Q&A;: Philip Parker, Developer of Automated Authoring Platform

Mac Slocum
April 30, 2008 | Permalink | Comments (0) | Listen

Philip Parker, founder of ICON Group International and a management science professor at Insead, has developed a patented approach to publishing that combines databases and programming with editorial management -- sometimes via humans, sometimes via computers. ICON Group produces books in 17 genres, including health care, business, reference and crosswords.

In this Q&A;, Parker discusses ICON Group's computer-driven process.

How do you identify book topics?

Based on personal and research interests, I select a genre. Once a genre is selected, I do all titles in that genre (e.g. all trade categories that are officially recognized).

Are writers, editors, or designers involved at any point?

Depends on the genre, but yes, all are relied on heavily at many stages. Health guides are written by medical professionals and hand edited. The business reports have highly edited sections, but 90 percent is computer based.

What types of sources/databases do you pull information from? Are there data sources you don't currently have access to that you think hold promise for this type of publishing?

Depends on the genre. I use the sources that are used by regular authors. For example, an economist uses well established sources to do econometrics, I use the same sources. Many companies and governments have under-utilized data sources and databases that may yield interesting genres; I have worked on the ones that I found of interest to me. I have a huge store of proprietary data. If I use a government source, this is cited, and will vary by genre (e.g. CDC for infectious disease information).

You were part of a print-on-demand (POD) panel at TOC '08. Are all Icon Group books POD? What POD service(s) do you use?

No, not all are print-on-demand. We use LSI [Lighting Source] and Booksurge for POD. We do some POD ourselves for specialized orders.

Could your company -- or a similar company -- function without POD?

Yes, in fact, most of our titles are not POD, but electronic via subscription for large libraries -- corporate and non-corporate.

Are all books also made available as ebooks? What ebook formats do you use?

Yes. PDF, DOC, Mobipocket (coming soon), Pocket PC.

Do researchers or clients ask you to prepare specific books?

Yes. We are able to do financial and labor studies on demand.

Mike Maznick says there's some fairly negative feedback on some of the titles. Is that a consequence of the automated nature of the content creation? Do you feel confident people buying these books know they're generated? Or does that not matter?

All publishers have negative and positive comments (e.g. O'Reilly). I would find it strange if our titles did not. Of the titles we have on Amazon, some 50/210,000 have real comments. Many are satirical. Of the ones from actual buyers, all publishers will receive negative and positive feedback (both can be not real, as Amazon comments are almost wiki based; posted by various people, including affiliates who are trying to sell titles).

I do not track the feedback on Amazon, but I imagine of the 17 genres (crosswords, classics, trade, outlooks, etc.), the negative ones are probably only on the health care guides, which are sold mostly to libraries and patient associations. Of all the genres, this one [health care] is not "generated by computer" -- all the text is written by professionals. The computer is used for formatting and doing the index, and compiling the glossaries.

I have a feeling that the low ratings are because the person does not like the content, thinks that better content or similar content is available elsewhere (e.g. the Internet) or was hoping for more. The health guides are clearly marked as Internet guides, and they cite Internet sources. All of the guides are vetted (by librarians, etc.). If people are dissatisfied because they think the computer wrote the text in the books, then they are dissatisfied for the wrong reason, which is unfortunate.

Many patient associations have not only reviewed the books, but also recommended them to patients and families. On balance, I think it better to make these available to patients with rare diseases who wish to better know how to navigate the Internet, beyond a Google search. For the other genres, I have never received negative feedback, only positive feedback or questions about methodology.

What is your most popular title? How many copies were sold?

Our trade reports, which are purchased by consulting firms, investment banks, and companies involved in international trade. This series is very popular. We gauge sales by series, not by individual titles. Traditional publishers think in terms of individual titles.

On average, how many copies of a single title do you sell?

There are thousands held by libraries (this is public data at World Cat). Some firms subscribe to all titles. Again, we often sell series. Some [titles] sell hundreds, some sell just a few, as a part of a series sale. The prices seen on Amazon are one-off -- we sell few or none of these.

For a typical title, what percentage of the total retail sale is profit?

We do not have a typical title. ICON Group as a whole makes no "profit" -- all resources are plowed into R&D; for new genres. The margins of the books at retail -- as opposed to profit -- are very low for the POD titles, and higher for the business titles. The margins for the low-priced products follow the industry, though we have lower margins as POD can be expensive compared to short-run printing.

A recent New York Times article says that each book costs you "about 12 cents in electricity." What other costs are involved in the process?

It can cost hundreds of thousands of dollars, or more, to set up a genre (programming, licensing, editing, research/analysis, etc.). Many [genres] take about 1 year to create, some take 3 to 5 years. I have been doing this for about 8 years now.

How is pricing determined?

Same as in the publishing industry. In some genres we try to equate marginal revenues to marginal costs. On lower-priced POD we make sure we cover the basic costs. On higher end, we try to be substantially below related titles (e.g. trade and outlook, and other business reports). The latter [higher end] are really not sold via Amazon much, but rather through MarketResearch.com, EBSCO (content inclusion), NetLibrary and traditional channels for those markets (direct sales).

How many titles do you plan to develop this year?

Depends on the genre. For Mobipocket (mobile books), we plan on about 68,000 titles. For others, maybe around 50,000. We are working heavily on my dictionary and animations.

Related:

• Video: "Patent on 'Long Tail' for automated content authorship"
• Automated Authoring Frequently Asked Questions
• Measuring Success on Self-Published Titles

Q&A; with WEbook President Sue Heilbronner

Mac Slocum
April 15, 2008 | Permalink | Comments (1) | Listen

WEbook is a new Web-based platform that blends traditional writing workshops with Web 2.0 functionality. Authors and groups can use the site to develop manuscripts, novels, screenplays and other publishable content, and if their efforts are well received, the projects are published by WEbook.

In the following Q&A;, WEbook president Sue Heilbronner offers further details on the company and its collaborative process.

How did WEbook start?

WEbook is the vision of Itai Kohavi, our founder and CEO. In addition to being a twice successful entrepreneur in the technology space, Itai is a twice-published author. When he "completed" his manuscript of his third project -- certain that it was in good shape -- he sent it to a few sharp friends for their feedback. The critiques he received were superb and comprehensive, but he realized that if he had the reactions and input throughout the writing process, he would have produced a far better written product in far less time. In addition, he would have enjoyed working together with friends and fellow writers. He looked online to see if anyone had created a cooperative publishing tool that would have met his needs. Finding nothing suitable, he conceived of WEbook, doing away with the age-old vision of the lonely author and embarking on a community-sourced content creation environment for book publishing.

Is the platform based on wikis? Blogs? Is it proprietary?

Our platform is proprietary and was built for this purpose. It is based in part on wikis, with additional focus on inline comments, inline ratings, and versioning.

How many people are currently participating in projects?

As of April 9, the day WEbook opened its public beta following the alpha, there are 750 registered users of WEbook. Many of the alpha users were recruited in to the process to help provide valuable feedback on the platform and prove that collaborative authoring works.

How are projects created and managed?

Projects are created by an instigator who has an idea for a new book, collection, story, screenplay, etc. That person, the "Project Leader," has the ability to invite others to participate in the writing, give feedback, or both. She sets exposure level and permissions for the project when she establishes it. The Project Leader is charged with managing the project, but in many cases the work can take on a life of its own, with other writers or reactors moving it ahead at a rapid pace.

Are all projects public?

No. WEbook felt it was extremely important to balance the interests of writers -- who feel very strongly about the ownership and protection of their written work -- with the wish of WEbook to create a vibrant community. WEbook allows a project leader or author to make a project private. In doing so, she can be the only member of the project or elect to invite a few friends. A "private" project can stay that way so long as it stays under 35 people (not coincidentally the size of the largest possible creative writing class). At member number 36, the project is effectively a WEbook public project. Users are made aware of this when their project hits that limit, and there are implications of this decision within the Terms of Use and the rights allocation.

Do authors maintain copyright?

The Terms of Use for WEbook required some really new thinking, as the model doesn't exist anywhere else on the Web, and we needed terms that departed in important ways from copyright law in order to make the process possible. Authors who work on private projects with fewer than 35 people retain rights throughout. Once they hit that 36th member, put their work into the public realm, or submit their work for contention as a published WEbook, WEbook takes a six-month option to publish. If WEbook does indeed publish a book, rights are transferred to WEbook, the publisher. If WEbook does not publish, rights revert to the author. That's the simple version. The WEbook Terms of Use are more detailed. We're also producing a short, snappy video to highlight key issues of this all-important topic.

What is the revenue split with authors?

Authors and substantial contributors receive a total combined royalty of 5 percent of net sales.

How are substantial contributors determined? Is it a quantifiable level (i.e. they posted x number of times)?

We use a formula to determine materiality. It has a few softer inputs beyond quantity, which constitute attempts to create a measure of quality and significance to the ultimate work. This is not fully refined, and we expect this algorithm to be a continually moving process as the site evolves. Ultimately, users will see a measure of how they stack up against the algorithm to give them motivation and transparency. We also are intrigued by the idea of giving authors marketing tools to motivate users to contribute to their projects. This might have royalty implications in the future.

Have you found certain topics that are suited for collaboration?

Our goal is to provide a platform that adapts and grows in the direction the community sets. Our first book is Pandora, a fictional thriller written by 17 authors and 17 other contributors and editors. We believe a novel is the highest challenge for collaborative writing, and we took it with the alpha community to prove the concept and learn how writers could write together in the hardest scenario -- a continuous work of fiction.

That said, we think the majority of the WEbook successes will fall into the category of non-fiction or fiction collections. Topics that resonate most with users and, presumably, with the reading public, will be those that are enhanced in coverage because they have input from a community. So, for example, we have a project on our site related to successes families have achieved at home with kids diagnosed on the Autism spectrum. On a far less serious note, there is a great collection of essays on 101 Things Every Guy Should Know How to Do and The First Year, a collection of harrowing essays from first-year teachers.

How will books be selected for publication?

The community will vote on projects that have been put into contention for publication.

How will voting be managed?

Voting will be done on the site. Ratings already occur there for in-progress works and submissions. You will need to be a member of the community to vote. The bar to join is low, joining is easy, and we feel that in order to give an important thumbs-up or down to a work, you should at least identify as a member of the community.

The community will be the overwhelming majority voice in what is selected for publishing. This makes sense, as we view the community as an ingrained base of potential buyers. That said, it would be disingenuous to say that WEbook will move ahead on 100 percent of their top selections or not move ahead on something that just missed the cut. We have strong writers on staff and in our adviser circle (mainly coming out of the alpha experience), and we intend to use them as an occasional input to ensure we're on a good quality path for constantly improving the brand and the work we bring to the market out of the site community.

What formats will books be available in?

WEbook plans to leverage the full range of existing formats -- paper books, ebooks, audiobooks -- as well as delve into new potential formats, including mobile dissemination and unique, customizable downloads.

Will the books be made available through retailers?

Pandora will be available at Amazon.com, BN.com and other Ingram-related online retailers. As WEbook establishes critical mass, a following, and a brand, we will pursue favorable distribution opportunities with brick-and-mortar retailers.

Related Stories:

• BusinessWeek: "WEbook Puts a Crowd in the Author's Seat"
• Web Worker Daily: "WEbook: Book Publishing By the Masses"
• Writing a Book with Google Docs
• Haskell Book Author Adapts to Reader Comments While Writing