as you’ve said, constantly over the last 10 years as any of us growing businesses have to. What do you think is evolving right now? What do you feel is the next evolution? What are the most interesting developments you think in publishing right now that are on your radar?

Andrew Rhomberg 31:18
Oh my, there are quite a few from our point of view. Audiobook is rising and rising and rising, and we’ve started supporting audiobook data collection, snippets and similar. We’ve gone heavily into illustrated books as well. On the client side, we see a lot of consolidation both among publishers, also fewer suppliers. So the book, the industry is consolidating across every axis imaginable. At the moment, it is globalizing. Harper Collins was probably the first to say I want to be everywhere.

Andrew Rhomberg 31:54
We’re seeing that across retailing, we already saw it, Kobo, Amazon, Apple, are all global companies and publishing is starting to look more globally as opposed to the old territorial approach. Ireland, UK, South Africa, Australia, New Zealand, were one, one thing. Whereas you should say I’m looking at English language. That is gradually changing. Much more direct selling by publishers, which used to be total taboo. Now they sell more direct. Until a year ago, I would have said publishers are only focused on creating demand themselves. That was already a big leap. Now we’re also seeing potentially more direct selling, because of the pandemic.

Arthur Attwell 32:37

Andrew Rhomberg 32:38
If the pandemic has taught us anything, it’s just how resilient reading and publishing actually is. It’s come through the pandemic extremely well. As I like to joke, reading is a very socially distanced activity. At the extreme end, if we look at Inkitt, WhatsApp and others, we are seeing this tendency of thinking more about millennial users, and how to potentially build in chat functionality overlaid on books and think about how some of these things are merging. Interactive books have never taken off, or the enhance book. About 10 years ago, when the two of us got started, a company called Read Medal was around from Berlin who were all about social reading, you know, Facebook was big, rising and still had a positive aura at the time and people thought reading would be social and it wasn’t. Reading has never been really social in that sense.

Andrew Rhomberg 33:40
But with the rise of Instagram and WhatsApp, I think we are starting to see now new influences that are creeping into publishing. We’re seeing some of the first publishers rethinking, especially for the young reader, so young adult, up to 25, 30. How are they being shaped differently? How do conversations and communications around books happen? How do these things interflow? I don’t see any influence to say virtual reality of gaming into publishing. There has been no trend of that nature that has really been persistent, I think. But I think the way we interact with online is starting to creep into the young segments and some publishers, especially at the children’s and young adult book end, are starting to look as different to trade publishing as an academic or professional publisher is to trade publishing. We are seeing a bit of a cleavage there. That’s a relatively newer trend I’ve seen, the last year or two. We still haven’t ourselves fully figured it out. So that’s a new frontier about to happen. I think we always think of publishing as uniform and very, very conservative and it is conservative, but within the spectrum of publishing, you now have people who are much more willing to take risks versus others who are generally hyper-conservative. I think in recent months, we’ve seen those differences become much more pronounced.

Arthur Attwell 35:15
Yeah, just when I feel like we’re beginning to figure out this publishing business, an entirely new vista opens up, and we seem to start all over again. Andrew, thank you so much. This has been absolutely fascinating. Thank you so much for joining me.

Andrew Rhomberg 35:28
It’s been my pleasure.

Arthur Attwell 35:30
Thank you for listening. Please make sure you subscribe, and it would be such a help if you tell a friend about the show. Also, don’t forget to send us your own book-making topics and conundrums at where I’ll also post links to things we talked about today.

Arthur Attwell 35:47
How Books Are Made is supported by Electric Book Works, where my team and I make books all day, every day, in mostly-sunny Cape Town, South Africa.