Thank you very much, Ann and SK, for this post. It reads like a brief history of decision-making at Duke University Press (http://www.dukeupress.edu) over the past five years.

(Apologies in advance for the length that follows. TL;DR: it’s really, really hard, but we think it’s worth it.)

We publish 120 books a year and 48 scholarly journal titles, with strong emphasis in the humanities, social sciences, and mathematics. The good news is that our customers–scholars, librarians, book authors, journal editors–align very nicely around what we publish, issue and book, chapter, and article. This is such an overwhelming positive that we have decided to integrate our journal and book content so that it is out in the world for all to find and sample and connect and use. In 2014, we put our books online on the same platform as our journals, and provided cross-content searching, browsing, access, and administration with more integration possible in the future (http://read.dukeupress.edu). We’re very happy with the decision, and we’ve received really positive feedback from the scholarly community. BUT…

The not-so-good news is that the world (or, at least, our corner of the world) is not yet built for container-less content. As researchers, scholars love the idea of “give me everything, regardless of format.” But as journal editors, they love their themed, curated, and lovingly packaged printed issues. As book authors, they are heavily invested in the self-contained and irreducible argument built across chapters at a carefully calibrated pace and presented with a show-stopping cover design.

Our market research has shown that while librarians like the idea of a combined book and journal database of content, few of them have the systems or processes in place to acquire and manage them. Resource-strapped librarians are attracted to the prospect of fewer platforms to support and administer, but their budgeting, purchasing, and cataloging practices are largely set up along format lines. In the humanities, the most popular platform for institutional purchase of ebooks is ebrary. It doesn’t host journals at all. Journal platforms with stellar reputations among journal-focused electronic resource librarians, such as HighWire and Atypon, are unknown to book-focused librarians.

And that’s just the outside the walls of the Press. We’ve encountered the challenge of a containered world in virtually every area of our internal production and distribution operations. The order and inventory systems that well support book sales do not handle subscriptions and electronic access very easily, and vice versa. The peer review and acquisition tools sets are different, the production management tool sets are different, and the approaches to managing content and assets are different. Dominant journal content standards (XML, specifically JATS) and dominant book standards (HTML5, specifically EPUB and EPUB-like formats) are not the same. Yes, you can do books with BITS, but if your market is largely the iBooks and Kindle reader, that’s a lot of expense and overhead, even if your journals division is already doing it. With virtually every publishing operation decision, we weigh which side of the house gets to be the square peg in the round hole this time.

So why are we trying to integrate our book and journal content and, to some extent, publishing operations? Because we are betting that the original, overwhelmingly positive proposition–there is a strong, focused, engaged community that writes, reads, and buys what we publish–will win out over time. We are betting that most of these challenges are outweighed by the strength, synergy, and value of our body of work. We are betting that the drive to operational efficiencies at libraries will make our combined, online offering more and more compelling. We are betting that scholars’ desires as editors and authors will come to more closely mirror their desires as researchers. But we are also betting that there will always be a significant number of people who want to read a themed issue in the original article order and who want to follow that long argument from page 1 to page 322.

As Angela says, this is not for the faint of heart.

Alix’s comment and those of David Crotty, together seem to make the most sense. One has to think about what is on the horizon since changing the publishing world is glacial and incremental. There are several pieces in motion outside of the pub/perish world:

a) The cost of research is increasing and the funding is becoming less causing more institutions, globally, to seek collaboration across disciplines and more links outside of The Academy

b) Research shows that, at least, thru 2030, the demand for qualified post secondary graduates shows that the need will be about 1:2:7 (masters+: bachelors: skill and experienced) Hence the shift in academic hires at the post secondary level for faculty

c) “Deep Learning” and similar software (eg that used by Google, Amazon, Facebook, etc and now available to searchers in all sector) will search and anticipate needs in all areas, can parse down to the words in a document, and will soon be able to interpret in the form of documents, stripping out the persiflage which entangles the world of research publishing. As long as the knowledge is digital, it can be scanned and interpreted based on a users need and avoids the human having to read all the documents, regardless of whether it is in monograph or journal formats. It also creates a lot of meta data linking relevant (the operational term) materials where accessible to the engine.

With the above:

a) The production and use of information will change and the needed knowledge can be deconstructed, analyzed and re-assembled without researchers having to laboriously toil thru the literature to tease out relevance. How machine accessible systems will impact the business model of publishers has yet to be addressed fully with regards to where the material lies and where the marketable value may be

b) Impact factors, so dear to the promotion/tenure academics and the publishers’ key to use and thus profitability will need to be carefully weighed

c) The continued proliferation of books and journals as opposed to providing other ways to validate value (to the producers (authors and publishers) and the end users (regardless of affiliation) may need to be carefully examined given the ability to “slice/dice” and reassemble across the knowledge creating/using environment.

d) Basically, there lies the capability, if all materials were assembled into a stack, for Machine Intelligence to be able to sort thru, send out documents for human review/validation and then assemble to meet the needs of the end user.

Publishers are slowly creeping up on this, as most of the above comments note and which we can see as the major houses are moving in this direction, adding some software that approaches the issues but is not disruptive to the current model. Unfortunately, the rapid development of deep learning systems, particularly by those in all sectors who need critical access, may take the system out of the hands of the creators and publishers as materials are extracted and reassembled outside of the system. The economics of The Academy (teaching and research) and the publishing industry may increasingly finding themselves talking to each other which will probably become fiscally problematic.

Our experience here at University of California Press echoes those of others, notably Allison above – perhaps not surprisingly as we’re both university presses serving a similar audience. Over the past 4 years we have largely integrated our book and journal publishing divisions – thought Robert’s preference for “integrated and separate” probably best describes the actual structure. This change has delivered enormous and diverse benefits, some organizational and some more strategic:

1. Cost and efficiency: Our book and journal divisions were totally separate – not just publishing staff, but also finance, IT and building. We have delivered very significant efficiencies and cost reductions through integration.

2. Better use of talent: we had really talented staff buried in our journals division who are now making important contributions at a press-wide level. And we’ve been able to provide career paths for people we might otherwise have lost.

3. More informed management decision-making: in the past many of the senior management team understood little about our journals business, but we’re now making decisions for the future of the Press that are informed by an understanding of each part of our business.

4. A more informed and flexible team of staff: we’ve become more of a matrix organization, and many of our staff regularly work in cross-functional teams on specific projects which has increased their knowledge – and often motivation.

That said, we’ve learned some lessons:

– The internal change – both process and culture – can be painful, and required a great deal of our attention over a period of time to be successful.

– Integration alone doesn’t suddenly create new market and revenue opportunities. Rather, it’s the foundation on which we build through through a deeper understanding of our customers’ and users needs as we move forward (as noted my many above).

– Understanding what customers want isn’t as easy as it sounds. As a traditional book and journal publisher, we’ve had to develop and hire a whole new range of skills in project management and iterative product development.

– Much of our traditional market is in the humanities and social sciences which are, as Allison notes, still some way behind in their understanding and acceptance of new forms of publishing, especially digital. In these fields at least, the printed monograph as gold standard in tenure/promotion remains a core challenge and although we’re trying to push boundaries, our ambitions are tempered somewhat by where our audience is.

That said, we’re still convinced (and our evidence to date supports) that this is the right long-term play in strategic terms.

Remarkably, though several commenters come from university presses, no one has yet mentioned Project Muse. When i came to Penn State Press as its new director in 1989, the Press was publishing about 30 books a year, mostly in the humanities, and 8 journals. All of the journals were in fields in which the Press also published books. It was evident that there were some real marketing efficiencies to be achieved through this kind of integration. The Press’s new books could be advertised in the Press’s journals, and the journals could be marketed at the same scholarly conventions where the Press had book exhibits. I continued this approach for that very reason, and though we added a few journals along the way, they were still connected tightly to our book editorial acquisition program. In 1995 our journals manager, Sue Lewis, was hired by Johns Hopkins to work with Michael Jensen to design and launch Project Muse, with the support of a Mellon grant. Being keenly interested in this experiment in e-journal publishing and connected to it via our former employee, Penn State Press in 2000 became the first press to have its journals accepted into Project Muse, which up until then published only its own journals through the platform. Joining Muse was the only way in which a small press like ours could afford to go digital. Our print subscriptions had been steadily eroding and we knew we had to make the transition, but could not do it on our own. Our journals were key to the Press’s financial survival because they generated a surplus that we used to subsidize the money-losing monograph publishing we did. It was crucial to maintain that revenue stream, and Project Muse allowed us to do so, eventually generating about two-thirds of our overall journal revenues by the time I retired in mid-2009. I also foresaw that Muse could help us get into ebook publishing, and as a member of the Muse advisory board, I kept urging JHU Press to add books to its aggregation, which it finally did, also following my recommendation that the books and journals not be siloed on the site but cross-searchable. This seems to have worked very well, but perhaps someone from Hopkins can provide more details. Anyway, it solved our Press’s problem for going digital with books just as it had for journals.Whether this will ultimately be sustainable as a publishing model over the long haul remains to be seen, but it was our salvation at the time. My successor as director at Penn State Press went a different route and tripled the size of the journals list, not concerned about keeping it editorially connected to the book program. I can certainly understand the financial considerations behind this different approach, but it did lead to some separation, contentwise, between the two programs. I think this is a matter of strategic preference, and I was comfortable with my approach as I am sure he is with his.

With respect to chapter level metadata, isn’t one work around to work backwards from the index (if present), with perhaps a modicum of editing? [I’ve suggested this in the past but not seen it in action.] Restriction for platforms is that it works best in terms of consistency at a book level though may become tricky across a series / platform. That’s where aggregate indexes can come into use, perhaps with some data/text mining tools to create further links. Only worries then are overdoing the links (useful vs. non-useful) or the semantics (meaningful vs. non-meaningful)… Starts to make you wonder where to draw the line and how to measure value/ROI.