How Many Hurdles Must an E-Book Clear Before It Goes on Sale in China?

The digital age is testing the Chinese publishing industry's established processes. To understand why bringing an e-book to market remains so difficult, QDaily spoke with Fu Mengruo, Kindle's technical director at Amazon China.

Do you skim Qidian novels on your phone, or would you rather read Umberto Eco on a Kindle?

Whatever the difference in “sophistication,” admit it: e-books are probably taking up more and more of your reading time.

That shift is prompting publishers to move onto digital platforms. At the first Kindle E-Book Technology Summit, held by Amazon China last week, e-book staff from major publishers packed a small conference room. The turnout caught even the host, Amazon's Kindle team, off guard.

Technical leads from publishers and the Amazon team had gathered to share what they had learned about making e-books. Publishers welcomed the exchange because the industry has little experience with e-book production and even fewer proven solutions. Missing content management systems and limits on time and money are among the problems they face. After the summit, the feedback board was covered with notes asking for more events like it—an unusual sight.

The Kindle Store launched in China in December 2012. A year earlier, Fu Mengruo, Kindle's technical director for customer technology at Amazon China, had begun leading a team that introduced Chinese publishers to tools and better ways of working. E-book publishing in China remains largely uncharted territory, and the team has continued to help publishers convert books and train staff. This is Fu's first interview with the media.

How many hurdles does an e-book face between the choice of a title and its release?

Where is the book? Can it be published?

For now, much of the publishing industry is still at the stage of turning print books into e-books. Books are not prepared for digital publication from the outset. A publisher releases a print edition, decides it might also suit a digital platform, and only then starts converting it.

Before that seemingly simple process can begin, publishers face two questions that ought not to be problems: Where is the book? And can it be published?

Both questions stem from outdated, inefficient content management systems.

First, a publisher may not know where a book's source files are. Every publisher uses a content management system to track its books to some degree, and many have dedicated resource management departments. But how well those systems are used varies widely. Poor content management can leave files incomplete or inaccurate, and earlier versions may be impossible to find when a book needs revising.

Rights are an even bigger problem. At CITIC Press, for example, the concept of the “right of communication through information networks” was still fairly new between 2005 and 2008, and contracts specified only the right to publish online. “We need to discuss with our legal department whether we can use contracts like these,” Cao Lingling, deputy director of channels at CITIC Press's E-Book Center, told QDaily.

In the early days of CITIC's e-book business, Cao said, every title “needed a description, a price and a summary before it could go on sale. We spent a huge amount of time sorting out that information and the spreadsheets.”

Without an effective way to track each book, publishers spend a great deal of time before release collecting and correcting files and related information. Much of that work could be avoided.

You can read the source files. Can a machine?

Getting the source files is only the beginning. Next, publishers must extract accurate, usable text and images. Software generally identifies the text, which can then be turned into the e-book format sold to readers. It sounds simple. In practice, it is another headache.

Publishers use different typesetting systems, so the files they need to export come in different formats and follow different standards. That makes collecting content difficult to automate and prone to errors.

“We receive files in all sorts of formats,” Fu told QDaily. “Some are PDFs you can copy text from; some are PDFs you can't copy from. Some are text files with no layout information; some are Word files. Many originate in two common typesetting programs, Founder and InDesign. Some files arrive with garbled characters or missing paragraph formatting. That makes the next conversion step very difficult. We then have to ask the publisher whether it can provide better files so we can move on with production.”

Accurate text is only the minimum requirement. Images pose another risk. “The images we get are the ones available after the print book has been finalized,” Fu said. “Print-quality images are high resolution, but for various reasons we may not be able to get the high-resolution files. We do our best to present the images well, but sometimes their quality means we have to leave them out.”

Sometimes getting the files at all is a struggle. In one case, a publisher's head office authorized Amazon to convert a book, but its branch office knew nothing about it. The reports and applications required to obtain usable files took considerable effort.

These quality problems led Amazon to add an inspection stage to its internal conversion process. It checks files from publishers for missing content, unclear images and typos to decide whether they are suitable for conversion. The check is necessary, but it takes more time.

Fu sees verification when content is collected as the part of the publishing process most in need of improvement. “It's a shared challenge for publishers, conversion companies, Amazon and other platforms,” he said.

File conversion takes considerable time and labor

Once publishers have the raw text and images, they must convert the files. Using various programs, they rearrange and package the material into the e-book readers will buy.

Many publishers are unfamiliar with conversion, so people do much of the work that could be automated. Developing a more automated process takes money and staff, posing both a technical and a strategic challenge.

At a publisher that places little emphasis on e-books, an intern might spend a day or two converting one title. A book with many tables, formulas or images can take far longer.

Publishers seeking to expand their e-book businesses face a much larger workload. They generally build in-house teams or hire conversion companies.

In-house teams have advantages in quality and speed. But until e-books become a substantial source of revenue, not every publisher can afford a dedicated team. Many outsource the work instead.

Conversion companies can be another weak link. Their methods and the quality of their work vary, and not all meet publishing standards. Outsourcing also creates a risk of file leaks, a loss publishers can ill afford.

Fu said Amazon does some conversion work for publishers free of charge because their technical resources and capacity are limited. Asked whether Amazon would eventually want publishers to do more of it themselves, he said both approaches would continue.

“We very much hope publishers will put more into their e-book businesses, both in effort and funding. At the same time, Amazon knows it needs to do this alongside them. If a book is particularly complex or a publisher is short on capacity, for example, we're happy to share some of the cost or help with the conversion.”

Print-layout experience has to be reconsidered

Layout design is part of file conversion. It attracts the most debate—and the most attention.

To design a reading experience for a different medium, publishers must rethink the arrangement of text and images, as well as the styles of headings and paragraphs.

1. Making an e-book? It's more like building a website

Take EPUB, a common e-book format. An e-book is essentially written in HTML. Making one resembles building a small website, with HTML defining its structure and styles.

E-book production is closer to web design than print publishing, and that presents technical difficulties for publishers with little experience of it.

Some features of this design also have advantages. An e-book note is a form of hyperlink: tap it to jump to the relevant passage, then return just as easily. In a print book, you have to turn pages to find it.

Fu has a view of what e-book design should do. He repeatedly stressed that “a book is, first and foremost, something to read.” He added: “Our e-books today have almost no design, or very little. We still use print books as the model. Everyone judges an e-book by the same standard: Why can't it look just like the print book? Print design suits reading a physical book. You can't simply carry its layout over to an e-book.”

2. With so many layout options, where should publishers draw the line?

At the summit, the Amazon team demonstrated several layout techniques: drop caps, notes, borders and mixed text-and-image layouts. Researching them had taken a great deal of time.

Designing and coding these effects takes time, but the more important question is how they appear on mobile reading devices. A Kindle has a six-inch grayscale screen, while mobile readers generally offer even less space and show one page at a time. Complex layouts such as images spread across two pages or text wrapped around images do not translate well to e-books.

Publishers need judgment to decide which effects to use. Choosing between a simple layout and a more elaborate one, and deciding how much detail matters, determines how much time they spend. A more carefully finished book may require more complex CSS and will certainly require more thought about the details.

3. Layout tools? None quite fits

Tools are another problem. Chinese publishers have yet to settle on software or build broad experience using it.

Fu said publishers often ask whether a single, integrated program could let them make e-books without knowing EPUB or HTML, or whether a graphical interface could make the process easy enough for anyone to follow.

“We've seen commercial software and platforms offer layout tools that don't require CSS,” he said. “But someone still has to spend the time and effort designing the styles. Once that's done, applying them can be simple—rather like using a format painter to make a top-level heading or a drop cap.” The Sigil EPUB editor, an open-source project under Google, is one example; a number of programs on the market show its influence.

For now, these tools vary widely in how hard they are to learn, what they can produce and how well they support Chinese. How much a publisher values e-books determines whether it cares about the finished result—and whether it devotes resources to learning and trying the software.

4. Is interactivity the future of e-books?

Interactivity raises another design question. In the United States, Amazon offers KEAV (Kindle Edition of Audio and Video), a tool that lets publishers embed audio and video in books. Hardware limits mean the content can be played only on Fire tablets.

As publications such as The New York Times and The Washington Post expand on mobile platforms, both have made infographics and interactive content strategic priorities. If e-books are to keep up, they will have to address the same question.

“Amazon Kindle has noticed the trend and the demand, and we are considering it,” Fu said. “There have been discussions in the United States, but there's no detailed timetable yet.”

Where is the balance between cost and quality?

Two factors constrain every part of e-book production: time and cost.

“Generally, a publisher's ideal price for converting a book is around 100 yuan,” Fu said. “Our standards for e-books are higher, so our costs are certainly much higher than that. Conversion companies offer even lower prices—some quote just over 20 yuan for a plain-text book.” The sums publishers expect to pay are startlingly small.

Much of the cost of making e-books lies in the initial investment of resources and staff. Beyond the time spent converting a book, recognizing text, turning it into XML, creating CSS styles and laying out pages may require publishers to buy several programs or even hire programmers to build tools of their own.

Meanwhile, publishers want faster turnaround. Responding more quickly helps them meet demand, beat competitors and earn more. At the summit, Cao of CITIC Press asked, “The fastest we can convert a book right now is one day. Has anyone found a way to do it faster?”

Most publishers want to balance time and cost against quality. Should a book be intricate and polished, or simple but still readable? If you want to put one on sale tomorrow, you may be able to offer only a basic reading experience. Will readers accept it?

The market is too small, and piracy is widespread

Clearing these hurdles and getting an e-book on sale is only the beginning. Marketing and the market itself are the final tests. For now, the e-book market remains small, and piracy holds it back.

Building a better process and making e-books easier to publish takes more money and staff, yet the market is not large enough to command publishers' full attention. Without that attention, the market struggles to grow. As with many new business models, each side is waiting for the other.

Cao told QDaily that CITIC had had an e-book department since 2009 and initially worked with mobile operators. “We've put 300 to 400 titles online and plan to add 1,000 this year. China's print-book market is worth about 30 billion yuan, while e-books together may be worth over 100 million yuan. Our first goal is to bring e-book sales up to one-tenth of print-book sales.”

Other publishers are waiting for the market to develop before competing, though they may miss an early opportunity.

The e-book market cannot escape piracy. Pirated copies may have helped a generation of Chinese readers develop the habit of reading e-books, but the losses they cause are more than the young market can bear. Before publishers can release print and digital editions together, they must address this major concern.

Encryption and decryption, piracy and legitimate editions, continue to follow each other. Kindle's technical protections keep improving, but new ways to defeat them keep appearing.

Apple's App Store, Netflix and the gaming platform Steam are among the more successful recent examples of services that have fought piracy. An easier, more natural way to buy is the best response—a goal Kindle and other e-book platforms have long pursued.

Releasing print and digital editions together means changing the whole process

Treating an e-book as a digital copy of a print book is the root of these production problems. Usually, the digital edition is an extra step after the print book is finished. The process was designed for print from the start, which is why making the e-book is so difficult. That one decision gives rise to the problems of rights, content management and collection, design, time and cost.

A more sensible approach would be to consider the e-book while designing, typesetting and printing the physical edition.

Fu described his ideal process: “At the stage when we choose a title, we'd think about who would read its e-book. During typesetting, we'd consider whether the layout works as an e-book. If it doesn't, we'd create a separate digital layout. If both editions then go through three rounds of editorial review and three rounds of proofreading, both will benefit. With a sound management system, the material we receive would be easy to convert. Ideally, we could release the print and digital editions at the same time.”

Fu is an idealist about e-books. As he put it, the reading experience comes first, then cost. What reaches the reader is a file of only a few megabytes, but publishing an e-book is still a long, painful process.

Judging by the summit, everyone in the book industry is trying to change that.

Cover image from Digital Trends