We cannot verify your location
Browse Book and Journal Content on Project MUSE

Download PDF

Creating a Consolidated Online Catalogue for the University Press Community

This essay derives from a feasibility study into the possibility of creating a consolidated online catalogue for university presses and was underwritten by the Andrew W. Mellon Foundation and sponsored by The Monterey Institute for Technology and Education; the essay is an abridgement of the report submitted to Mellon and MITE. For more information on the background, see the acknowledgments and disclosures at the end of this paper.

The aim of this study was to identify a means by which university presses could sell more books and, specifically, whether the creation of a consolidated online catalogue of press titles would help in this regard. The presses at this time vary considerably in terms of size, resources, and familiarity with online marketing, though all of them are active to some degree online and most identify Amazon as their largest customer. Despite this activity, however, and even for the very largest presses, a press-wide catalogue would augment sales by exploiting greater scale, enabling more effective search-engine and other online marketing, by opening up new promotional vehicles, and by strengthening individual press brands by bringing more robust technology to each institution's efforts.


university press, scholarly publishing, direct marketing, metadata, ONIX, search engine optimization, Google

'Everybody has [electronic] copies of our books except for us.'

—Comment by a university press director


Since it was first conceived five years ago as part of a consulting project for a not-for-profit organization operating in the area of scholarly [End Page 385] communications, ScholarsCatalog (originally Academilogue) has had the same aim: to help academic publishers, and university presses in particular, sell more books. Much has changed in the publishing environment since that time, of course, and ScholarsCatalog has evolved with those changes; it will continue to evolve, as publishing is nothing if not dynamic. It is worth remembering that today's investments may be tomorrow's legacy systems, potentially weighing down future generations of innovation. ScholarsCatalog's 'cloud computing' architecture is a hedge against such situations, as it outsources the burden of keeping up with new technology onto the service provider, GiantChair, creator and manager of the ScholarsCatalog service.

An important change to the original plan for ScholarsCatalog is that e-books, initially something that was mostly on the horizon, are now an established and rapidly growing part of the book business. ScholarsCatalog will be hosting e-books at launch, just as GiantChair now hosts e-books for its other clients and services, including the French model for ScholarsCatalog, Le Comptoir des presses d'universités.1 ScholarsCatalog already hosts Google Previews and will participate in the Google Editions program as soon as that program launches. PDF and MP3 are already supported, and other e-book formats will be added to the roster soon. Thus, any press participating in ScholarsCatalog will enter the e-book business rapidly and without having to develop much technical expertise.

One area that has not progressed markedly over the years is the quality and availability of book metadata. This is a complicated and exasperating problem, the implications of which may not be fully understood by many publishers. In the online world, a book's metadata are its principal marketing support, and incomplete or poorly formatted metadata lead to low search-engine rankings and lost sales. As part of the development of ScholarsCatalog, GiantChair has been working with a number of publishers to put their metadata into the ONIX format, creating automated tools to streamline the process. ScholarsCatalog and all publishers would benefit from participation in the Book Industry Study Group's ONIX certification program.2

Publishers working with ScholarsCatalog have recommended a great number of additional features. We are getting close to imposing a 'feature freeze' so that we can set a date for opening ScholarsCatalog to the public. But most of the suggestions are very good ones, and we don't want to leave all of them to Version 2.0. We are working now, for [End Page 386] example, on making the GiantChair 'BackOffice' metadata management tool available to more publishers; identifying ways to strengthen individual publishers' branding on the site; and probing the issues surrounding fulfilment of print book orders for titles from more than one publisher. It appears doubtful that ScholarsCatalog will ever be 'finished'; Internet marketing is a work in progress, and ScholarsCatalog will be adapting to the new circumstances of the online environment as they appear.

I. Background and Scope

This project has its roots in a series of discussions concerning the fate of university press publishing. The press world is under a great deal of strain at this time, strain that is worsening under the collapse of the economy. The presses' problems are being addressed by many people in a variety of ways, the most common being to identify ways to rein in expenses. This project is taking a different tack: rather than seeking a means to reduce costs or scale back programs (both of which may be appropriate actions in some instances), the aim of this report is to identify ways for the presses to sell more books. Since the presses almost always make a gross profit on every copy of every book sold (typically $0.58 on every sales dollar, an extrapolation from AAUP data), selling more copies would generate a greater contribution to the presses' overhead. In effect, the goal of this project is to identify a means by which the presses can market their way out of their financial problems.

The presses are involved in a number of activities, the chief of which is publishing books. Among the other activities are the publishing of journals and the provision of various services (e.g., warehousing) for third parties. Other activities are also beginning to appear; at least one press, for example, is exploring the possibility of publishing scientific data sets. The scope of this project, however, is books alone.

Press books themselves come in a variety of packages: different formats (hardcover, paperback); different media (print or electronic); and different marketing offerings (books sold on a stand-alone basis or books as part of digital aggregations). The primary focus here is the printed book, though much of the discussion applies (or will come to apply) to e-books and, to some extent, to aggregations. The reason for this focus is simply that that is what the presses sell today, and that is what most customers prefer. Of course, this is now changing, and as those changes occur, an online catalogue will evolve with them. [End Page 387]

The Association of American University Presses (AAUP) lists 126 member presses in its Directory 2008: Information on 126 Scholarly Presses in the U.S., Canada, and Overseas. These presses include publishers formally linked to parent universities (e.g., Harvard University Press, The University of California Press), presses headquartered outside the U.S. (e.g., Oxford University Press or OUP), and scholarly publishers not affiliated with a university (e.g., The Minnesota Historical Society, The Jewish Publication Society). For purposes of this project, this is a useful typology, which I will return to below. Beyond these not-for-profit (NFP) academic book publishers are commercial firms (e.g., Palgrave/Macmillan) whose publications are of interest to the research community and often individual titles (e.g., a political memoir by Henry Kissinger) published by trade houses (e.g., Random House). Taken together, all these publications, from a university press monograph at the centre to a stray book from Random House on the perimeter, constitute the relevant field for an online catalogue for academic books, but it was only a subset of the AAUP's extended membership that participated in this study.

The questions put to the interviewees essentially came down to these three:

  • • What are your press's current marketing practices, especially insofar as online marketing is concerned? How do other entities (e.g., vendors) fit into these practices?

  • • What is on your wish list for marketing? If you had extra money, how would you spend it?

  • • What are your thoughts about the utility of a comprehensive online catalogue for academic books? Would you participate in such a venture were it to be launched?

The organization of this memorandum pretty much tracks these three questions.

I do wish to emphasize that this project began with a hypothesis (A consolidated online catalogue would help to sell more academic books) and all the research was conducted to test that hypothesis. Someone simply determined to expose more data about the press world would have asked different questions. Indeed, there are many such questions that I was tempted to probe while working on this project, especially with regard to the publishing programs themselves (what books are the presses [End Page 388] publishing and why), but the focus here has been narrow: On the basis of what can be determined about the presses, is there a case to recommend an investment to build a consolidated online catalogue?

I have coined the term 'Academilogue,' a compression of 'academic' and 'catalogue,' to use as the name of the prospective online catalogue. This name is a placeholder. No one likes it, but there was a need to call it something. (A prototype site is now under construction, with the name ScholarsCatalog.)

II. Statistical Snapshot

How big is the university press sector? At what scale would an online catalogue have to be built to accommodate all the press titles?

The statistics on the book industry in general are not very reliable. This is because the industry's sheer diversity and the highly specialized character of some book programs make it hard to track down all the participants. The figures for the university press world are, if anything, better than the average, in large part because the AAUP has put a great deal of effort into this in recent years. But nonetheless, there are problems with the data and also some exceptional situations that must be kept in mind.

The largest of the exceptional situations is the scope of the programs at OUP and Cambridge, which skew all the figures. In the interviews OUP reported publishing 2500 new books each year, Cambridge 1500. Between these two publishers alone the total number of titles in print is about 46,000. And here we encounter our first statistical conundrum, in that the total for OUP and Cambridge cited in the AAUP directory is only about 34,000, a discrepancy I will not try to reconcile. OUP and Cambridge are also profitable, OUP immensely so, and are significant contributors to the finances of their parent institutions, and thus the case to 'help' them is much harder to make. The programs of these presses are also diverse and go far beyond the scope of this report. For example, OUP's ESL program is a money machine; and were they to be auctioned off, the journals programs of both of these presses would attract many bidders from the commercial sector.

Putting OUP and Cambridge to the side, the press world is fragmented and diverse. The six largest American presses all cluster around $20 million in books sales (Yale is a bit larger); a reasonable estimate is that each ships around 1 million units a year at an average price received [End Page 389] of about $20. The status and fortunes of these presses are very different from presses that publish, say, 50-100 titles each year, a size too small to permit a press to make meaningful investments in business development and technology.

Overall, the presses publish approximately 12,000 new titles each year. I have tried to derive the total number of press titles in print from the AAUP directory, which gives us a rough idea but is not definitive. Adding up all the presses formally attached to a university, but excluding OUP and Cambridge, yields a total figure of just under 90,000 titles. The problems with this estimate are that there are university presses that don't belong to the AAUP, we don't know if the presses all report figures the same way (I suspect that they don't), and we don't know if by 'in print' a press includes POD titles (I would). Nevertheless, 90,000 is a useful working number for press titles in print. And if we add in OUP and Cambridge, an estimate of around 125,000 titles in print is defensible.

When we get to the academic publishers that are not formally attached to a university, the figures are much less useful. The AAUP's affiliates cite figures that, when added together, come to just over 10,000 titles in print. This does not include the numbers for the RAND Corporation, which lists 18,000 titles, a figure too high to be for books alone, but which may include other media and white papers. The bigger problem with non-aligned academic publishers is that very few of them belong to the AAUP and hence their output is not listed in the directory. As one press director remarked to me, Washington, DC, is swarming with not-for-profit publishers of books of intellectual merit. If we could get them all into the online catalogue, the number of titles would be much larger. And this is before we begin to count the academic titles from commercial publishers, who may be invited to participate on a title-by-title basis.

So our working figure is 125,000 press titles in print—and growing. This is a very significant number in that the total number of titles in Amazon's catalogue is believed to be around 2.5 million, and Ingram, the leading wholesaler, actively warehouses about 1.25 million titles. (An Ingram executive told me that the total number of books in print is probably 1.5 million. One wonders how Amazon derives its figures.) Barnes & Noble keeps just over 1 million titles in inventory, with a declared policy of entering into its systems at least one copy of any [End Page 390] book a publisher submits to it. (Who would not want to be part of B&N's distribution system?) However small the overall sales figures of the presses may be, in terms of title output and their impact on the overall infrastructure of the publishing industry, the presses constitute a sizable segment.

Parenthetically, it should be noted that it would be very handy to be able to say what percentage of all new books are published by university presses (that is, 12,000 new titles is what percentage of the total?). The problem is that estimates of new book output are all over the place. Ten years ago the number of new books published in the US was thought by some observers to be about 60,000 (probably based on the number of registered ISBNs and Bowker's Books in Print). Some analysts still quote figures in that range today, but there are also estimates that reach to 250,000 titles a year and more. I suspect the truth is somewhere in the middle. The figure of 60,000 probably understated the total by overlooking all the niche publishers that did not bother to register their books with Books in Print. The larger number almost certainly includes multiple editions of the same title (hardcover and paperback, British and American editions of the same book, etc.) and a huge and growing number of books that are self-published, an outgrowth of low-cost digital publishing technologies and the 'Long Tail' marketing capability of Amazon.

It is therefore probably not unreasonable to say that university presses constitute 5-10 per cent of the total new book output in the US, provided that we don't push at that figure too hard. That figure is important when one considers what a comprehensive online press catalogue would look like. And when we add backlist to the mix, the presses' role looms even larger, since the presses have a much deeper commitment to keeping books in print than do commercial publishers.

When we move to revenues, the presses' share of the total book industry is, of course, much smaller. In 2007 the AAUP surveyed its membership and concluded that the total book revenues of the 63 participating presses were $277 million; these figures did not include OUP and Cambridge, but presumably did include all of the large American presses. Extrapolating these figures to allow for the presses that did not participate in the survey, a reasonable estimate for American press book revenue is around $310 million. Adding OUP and Cambridge [End Page 391] in, the total in 2007 came to $530 million. That's out of an industry of about $30 billion.

On the bottom line the participating presses in 2007 reported parental subsidies of about 8 per cent of net revenue. They also reported receiving just over 4 per cent of net revenue in philanthropic support and income from endowments. With these various subventions added to revenue, the participating presses generated a surplus of 3 per cent. Allowing for extrapolations, back-of-the-envelope adjustments, and so forth, it's not unfair to say that the presses 'lose' about 10 per cent or $31 million a year on operations (at least in 2007), without factoring in the large surpluses at OUP and Cambridge.

I suspect that the subsidy is larger than that, however, in that sometimes support for a press sneaks in without acknowledgment on the income statement. For example, some presses get warehouse space from their parents at no charge; others do not have to pay interest on working capital. Or there might be an income statement where the plant cost (the cost for creating the master copy, which includes typesetting, copyediting, etc.) has been reduced by a grant from the author's university; a press might use this money to offset plant rather than show the subvention 'below the line' in order to make the numbers look good. One truly bizarre situation is at an institution that puts all the cost of employees' medical benefits on a separate, university-wide account, removing this cost from the press's income statement (and all departmental budgets). This is a terrible business decision in that it corrupts decisions concerning insourcing and outsourcing certain tasks (e.g., copy-editing) by making in-house staff seem less expensive. Consider the arithmetic: a press staff of twenty-five people, medical insurance of $10,000/employee, an unacknowledged subsidy of $250,000. For this particular institution's press, the hidden subsidy comes to around 7 per cent of revenue.

I would therefore gross up the estimate of the total subsidy to the presses' book programs to around $35 million a year (leaving OUP and Cambridge out), a bit more than 10 per cent of net sales. To put this in perspective: the ARL lists eighteen libraries with total budgets of this size or larger—but we should not have to choose between operating a library and running the entire university press system. The subsidy for presses goes toward a sizeable slice of the total US book output each year, and an even larger piece of the academic book market. For all the gnashing of teeth about the woes of the presses, the conclusion seems [End Page 392] inescapable to me that the university press segment provides the most efficient use of capital of any part of the world of scholarly communications today.

III. Results of the Survey of University Presses

Data for this project was gathered by distributing a questionnaire to the presses. In most instances the presses filled out the questionnaires in advance of phone interviews. The aim of the questionnaire was to explore what capabilities the presses would bring to a shared online marketing project. As expected, the presses' resources varied considerably in this regard, with one press, for example, noting that it had several people on staff whose job was to perfect the metadata that was sent to marketing partners, while another press had no expertise in the ONIX standard and relied on sending out information in Excel spreadsheets, with the hope that the marketing partners would transform this information judiciously. Simply averaging all these responses together does not tell us much about the ability to launch a new service.

In this section I will summarize in a highly compressed form what the presses said in response to the questionnaire. There is, of course, no 'average' press.

I should note that there was no attempt to survey editorial programs. The point of this project was not to persuade presses to adjust their editorial programs so that they would publish books that were more saleable, but to help the presses sell more copies of books that they already publish.

A. Production Issues

By 'production' I mean all stages of the publishing process that follow the acquisition step (which includes revision in response to peer review) and result in the master copy of the text, which is then used for manufacturing (which consists of printing and binding). The steps for production include copy-editing, composition, and production editing (e.g., trafficking proofs). Sometimes this process is called 'prepress,' an increasingly misleading term in the Internet age. All the costs for production are typically gathered together on the income statement and are distinguished from the variable costs for manufacturing.

The production process is packed with digital technology, but the technology is not always integrated. For some presses the production [End Page 393] process yields a text in a format that can then be used for either print or online distribution, but for many presses an online text is part of a post-production process—that is, after the master for book printing is created, there is another step to make the text Web-ready. For some older books, the Web-ready version may literally follow manufacturing: a book is printed (or was printed) and warehoused and only later is picked up and put through the digitization process. The bulk of the Google book-digitization project is of this (the post-production) kind.

There does not seem to be a widely used software platform for managing the workflow of the production process for books—in contrast to journals, where a number of rivals compete (e.g., ScholarOne, Editorial Manager, BenchPress).3 There are several reasons that the workflow is, for most presses, not fully integrated. First, there are a number of applications that are simply entrenched in the marketplace (e.g., Microsoft Word, InDesign, Acrobat/PDF), and any total solution would have to work around these mainstays. Second, books are simply harder to produce than journals—more steps, more back-and-forth with authors, fewer standardized formats—and a workflow process would have to accommodate this complexity. Third, on the marketing side there has been far less pressure to publish in an electronic format, thus mandating that a print workflow remains in place.

There were few surprises in the presses' responses to the questions on production. In recent years digitization has penetrated the process further—so, for example, to the question, 'What percentage of your books are in some kind of digital format?' a typical response was, 'All books since 2000.' For some presses the number of titles in digital format was around 25 per cent, but the bulk of the presses reported figures closer to 50 per cent of all books in print. A few presses have 100 per cent of their books in print in some digital form.

Far and away the leading format is PDF. A small number of presses are putting files into XML and many presses are eyeing this seriously. Only a small number of publishers have complete end-to-end XML workflows (this is true for commercial publishers, too). The predominance of PDF is significant in that a file in PDF requires another transformation before it can be fully exploited on the Web (for example, for search-engine optimization and for becoming the centrepiece of online community-building), though there are vendors (e.g., Olive Software) that provide these conversion services. With the growing interest in e-reading [End Page 394] devices (e.g., the Amazon Kindle), PDFs are likely to require another conversion before becoming useful (no press that I spoke to has budgeted for this).

All of the vendors cited by the presses have become mainstream suppliers, with Ingram's Lightening Source at the top of the list. Other vendors include IBT, CodeMantra, and BiblioVault. I was surprised not to hear the name of R.R. Donnelley, which has a huge prepress division. All the presses have some kind of POD arrangement in place. (It is astonishing how quickly POD has gone from an exotic capability to a routine affair.) Some presses were reluctant to share vendors' names for competitive reasons. Most of the presses expressed a desire to assert more control over the management of their digital content, but only one reported a program to do so.

Book production is not the same as catalogue production, of course. Most of the presses have created a marketing database, where information on titles is stored, later to be output for catalogues, whether in print or online. PDF is the preferred format, though many presses have some or all of their catalogues online in HTML. In some instances a press's online catalogue is a combination of HTML (for newer titles) and PDF.

Most of the presses (the figure I heard a couple of times was 70 per cent, but I doubt anybody has actually tallied this) send the digital files for catalogues to a third party, which then processes the data into the ONIX format and sends it on to various marketing partners. Some presses do more, some less, in assisting in the creation of metadata for online marketing partners. It is noteworthy that the presses that reported the greatest emphasis on developing this metadata also reported the most success in selling books through Amazon. One form of this, for example, is that Amazon would buy directly from the publisher instead of from a wholesaler, thereby improving the publisher's margin.

With regard to creating a comprehensive online catalogue (Academilogue), many of the presses are going to need assistance to get their information into a form suitable for online presentation. For catalogue metadata the presses will either have to work with third parties (as some of them already are) or they will have to have the new catalogue service assist in creating the metadata for them. Thus part of the expense structure will be for the service itself, part for the process of preparing information for the service. [End Page 395]

If we think more ambitiously about getting the full texts of the books themselves online, the problems and costs mount. Books (perhaps half of the total) not already in digital form would have to be digitized—a considerable expense. Books that are in a suboptimal format (e.g., PDF) would have to be converted. For new books a complete XML workflow should be implemented, though it should be clear that some presses think that this is not worth the money. None of this is easy or inexpensive, and in some instances (books published many years ago that remain in print, but for which there is little ongoing demand) it is difficult to argue the case for making an effort.

One intriguing possibility is to use the published API (application program interface) for Google Book Search (GBS) to display the text online. Any publisher that is comfortable with the GBS terms of service could then display a searchable text online as though it were on the publisher's site or at http://academilogue.org. (The data would still reside at Google, but it would appear to reside at the publisher or with Academilogue.) I have seen this capability demonstrated, and it is impressive—and inexpensive. Writing to the GBS API requires more technical capability than most presses have, of course, which strengthens the case for a consolidated service.

B. Marketing

(1) Online Activity

If there is a single principal finding for this project, it is that online bookselling now comprises 25 per cent of total press sales volume (measured in dollars), up from zero ten years ago. Of course, not all presses report a figure of exactly 25 per cent; the range begins as low as 10 per cent and goes over 30 per cent, but taken together, the figure of 25 per cent is pretty sound. During the ten-year period that press sales began to migrate online, overall sales, when adjusted for inflation, have been essentially flat. So, zero to one-fourth of the business in ten years in a flat market. The presses enjoy such large sales online because their titles are quintessentially 'Long Tail' products. For other publishing segments, the online market share is generally lower (probably 10 per cent for trade books, less for new college texts).

The phrase 'online bookselling' is misleading in one respect, and that is that almost all of the online sales are through Amazon. (The press comments about Amazon were so extensive that I have put Amazon [End Page 396] into its own section below.) Amazon sources press titles directly from the presses and through wholesalers (principally Ingram and Baker & Taylor). An intriguing and unexpected finding was that some of the presses are beginning to see meaningful sales from their own Web sites. One press, with total book sales just under $20 million, reported that sales from its own Web site were 'only a few hundred thousand dollars.' Only? Another publisher told me that the sales from his company's Web site were large enough to motivate him to study the situation. He found (not surprisingly) that every book on display from his Web site also appeared on Amazon, and in virtually every case, the Amazon price was lower (he factored in shipping and handling). Since the presses earn a higher margin on books sold from their own sites, there is a niche marketing opportunity here. The key is having a well-designed Web site.

Perhaps the most disappointing finding of the survey is how few presses were analysing traffic to their own Web sites. A small number of presses could immediately quote usage figures, but a more common response was to say that they don't have the means to evaluate Web traffic or that they have the data, but don't have the resources to study it. (For those presses that studied usage, the tool of choice is Google Analytics.) Indeed, some presses seemed unaware that studying Web usage was something that a publisher might want to do.

The situation with analysing Web traffic gets curiouser and curiouser. All the presses have Web sites; some of them are fairly sophisticated. Many of the presses have started blogs, and a few have initiated podcasts. But what is the traffic for the blogs, who reads them, and what are the presses attempting to accomplish with them? If a podcast falls in the forest and there is no one there to listen to it. . . . There were many times that the presses' involvement with trendy new media tools seemed to be taking place in the absence of any kind of marketing plan.

One interesting item that emerged during the course of this study was the news that Harvard University Press planned to create videos for its authors. Scattered comments from the other presses were of the variety of, 'Well, Harvard would do that'—implying that online video was something that only a particularly well-endowed press could consider. Here I wonder if some of the presses (and I emphasize the word 'some,' as there are presses, even small ones, that are thinking innovatively about digital media for marketing purposes) are too easily intimidated. One could imagine a program by which a student intern shoots [End Page 397] five-minute Web videos of authors talking about their books. The videos are uploaded to YouTube, with transcripts for better search-engine discovery and a pointer to the publisher's Web site. A more technically advanced press could, in addition, put the video on its own site; and an even more sophisticated press could publish the video's API and invite users to syndicate the video across the Web. A Hollywood budget is not required to do this. This is the kind of thing (and there are so many with digital media) where it may be better simply to do it and see how it works (including measuring the outcome) than to plan it carefully in advance.

The survey made clear that for most of the presses, more aggressive online marketing was limited by constraints on resources—always allowing for the exceptions of the larger presses and a few small presses with particularly 'geeky' directors. This became apparent in the responses to the questions on the staffing of marketing departments. A smallish press might report a marketing department of three to five people. In some instances one individual was responsible for managing the Web site and other aspects of online marketing, but the most common response was that every member of the marketing department worked on both bricks-and-mortar and online marketing. For virtually all the presses, online marketing was an activity that had been layered onto traditional marketing efforts. (A typical response: online marketing has 'about one-half FTE' or 'one FTE.') Thus the presses' marketing resources, never robust to begin with, are now being stretched across multiple media.

This finding (that there rarely is a large amount of dedicated headcount for online activity) overturns one hypothesis concerning the benefits of a consolidated online marketing service. It had been my hope that Academilogue would permit the presses (the smaller ones, at any rate) to outsource a large portion of their online marketing, reducing internal headcount and thus making the development of Academilogue in some respects cost-neutral. But you can't fire a half-person. And if online marketing activity is integrated with traditional marketing activity, it is hard to see how the presses would grapple with restructuring in order to realize any cost savings. Thus if Academilogue is to succeed, it must do so on the top line—increased revenue—not by rationalizing the cost structure. [End Page 398]

A related comment from some of the presses was to the effect that 'We would never give up our own Web site' (not that they would be asked to). The concern was loss of control and also a fear that the funding for a consolidated online service would disappear, leaving a press without an online marketing arm. This is entirely based on a misunderstanding of what Academilogue is setting out to do, but it points to the need for greater explanation of what the service would be, what its aims are, and how it would fit into a press's current online activity. To bring the presses along on this point will require an investment in graphic design for mock-ups and perhaps even a limited online demo site. Academilogue, that is, must be a marketer of its own services to its customers, the university presses.

One footnote on online marketing: the academic publishers surveyed that are not formal university presses (e.g., OECD Press, Island Press, The Jewish Publication Society) seemed to be doing more online than the university presses themselves. This has to be classified as a general impression and not a definitive conclusion because the sample was small, though one press director told me that she believed that the niche or vertical markets of these publishers played a large role in the development of their digital plans. It may be that these specialized academic publishers more rapidly migrated to the Web because of the natural constituencies that gave rise to them (e.g., in the case of Island Press, people interested in environmental science). The university presses themselves don't have a natural community of users. The presses are somewhat marginalized on their own campuses (though people will debate the degree to which this is true) and the broad base of their publishing programs (typically ten or more areas) scatters their attention over a wide area. The Web is a direct-marketing medium and is most effective when a specific segment of the population can be targeted.

(2) The Special Case of Amazon

Amazon is the single largest element in university press publishing today. To recap some of the 'statistical snapshot' in Part II, based on the information provided by the presses, online booksellers (principally Amazon) now comprise about 25 per cent of total dollar volume for the presses. That figure includes estimates of the sales that the presses make to wholesalers (primarily Ingram and Baker & Taylor) that are then shipped to Amazon (and Amazon then sells these books to its [End Page 399] customers). The significance of this figure (that is, the 25 per cent of total volume) is that ten years ago, Amazon's sales of press titles were literally zero. During that decade, the total press sales, when adjusted for inflation, probably have not grown at all. Thus Amazon has gone from 0 to 25 per cent in ten years in a flat market.

It is difficult to benchmark these figures against the book industry as a whole, as Amazon is notoriously unforthcoming about its operations. Industry sources estimate that trade book publishers sell a smaller percentage of their books than do academic publishers through online booksellers, perhaps something in the 10- to 15-per-cent range. Amazon's share for new college texts is lower than that (the network of college bookstores is still robust, though declining in part from competition with Amazon). A fair generalization would be to say that 'Long Tail' titles sell better online than elsewhere, and academic titles are certainly on the Long Tail. Amazon is thus a potent force to the book industry as a whole, but its overall impact is greatest in the segment where university presses operate. The unfortunate irony is that while Amazon is very important to the presses, the presses represent only a small portion of Amazon's total business.

Although Amazon is best known as a consumer bookseller, it also has programs for schools and libraries. In the K-12 supplemental-book market it competes with the Follett wholesale company (and some publishers' direct sales). In public libraries it competes with Baker & Taylor and Ingram. In academic libraries it competes with Baker & Taylor (Yankee Book Peddler is owned by B&T), Ingram/Coutts, and Blackwell. In college texts it competes with college bookstores. There is no breakdown available of where Amazon's sales ultimately wind up, so a press can only guess where that 25 per cent of total press volume sold to Amazon goes. It is probable that most of those sales end up with individuals, but some find their way to academic libraries, especially the smaller ones.

I spoke with a number of librarians (and previewed the question on the liblicense mailgroup) about Amazon's role in libraries. It is fair to say that the largest research libraries rarely if ever buy anything from Amazon, but smaller libraries purchase books from Amazon when they need something right away (Amazon's fulfilment is believed to be the fastest in the industry) or when they are shopping by price and are [End Page 400] willing to give up some of the special services (e.g., library cataloguing data) provided by the traditional library wholesalers.

Inasmuch as the presses publish books that are used in college classrooms, Amazon's role in college texts is important as well. The text market is 'formal' (books surveying a field, typically for underclassmen, published for the most part by the industry giants: Pearson, Cengage, McGraw Hill, John Wiley, and Von Holtzbrinck/Macmillan) and 'informal' (books that find their way into the curriculum typically after being published with other markets in mind). A good example of informal texts, which literally sell millions of copies overall, are literary classics. Few of the presses publish what can be called formal texts (MIT probably does the most of any press), but many press titles end up in college classes, typically as moderately priced paperbacks. Press titles also often appear in library reserve rooms as supplemental reading.4

What is Amazon's role here? For library reserve rooms, the situation is the same as for library sales in general: little Amazon activity with the largest libraries, an unspecified amount of activity with smaller libraries. For books used in classrooms, however, Amazon is playing an increasingly large role. Amazon discounts titles as much as it can and thus puts pressure on traditional college bookstores. And of course it is significant that college texts are purchased by students, though adopted by instructors, and students freely shop around the Internet for the best price. Thus some portion of the presses' sales to Amazon end up in college classrooms.

Much to publishers' consternation, Amazon has also become a large purveyor of used books, though Amazon is hardly the only one. A search on Amazon is likely to result in a page that says that 'new and used' copies are available, for prices beginning at a low figure. This is a benefit to consumers and impecunious students, but it also increases the burden on the publishers, who derive no income from used-book sales. Since the university press sector operates at an average deficit of 10 per cent of revenue, there is no way to underwrite this alleged benefit to consumers.

While figures vary from press to press, it is clear that Amazon is leading to some gains in sales (mostly by market-shifting), some losses in sales (through used books), and also to margin erosion. Margin erosion occurs in various ways; among the most common for the presses is when a publisher sells a book at a college discount (20 per cent) to college bookstores, but for roughly 46 per cent to retailers, including Amazon. [End Page 401] Amazon discounts that title to customers, and students buy it online, bypassing the campus bookstore. Thus a publisher could literally sell the same number of copies of a book, but experience a 26-per-cent decline per copy in income. I won't work through the thicket of publishers' discount schedules here, but it is safe to conclude that Amazon is putting downward pressure on the presses' economics even as Amazon sells more and more press books. It is not far-fetched to anticipate a time when the college discount disappears entirely, costing the presses literally millions in income.

All the presses report that Amazon is their biggest or second biggest customer. (Baker & Taylor is the most frequently cited rival.) Amazon has achieved this position because it is exceedingly good at what it does. Amazon was not the first e-commerce site, but it took e-commerce to a level where it truly was competitive with first-rate bricks-and-mortar retailers. (It is an odd fact that Amazon started out in books, which are an almost negligible retail category. Amazon, of course, has grown far beyond books and even, through its platform services, beyond retail.) Amazon created an Affiliates program, which syndicated the Amazon storefront across the entire World Wide Web. Amazon initiated the idea of community tools for the book industry and presents user-generated reviews next to many books. Amazon took a strong position in used books and is now a leader in print on demand. The discovery tools (including Look Inside [the book]) are the best in the industry. Search-engine optimization is outstanding: Virtually any Google search for a book results in a link to Amazon at the top of the page. This list goes on and on. At times during the interviews with press personnel, it was hard to move the conversation beyond Amazon: All Amazon All the Time.

Power corrupts. Almost without exception, the presses are uncomfortable with Amazon's growing strength, in part because Amazon is beginning to throw its weight around. This takes various forms. For one, Amazon is putting increased pressure on the presses for deeper discounts. A second point: Amazon now looks to the presses for cooperative advertising support, which affects the way books appear on the Amazon site. Amazon is also piloting a semi-automated service called Vendor Central, which, to put it mildly, the presses detest. Vendor Central is designed to be the only way a publisher can communicate with Amazon. A publisher posts a question or comment on a message [End Page 402] board and waits for an Amazon representative to respond, if Amazon deigns to respond. It is no longer possible to pick up the phone or send an e-mail to contact a buyer.

Anyone experienced with retail channels will not be surprised by Amazon's evolution, from 'nice guy' and innovator to a 'breaking heads' Wal-Mart. Arrogance is a function of muscle. A fair statement to make about Amazon is that as it has grown, it has become more and more like its peers with regard to its business practices. Thus, in some respects grumbling about Amazon must be put into context: Amazon is simply like other accounts, but larger.

There are other issues with Amazon, however. Publishers send metadata to Amazon for the Amazon online catalogue, but Amazon controls what actually appears on the site, and Amazon, some presses report, is not always quick about making corrections. The most obvious missing or hard-to-find element is the publisher's name (Amazon does not want customers to click to the publisher's site). Oddly, except for those titles that use the Look Inside (the book) feature, it is often a challenge to find information about the books on the site, though there may be reviews, sometimes user created. Would it be so hard to display prominently a 250-word abstract, created by the publisher, before all those references to how to buy used books? Why must a user dig through so many invitations to purchase other books before coming to a description? The publisher is cut off from the actual merchandising of the product. Certainly no publisher would have used books and new displayed side by side.

Thus, Amazon is essential—and imperious. Depending on how the question is asked, publishers give different answers about the desirability of creating an online catalogue to offset Amazon's dominance. 'Would you like to see an online university press catalogue?' Answer: 'Amazon is doing that already.' Alternative question: 'Would you like to see a service that provides alternatives to Amazon?' Answer: 'Yes. Amazon is getting too powerful.' The metaphor of Finlandization occurred to me as I was probing this topic.

If academic publishers controlled Amazon, there would be little reason to create a new online catalogue. It is precisely because academic publishers have little influence on Amazon that a new online catalogue is desirable. [End Page 403]

(3) Google

Google is the hardest item to discuss in the context of press marketing. Google is vast and dynamic; it truly is the outstanding organization of the decade. I do not pretend to understand all of the implications of some of Google's initiatives, and to judge from the number of conferences, white papers, and blog posts on the recent proposed legal settlement between Google and the various publishing entities that were suing it, I am not alone. It is fair to assert that Google is, with Amazon, one of the two most important aspects of press marketing today and that it is likely to become even more potent in the coming years. On some items the presses spoke about Google with virtual unanimity; on others there was considerable disagreement. No one I spoke to expressed confidence in a particular vision of what the online marketing landscape would look like in a few years as a result of the Google legal settlement.

Although every staff member of every press knows that Google has a large suite of services, the three that occupy most of the presses' attention are Google Scholar, Google Book Search, and Google itself (that is, the consumer search engine found on the home page5). Google Scholar did not arise very often in the interviews, as it focuses on journals. The Google search engine was in the background throughout ('You can look it up on Google'). GBS is another matter, however, as it specifically targets book publishers.

GBS itself is not one thing, but a suite of services. The original GBS, which most presses participated in, digitized press titles and made them searchable online. Participating presses could earn advertising income when users searched on their book pages. No one in the survey deemed this program a success. Not surprisingly, much of the content of academic publishers is simply not the kind that will entice advertisers. Income from this program was termed 'none,' 'negligible,' or 'unimportant.'

More significant is the exposure GBS gives to books. Every press likes this. But here we run into the problem of not staying on top of the usage logs. Some presses reported that they studied the logs carefully; one commented that there were clear connections between the Web traffic that GBS delivered to the press's door and sales; but most presses either did not track the GBS usage or stated that sales attributable to GBS were negligible or unknown.

As in the discussion of Amazon, the unavoidable question is, If GBS is such a potent marketing force, where are the incremental sales? If press [End Page 404] sales are flat and GBS is helping to sell books, then the sales attributable to GBS could only be coming from other, increasingly disadvantaged sales channels. It is entirely possible that GBS, like Amazon, is simply shifting the business around but not enlarging it. Another way of looking at this is to say that without GBS and Amazon, press sales would have declined; and this, I believe, is true.

Speaking with my publisher's hat on, I would have preferred that GBS not include links to library copies of books, but based on my own experience using GBS, I am doubtful that the library links have cost publishers many sales. This is also the opinion of all the press directors who took up the question. That situation may change, however, as more and more presses, with and without Google's own proposed marketing programs, make the full texts of books available to academic libraries through digital aggregations. None of the presses I interviewed seems aware of this possible development. Someone could search on Google/GBS, find a title of interest, and then follow a link to a library, where the book could be sampled or read in its entirety. Sales of digital books to libraries, in other words, are likely to lead to reduced sales overall.

On the proposed Google-publisher legal settlement, the comments of the presses at the time of the interviews didn't seem uniform or conclusive. What no one knows is what it means for this to be a settlement between two parties and not a legal ruling. One wonders if Project Muse would have started on its ambitious book-aggregation project had the management known that the GBS agreement was in the offing. At a minimum (assuming the courts approve the settlement and that many publishers choose to participate in the new GBS aggregations), Google is going to establish a touchstone for all other aggregations with regard to technical features, terms of sale, and pricing. I suspect that some of the urgency at many presses at this time to create digital aggregations is motivated in part to avoid coming under Google's umbrella.

(4) Other Marketing Partners

The presses all work with multiple marketing partners, though some are more important than others. Besides Amazon and Google, Baker & Taylor and Ingram play the largest roles in the presses' fortunes. Presses with large trade programs also do significant business with Barnes & Noble. [End Page 405]

All the presses have long-term arrangements with both B&T and Ingram. Historically, B&T was the leading means to get press titles into libraries, though Blackwell and some smaller wholesalers were and are still significant players. The importance of B&T has declined along with the reduction of library purchases of press titles. B&T recently announced an arrangement with eBrary to provide a hosting platform for digital aggregations of books, but at this time no press has indicated that it is participating in the B&T program. B&T, of course, also services Amazon and, to a small extent, physical bookstores, and thus is important to the presses in that regard.

Ingram is a more complex case. Historically, Ingram was the presses' primary avenue to bookstores. Ingram continues to play that role today, though the market share has shifted and Amazon is now one of those bookstore accounts. With the advent of Ingram's Lightening Source service and the acquisition of Coutts (with its My iLibrary software platform), however, Ingram has diversified its business, becoming the leading provider of digital asset management and POD services for book publishers (competing with BiblioVault) and now a competitor to B&T in the library market.

The survey did not turn up any new or unexpected information about either Ingram or B&T. All the presses push metadata to these partners, who use it to build their online catalogues. (These catalogues are B2B, used for orders from bookstores and libraries, with no consumer face.) The one new item uncovered is for a catalogue under development that is designed to bypass both Ingram and B&T. This is being sponsored by the American Booksellers Association and is being built by a company called Above the Treeline. This new service will facilitate direct ordering by independent bookstores from publishers. The cost of participating in the ABA program may be too great for the presses to bear, however. Since independent bookstores are not a significant market segment for the presses, this service is not likely to be a high priority for them.

(5) Print Marketing

Press marketing is not limited to the online sphere, of course. In addition to the usual public-relations activity and management of trade accounts, the presses all create print catalogues, which are widely distributed. A typical press has two seasonal catalogues a year and a number of smaller, subject-specific catalogues. These catalogues are mailed to [End Page 406] bookstores, wholesalers, scholars, and librarians. A representative press may spend $50,000 per year on these activities, though some of the larger presses spend a great deal more. Almost no press now creates a comprehensive print catalogue of all books in print; if such a catalogue exists, it is now fully digital and appears either on a press's own Web site or as part of the more comprehensive catalogues of some of a press's trading partners (e.g., Baker & Taylor). The trend for print catalogues is to print fewer, but no press wants to eliminate them and most presses would like to create more subject-specific catalogues and special catalogues targeted for particular events (e.g., a catalogue created to accompany a visiting scholar's open lecture delivered on campus).

The print catalogues themselves rest upon well-established infrastructure of mailing-list rentals and databases of market information. The presses typically rent lists, mail catalogues, and tabulate results. This is important: insofar as the mailings permit some degree of order tracking, the presses can determine what catalogues warrant investment and where funds would be better used elsewhere. Some mailings, of course, are bound to be both untrackable and wasteful. I didn't survey enough librarians to be able to make this point conclusively, but I suspect that many if not most librarians consign press print catalogues to the trash bin upon receipt.

Could Academilogue help the presses with their current catalogues? Yes, I believe it can, and this is true whether the catalogue is online, in print, or for special occasions. If Professor Jones is delivering a guest lecture on campus on the origins of the American Civil War, the local university press could go to Academilogue and use its tools to create a Web-based catalogue tailored for the event; the catalogue could include books from multiple publishers, with the host press earning a tariff on books sold. The press could also download catalogue material and use it for POD: fifty copies of a special catalogue to be distributed to all attendees of Professor Jones's lecture. The American Management Association may provide a helpful example here. The AMA publishing arm sells books from tables at the back of the room where AMA seminars are held. Those tables are now the AMA's biggest book-publishing sales channel. Creating special-event catalogues would not reduce the press's $50,000 catalogue cost—it would add to it—but it would augment sales. [End Page 407]

C. Wish List

What would a press do if it had more resources? What is on the wish list? There were some surprising answers to these questions.

But, first, what was not surprising is that the presses almost universally would like to beef up their digital marketing efforts. They would add to their Web sites, include more blogs and podcasts, and perhaps hire someone to work in a more dedicated fashion on metadata. Some presses would also like to digitize more of their backlists (though one press director questioned this, noting that just about everything that had any potential of finding a market had already been digitized). They would like to have a complete XML workflow, and those that do not have ONIX capability (a minority) would like to get it.

The fascinating thing about these items is that they are not necessarily tied to an expectation that they would yield greater sales—and why would anyone increase marketing expenditures if it were not believed that sales would improve? So, to the question, 'How many more copies of books would you sell if you did this or that?' the answer was usually silence on the telephone. What would we get for a $100,000 investment in the Web site? A better image. And what is a better image for if not to sell more books? Again, silence on the phone.

Part of the reason for the presses' reticence about forecasting a return on a new marketing investment no doubt comes from simply not knowing the answer to the question. But it is likely that another factor is that many of the presses think of some marketing activities not so much as a way to sell books but as a means to impress and attract authors. Given the nature and structure of the book business, this is an eminently rational point of view. There are many, many books for a reader to choose from; therefore readers choose the best books, which they identify with the best authors. For a press to sell more books, it makes sense to try to attract the finest authors. Editorial decisions, in other words, are in some respects aspects of marketing strategy.

What we can call 'editorial marketing' presents both opportunities and problems for a consolidated catalogue. The first problem is that it is hard to develop meaningful metrics for a service when the desired outcome is to 'have authors think good things about us.' In the absence of such metrics, getting presses to pay for the use of the service is a challenge—whereas getting a publisher to pay $10,000 for a service that measurably delivers an incremental $100,000 in sales is easy. A more [End Page 408] insidious problem is that a shared catalogue (or, more to the point, a shared catalogue infrastructure) can help to make all authors and all presses look equally good. Some of the more distinguished presses may balk at helping the smaller presses get access to the same tools to attract authors.

The opportunity lies in the fact that the presses don't only compete with each other but also with other academic publishers and, for that matter, any claim on the attention and discretionary spending of the target audience. Adding such features as author pages, author blogs, and so on could make the press segment as a whole more competitive.

On the wish list for a small number of presses (but well articulated, suggesting that the topic has been actively studied for some time) is identifying means to increase sales in emerging markets. The two such markets most often cited were China and India. (Based on my own activities, Eastern Europe may be the bigger near-term opportunity.) Insofar as Academilogue would increase discovery of press titles—and, as an Internet service, Academilogue is inherently global—the proposed service could help to address this.

Also on the wish list was the creation of a means to sell digital aggregations of books to academic libraries. (This is discussed further in the section on libraries below.) A number of presses have started to work on such projects, often citing the example of Oxford Scholarship Online. Project Muse and perhaps JSTOR are also preparing offerings in this area. I believe there is indeed a modest opportunity for the presses for programs of this kind (modest because library budgets can stretch only so far and Google is bound to take a share of the market). The virtue of Academilogue with regard to such aggregations is that it will increase the amount of attention paid to specific titles included in such aggregations.

The unexpected items on the wish list all had to do with direct marketing, both print and electronic. Among the items:

  • • More money to mail more catalogues

  • • Access to mailing lists for catalogues that did not have to be rented for each use (e.g., a shared resource controlled by the AAUP)

  • • A well-scrubbed set of e-mail lists for particular disciplines—also not requiring use fees

Putting aside the utopian fantasy of no use fees (if the data get updated, there will be additional costs, thereby requiring ongoing fees), what's [End Page 409] interesting about this list is that it represents a clear problem of lack of capital for investment. (The term of art in the commercial world is 'pure balance sheet problem,' and such problems are regarded as silly—because it is assumed that you can always raise capital, though in the current economic climate, this is not necessarily true.) That is, the presses see an opportunity ('If we mail more catalogues to carefully targeted customers, we will get more sales, enough to more than pay for the increased mailing costs'), but lack the capital to pursue it. This is different from, 'We would like to expand our list in medieval literature, but we can't find a donor to subsidize it.' Allowing for the very big differences between the NFP and commercial worlds, a lack of investment capital can often find a remedy by crafting a credible business plan for sources of capital. So to the wish list I would add one unspoken item: a resource for financial modelling.

IV. Libraries

The research for this particular section of the report was supplemented by interviews with librarians and members of my professional network.

The role—the diminishing role—of academic libraries in the fortunes of university presses has been much discussed. The question for the presses today is, what, precisely, is the economic relationship with libraries and how might that arrangement be improved? I am not investigating here other arrangements with libraries—for example, the provision of IT services or drawing on library collections for editorial projects—but focusing entirely on the library as press customer.

The presses sell books to libraries through wholesalers: Baker & Taylor, Ingram, and Blackwell. Amazon has also entered this market, but the size of its library segment has not been made public. The preferred way to sell books to libraries is through standing-order and approval plans such as those managed by the Yankee Book Peddler division of Baker & Taylor. A standing-order plan, for example, might consist of a library's profile ('Send us all books that support our research program in Continental philosophy, but do not include titles from these publishers [list follows]'), and a set of terms (pricing, etc.). These plans have dropped off considerably over the years (an artefact of 'the serials crisis'), but are still in use by the large ARL institutions. Other library sales result from libraries' review of wholesaler and publisher's marketing materials, [End Page 410] including print and online catalogues. Books that are needed quickly, perhaps in response to a faculty request, may be ordered from Amazon.

Libraries now account for about 25 per cent of the presses' total sales, measured in dollars. The percentage would be lower if measured in units because libraries often buy books with more expensive bindings, though the trend is for even libraries to opt for less expensive formats. This was a hard number to derive from my research, and some press directors would argue strongly that the figure is lower, though most press staff are comfortable with expressing this as a range: 15-25 per cent. For some presses, the percentage is higher, sometimes much higher. With the exceptions of the two largest presses (OUP and Cambridge), generally speaking the larger presses report lower library sales than the smaller presses. It is the structure of this market segment, where sales go almost entirely through wholesalers, that makes it hard to get precise figures, since the wholesalers sell to multiple segments.

Sales to libraries continue to drop, and in the current economic climate, things could get very bad. It is not known, however, what the figures were like in 'the good old days,' say, twenty years ago. No one has sales records that go that far back, but there is anecdotal support for the estimate that library sales were once 40-50 per cent of press volume. Could they have been higher? I suspect they were, but there is no evidence one way or the other. It is clear that the number of bricks-and-mortar stores that stocked short-discount books (that is, books sold at a discount of around 32 per cent of list price instead of about 46 per cent for trade titles) has fallen off considerably, from around 750 twenty years ago to under 100 today—victims of online competition. But professional books stores are not libraries, even if they sometimes stocked books with library bindings.

So, without having firm evidence of how strong academic library sales once were, we have reasonable, useful information on how strong they are now. We also know that the trend line is not favourable, but there may be some offsetting good news buried in Amazon's internal sales reports.

How could the presses sell more books to libraries? Even if we accept the hypothesis of this report, that increased discovery will lead to greater sales, there are at least two areas where even the most robust marketing is not likely to have any effect. The first of these is with the larger ARL institutions, which already collect much of what the presses publish [End Page 411] every year through standing-order and approval plans. We could hardly expect that these libraries would begin to collect books in areas where there is no support from research faculty. Thus the larger ARLs do not represent a market for potential growth for the presses.

The second area where the growth opportunity is negligible is in highly specialized fields where the current marketing venues already saturate the market. One press shared an anecdote with me about an author of a linguistics monograph, who appeared on the Charlie Rose public-television show. Despite the exposure, the sales of that author's book did not rise at all. That press's view is that their current marketing activity already reaches all the prospective purchasers of that book, libraries and individual scholars alike, and no amount of augmented online marketing is going to boost sales.

The press with the Charlie Rose author appearance has a point (and a number of presses made the same point in less dramatic ways), but will that point hold forever? That press has saturated the market for linguistics monographs with outstanding marketing in existing channels. Professors in the field receive mailings; they advise librarians on what to purchase; they see reviews in linguistics journals; they may receive e-mail alerts of new titles in their field; they attend conferences where all the new publications are discussed; and so on. Times change, however, and marketing venues with them. Five years from now the environmentally unsound mailings may decline in number or disappear altogether; librarians may rely more on algorithmically generated online recommendations; the linguistics journals will increasingly focus on their online editions and be susceptible to new online marketing services, including enhanced search-engine discovery; e-mail alerts may have given way to RSS feeds, which are influenced by the many manifestations of Web 2.0 marketing; conferences may become Webinars; and so on. The question of what constitutes market saturation ultimately hinges as much on foresight on the future structure of the marketplace as it does on a hard-nosed analysis of the number of prospects for the purchase of highly specialized literature. Enhanced online marketing may be necessary just to maintain a market, if not to increase its overall size.

Where many of the presses believe they will find growth in the library sector is in the sale of digital aggregations of books. I agree with this, with some hedging and caveats. But let's summarize the opportunity first. [End Page 412]

The idea of digital aggregations is built largely on a JSTOR-like marketing plan. A large collection of material is put together, typically assembled by subject area. The material is in digital form and made searchable (and the forms of search keep expanding—e.g., the advent of Google Scholar). Every year new material is added to each collection. The collections are marketed directly to libraries, though sometimes intermediaries and consortia get involved. Sales are on a subscription basis, with the publisher having the obligation of maintaining a reliable data centre. Terms and pricing vary, but within a fairly narrow range. (New issues keep arising, of course. For example, what is the publisher's policy on preservation?) Oxford Scholarship Online is an example of a successful program in this area.

Part of the motivation for developing programs like these is that it changes the way librarians think about book purchases. Currently librarians purchase books on a one-time 'firm' basis—that is, once purchased the library owns the copy outright and the press has no further obligations to the customer concerning that title. With digital aggregations, the subscription model becomes operative (libraries may or may not have access to the titles permanently and publishers have an obligation to service the collection on an ongoing basis). Aggregations can thus be sold as serials and draw on a different library budget line, not necessarily a literally different 'pool' of money, but funds that a librarian thinks about differently. The benefit of this for the presses (the benefit for libraries is arguable) is that the ongoing nature of a subscription make it less likely that the sale will be cancelled in a coming year. A digital aggregation thus may become part of a library's 'standing' or ongoing budget, like subscriptions for STM journals.

There are many challenges for the presses in putting together a program like this, not least being that few have a sufficiently large list of titles in any one subject area. With a few exceptions, it is likely that presses wishing to pursue this strategy will band together to form large collections organized by subject. (Duke has just launched such a program by itself, but Duke's list is fairly narrowly focused; thus a Duke aggregation is something of a proxy for a topical aggregation.) Assuming the presses are available to deal with the considerable technical and operational issues, they still must contend with the following: [End Page 413]

  1. 1. As noted above, 75 per cent of press sales go to individuals. It is probable that most of these individuals are affiliated with universities. Thus a digital collection, coupled with remote access, may serve to cannibalize sales to individuals. This is precisely what happened to many society journals when libraries began offering remote access to digital collections; many of the society members dropped their print subscriptions in favour of the electronic (and free) edition. Thus growth in library digital aggregation sales may be offset to some degree by a decline in print individual sales. This creates a challenging situation for setting prices.

  2. 2. Where are the libraries going to find the money to pay for these aggregations? It may be that the first presses that offer such collections will be successful, but subsequent offers may be rebuffed by budget-strapped librarians.

  3. 3. The competition in this area is likely to be keen, and it may arise quickly. It is probable that in addition to collections prepared by clusters of presses, we will see new initiatives by such organizations as Ebrary, NetLibrary, ALPSP, and JSTOR. And hovering in the background is Google's entry into this market.

There is another dimension to selling to libraries that warrants discussion; I would characterize it as an 'elephant in the room' matter. The widespread belief is that libraries would buy more press titles if they had the money and that the presses have been squeezed out of library budgets by soaring prices of STM journals. I am sure there is some degree of truth to this, but what concerns me is the anecdotal reports that press titles circulate poorly. I do not have definitive information on this topic, but I have heard remarks like '40 per cent of university press books never circulate' and '25 per cent of all the books in our collection have not circulated once in the past ten years.' If this is true, what is the likelihood that libraries will support digital aggregations?

It would be useful to study the actual circulation of press titles, but in the absence of such a study, we may infer that there is limited circulation from reports that libraries are increasingly contemplating moving to comprehensive just-in-time or on-demand book purchases. One scenario for this is that a library would cancel all its standing-order plans. Then a catalogue of some kind would be made available on the library's resource site. Authorized members of the library's community would [End Page 414] go to the site and request purchases. Subject to guidelines on purchasing ('Your department has overspent its allotment this month'), books would be ordered from a wholesaler and entered into the library's collection. Consider a major research library's purchase of 10,000 new press titles every year on a just-in-case basis and then reflect on how that figure will be adjusted if that institution moves to on-demand. One likely unintended consequence is that fewer books will be purchased, fewer press titles will get published, and those titles that do get published will carry higher prices.

If the presses begin to migrate their library business to online aggregations, they will find their books and collections evaluated pretty much as journals are—that is, by various usage metrics. This means that the presses will have to become increasingly sophisticated about online marketing, a difficult task for the smaller presses. It also means that a comprehensive online catalogue of academic titles could be useful as one component of a broad suite of marketing services in maintaining and even building presses' library sales.

V. Utility of a Consolidated Online Catalogue

Despite the fact that all the presses have some degree of online marketing activity and almost all cite Amazon as their largest customer, in my view the bulk of the presses continue to underestimate the flexibility and potential of the online medium. There is a widely held view that the Internet is a static thing, that once you mount a Web site, you have 'figured out the Web.' One example of this was the comment that a combined press catalogue was tried ten years ago; it didn't work then, so why should it work now? I tracked that project down; it was started by Michael Jensen, Chuck Creesey, and Bruce Barton (and may have been funded by Mellon).6 In interviews with Jensen and Creesey I learned that that earlier project was literally pre-Web (it was built on Gopher) and was handicapped by the absence of an efficient means for the presses to transmit data (this was also pre-ONIX). Both Jensen and Creesey are eager to help out with a new catalogue project. The changing nature of the Internet is an invitation to rethink our marketing aspirations. (See Appendix A for the text of a blog post that previews some of these ideas in the context of POD.)

A consolidated catalogue will introduce scale to Web marketing, which is a very important thing for search-engine marketing (bigger sites [End Page 415] get a higher ranking in Google, though the world of search-engine optimization is Byzantine and ever-changing). Thus titles listed in the catalogue will more likely be discovered by users; and one assumes that greater discovery will lead to more sales. (Note exceptions to this in the section on libraries above.) A consolidated catalogue will also introduce scale to technical development, as a larger organization will be able to recruit top-notch engineering talent (or inspire that recruitment by a vendor) and distribute their work to the press community at large. And a consolidated catalogue will increase leverage with vendors: it is one thing for Duke University Press to negotiate with eBrary (as it did, and from what I have heard, apparently with great success), but it would be another matter entirely if eBrary were to be approached by over 100 academic publishers. Indeed, the scale of the operation could entice other vendors to enter the field at various points along the value chain.

What is the benchmark for the service's utility? As one press director astutely put it, the challenge for the service is that it has to add something to Amazon and Google. If the service can provide an offset to Amazon and Google, that's a good thing in its own right—because all publishers want multiple avenues to the ultimate customers. If the service adds sales for the presses, or enables sales at a higher margin, that would make virtually all the presses want to participate. Thus the metrics: sales channel diversification, increased volume, enhanced margin, and lower costs (through outsourcing). Any combination of these items is a good thing.

VI. The Evolution of the Idea

The initial idea for a consolidated online catalogue changed as this project developed. In part this was because of changes in the environment (e.g., commentary on the Google-publisher legal settlement, the explosive growth of certain e-book readers such as the Stanza for the iPhone), and in part this was in response to the many comments and suggestions made by the interviewees. So, for example, at the start of the research, the assumption was that an online catalogue would increase sales, lower individual presses' internal IT costs, and reduce print catalogue costs. I now see little short-term opportunity to reduce presses' IT costs, since there are so few personnel dedicated to online marketing to begin with. And no press so much as nibbled at the idea of creating print catalogues [End Page 416] with tools provided by Academilogue. I will summarize a revised feature set below, but first let's review some of the comments from the presses.

A key issue is that many of the presses simply want a project that does more. They want not only a catalogue, but full online transaction capability as well. They want the ability to create e-books. They want the ability to develop and market aggregations of e-books to libraries. And they want a resource that assists them in targeting undergraduate instructors for course adoptions. There is great pent-up demand for doing more things digitally. The problem the presses have is that they lack the resources to create these things on their own, and sometimes those absent resources include hard-headed strategic thinking.

On the issue of stand-alone e-books, I am sceptical about the presses' involvement at this time and do not see this as a useful activity for Academilogue. E-books are under 1 per cent of the book business today. In any event, e-book publishing is a problem that must and will be solved, but other publishers have a bigger stake in finding a solution. For the presses this is an opportunity to work with other people's money: Have Random House and Simon & Schuster create a consensus around e-book technologies and conventions, and then the presses can join the parade. Having said that, insofar as Academilogue increases demand for press titles, e-books would benefit.

For digital aggregations the opportunity is clearer and shorter-term. Libraries are purchasing such collections now. Creating and marketing such aggregations, however, is a very different matter from managing the marketing services that Academilogue proposes to provide; it will require different infrastructure from that of an online catalogue. But once again, if Academilogue increases discovery, it should increase demand, and thus even digital book aggregations should benefit.

The most interesting idea to come out of the press comments was to think of Academilogue as a platform, that is, a baseline of content onto which other services (including more content) could be layered. Once we have a catalogue in place, we could add a list of all college courses, with their instructors' names, schedule, and title selections. Presses would then mine this information and make recommendations to instructors as to appropriate press titles for required and recommended reading. The course list would also assist presses in identifying trends for new acquisitions. This layered service, in other words, would provide the kind of information about actual classroom adoptions that [End Page 417] at this time only the five leading commercial textbook publishes have access to in their proprietary databases. (There are vendors that provide some portion of this information, but these vendors do not have a faculty-facing service, and in any event, their listings are incomplete.) For this example, Academilogue provides only a piece of the solution, but it is an essential piece, as it is not possible to map books to courses unless there is a list of all the books.

Finally, many of the presses want to sell books, or sell more books, from their own Web sites in order to improve their margins. I resisted this idea until I began to hear reports of successful direct Web sales. I don't think anybody believes Academilogue could become a rival to Amazon, but the possibility of creating transaction capability now seems to me to have merit. There are complex issues here (the technology of e-commerce, the integration of fulfilment systems), but there also are benefits (e.g., Academilogue could help finance its operations in part by taking a percentage of sales, though less than Amazon does).

A press-wide online catalogue thus could evolve into a part of the infrastructure of the academy, assisting the presses themselves in strengthening their financial condition, enabling scholars and librarians to discover and purchase books, and helping teachers identify the best books for use in the classroom.

VII. Snapshot of the Feature Set

This is not the place to design a Web service, but it is useful to consider what features such a service should aim to provide.

A. Basic Architecture

The design philosophy should be 'build for the market, scale to the stars.' It is important that the feature set not get out too far in front of the market. If RSS feeds are common (as they are today), Academilogue should have them. If 3-D displays are the bleeding edge of computing, Academilogue should not bleed with them. New features should be considered by studying actual user activity—and here we have the whole issue of analysing user logs once again. For the initial release of the service, the feature set should be the minimum required to test the marketing proposition, and that is that the service should make press books discoverable online. [End Page 418]

But scaling the system is not an afterthought. Over time new features will be added, and the architecture of the system must support that. The system should be modular, and it should be technically possible to publish APIs (whether, for business reasons, the APIs are released or not is another matter). The underlying architecture must be able to accommodate the long-term goals of the service, which could include full e-commerce capability. This means that the design team must be sophisticated. Academilogue is more than a blog or a simple Web site.

B. Creating the Feature Set

For the first release the service should do the obvious and simply attempt to copy the de facto standard, a typical catalogue entry on Amazon. There will be some obvious differences (no used books, much less commercial clutter, no user-generated or unmoderated reviews, and, of course, the prominent display of the publisher's name), but we should not reinvent the wheel; Amazon is very good at what it does. For subsequent releases other elements can be added. I would like, for example, to see links to a network of author bibliographies, which in turn link back into Academilogue, and I would like to see new fields of information of interest only to people of an academic bent ('The publication of this book was made possible through the generous support of the Mellon/MacArthur/Sloan Foundations,' etc.). But we should resist the temptation to play 'this would be cool': A Web site should not be the sum of the whims or tastes of the people who create them, but the carefully articulated response to actual and forecast usage patterns. The era of 'mad genius' marketing is over; in the Internet age creativity is the work of a cold-blooded data analyst.

The site will definitely require internal search capability (why not Google?), but it should also display titles by topic (American history, French literature, etc.), which is how most people look for books. This is a response to the most frequent criticism of the project: Nobody looks for books just because they are from university presses. This is true: They look for things by area of interest. A well-designed service, however, can create new, composite areas of interest, as areas of study evolve and occasionally cross over into other fields. And the service could provide a page for topical categories created on the fly, as when a prominent scholar is scheduled to give a guest lecture at a university or perhaps [End Page 419] before a governmental audience ('Click here to see a list of the books cited in Professor Jones's bibliography').

While it may not necessarily be a feature of the initial release, the service should also aspire to connect press titles to current affairs. The AAUP's 'Books for Understanding' program7 is a good example of something like this. It would be improved, however, if the listings went beyond author and title and included summaries. It would also be desirable to have academic bloggers to comment on topical matters, with links to specific books—thus providing even more content to be indexed by search engines and drive up Web traffic. Academilogue, in other words, should be made to play the Web 2.0 game, but with a distinct academic flavour.

Unlike almost all marketing partners today, Academilogue would allow the presses themselves to take full control of the content for their titles. The service would provide a series of templates for each press to fill in (title, author, abstract, etc.), and the press and only the press would be able to edit the content. (The inability to affect the way their titles are presented on Amazon was the single most common complaint I heard during the interviews.) There are some important workflow matters to be sifted here: Does a press create content directly for Academilogue? Does a press push metadata to a third party, which in turn sends the material to Academilogue and other marketing partners? Does Academilogue take on the responsibility of pushing metadata to the other partners? All of this is going to have to be analysed carefully, with the likely result being, at least at first, a mixed workflow—not entirely efficient, but accommodating the needs of big and small presses alike.

If Academilogue does not have full transaction capability in its earliest releases, it will nonetheless have to point users to where they can purchase books. Each press must be able to make decisions for its own publications. One press may choose to put an Amazon button on the page (thus earning a commission from Amazon), another may direct traffic to its own warehouse or Ingram or elsewhere. There is an inherent problem with this, in that any user who wants to purchase books from more than one press may not be able to combine the books for shipping. There are workarounds for this (e.g., a special contract with B&T for fulfilment of any order for two or more titles), but it is probably not possible to match Amazon point for point. [End Page 420]

What titles to include? Ideally, the site would include all in-print university press titles. I would also like to see it include all out-of-print titles (assuming rights have been cleared), whether through POD or even digitization on demand. It is worth noting that since Academilogue is merely a catalogue, it does not have the challenge that Google does of creating a digital inventory for every title. Indeed, Academilogue can use the GBS API and display the text of a book on the Academilogue site, even as the data physically resides on the Google data centre.

But the service should go beyond university press titles to include books by other academic publishers, for-profits and not-for-profits alike. The goal is to create a single destination on the Web for all academic books, and for this to happen, critical mass is essential. Thus, the total catalogue will be smaller than Amazon's but greater than the sum of the titles published by members of the AAUP.

This raises the question of management and monitoring. For such things as determining what presses are invited to join and which are not, there will have to be an advisory committee, the equivalent of an editorial board. There will also have to be advisory boards for such things as proposing feature sets, establishing workflow, and so on. The management of Academilogue will have to work closely with these groups, and at the same time it will have to be solicitous of the concerns of its participating presses.

A reasonable summary of the initial feature set would thus likely look something like this:

  • • A comprehensive catalogue of all participating university presses

  • • Participation by invitation of other academic publishers or specific titles from other publishers (e.g., an important political biography published by Random House)

  • • A site that has been fully optimized for search engines and tagging and indexing services of all kinds

  • • Individual accounts for each publisher to manage its titles

  • • Templates for posting content (to create a somewhat uniform presentation)

  • • Links, at the publishers' option, for ordering, further information, etc.

  • • A trade service (that is, for bookstores and wholesalers, as distinct from the 'open' consumer service) with special information on ordering, discounts, etc. [End Page 421]

  • • A library service

  • • A page (or tab) for subsidiary rights

  • • A set of tools to enable the creation of custom online catalogues

  • • Tools (APIs) to allow the presses to 'call' or invoke the service and display information under the presses' own brands (eliminating the need for the presses to invest in their own online catalogues)

  • • Tools to enable the use of the content online to be downloaded for print production (e.g., a hard-copy catalogue to be presented to bricks-and-mortar bookstores)

VIII. Organizational Model

If one imagines fifty or more presses on one side, and a large technology company on the other, what structure would best facilitate a working relationship? BioOne may provide a useful model.

BioOne was founded to be something of a Project Muse for the life sciences. It is a unit tucked inside ARL, with a three-person professional staff. BioOne is responsible to its publishers—just under 100, publishing over 100 journals cumulatively. The services for these publishers, both print and electronic, are provided by Allen Press, one of the founders of BioOne. The BioOne staff essentially mediates between its members and Allen Press. Allen Press thus gets to deal with only a single set of contacts, and the publishers themselves, all small professional societies, get expertise from BioOne that none of them could afford on its own.

BioOne does not cover all the bases, of course. It is not creating software, for example; it is essentially a management operation. But the basic structure is sensible and worthy of emulation.

Another interesting model is that used in France by the French university presses. All of the French presses now participate in a consolidated online catalogue, Le Comptoir, mentioned above. The French presses work with a vendor named GiantChair,8 which has offices in Paris, New York, and San Francisco. GC provides the tools for the participating presses to enter and upload metadata about their books. The metadata then appears on the Le Comptoir Web site. The online catalogue has been carefully designed to optimize search-engine indexing, bringing a large amount of traffic to the site. Users who come to the site buy directly from it, bypassing other online booksellers; and this of course is good for the margins of the presses. Le Comptoir [End Page 422] works directly with its vendor (that is, GC), a structure I would not recommend, as it potentially diminishes the presses' long-term control of the consolidated site.

IX. Next Steps

For Academilogue to proceed, the next step (beyond coming up with a better name for it) is to identify a host institution, where a proposal for start-up funding can be developed. Such funding would then lead to a small BioOne-like organization, whose initial task would be to survey and select vendors. Once a principal vendor had been chosen, the new organization would seek to enlist the presses to participate and to establish the various review boards for such things as membership, policy, and technology.

X. Acknowledgments and Disclosures

This essay recapitulates in abbreviated form the findings in a study prepared for the Monterey Institute for Technology and Education. I would like to thank Gary Lopez and Nancy Cook of MITE for their support for this project. The study itself was made possible through the generous support of the Andrew W. Mellon Foundation.

A number of people assisted in the study and the documents derived from it—including, of course, over fifty interviewees who gave considerable time and effort to this project. In particular I would like to thank Philip Pochoda, Marlie Wasserman, Lynne Withey, and Ellen Faran, the directors of the university presses of Michigan, Rutgers, California, and MIT respectively. Karen Barch, COO of CDC (the distribution arm of the University of Chicago Press) assisted me in formulating the key operational requirements. Michael Jensen of the National Academies Press has been a reliable and creative partner throughout. Finally, my special thanks to Terry Ehling of Cornell University, who volunteered to begin the implementation of the project. [End Page 423]

Joseph J. Esposito

Joseph J. Esposito joined GiantChair in January 2010. GiantChair provides direct marketing services to the publishing industry on the Internet. Prior to joining GC he ran his own consulting firm, which focused on strategy for digital media. His clients ranged across the for-profit and not-for-profit sectors and included many academic publishers.


1. Le Comptoir des presses d'universités, http://lcdpu.fr/

2. Book Industry Study Group, http://www.bisg.org/

3. This may be changing, however: see John Willinsky, 'Toward the Design of an Open Monograph Press, The Journal of Electronic Publishing 12, 1 (February 2009), doi:10.3998/3336451.0012.103.

4. See Ann Okerson and Paul Conway, BYTES (Books You Teach Each Semester):Final Report to the Andrew W. Mellon Foundation (NERL 2001), available at http://www.library.yale.edu/~okerson/BYTESFinal070401.PDF.

5. Google search page, http://www.google.com/

6. Michael Jensen is now working with me on the planning of the implementation of this project.

7. Association of American University Presses, 'Books for Understanding,' http://aaupnet.org/booksforunderstanding.html

Appendix A. Blog Post on the Implications of an Online Pod Catalogue Created by Random House UK

[This post appeared at the group blog Publishing Frontier (http://pubfrontier.com) upon the occasion of an announcement by Random House and was designed to elicit comments on the utility of an online catalogue.]

Random House and Its Very Own Print-on-demand Web Site

Posted: November 29th, 2008, by Joseph J. Esposito (edit)

Random House has announced that it will be creating a Web site to market selected titles as print on demand. This has come under criticism in a number of quarters, not because POD is not fully appreciated but because of the truism that no trade publisher has a brand that means anything to a consumer. Thus, RH or any other trade publisher is making a mistake if it believes that consumers will go to the RH Web site. Rather (the argument goes) RH should participate in an aggregation with other publishers, re-creating for POD (or e-books, for that matter) the kinds of aggregation already familiar in the bricks-and-mortar world (e.g., Barnes & Noble) or online (e.g., Amazon). Therefore, POD is great, Web sites are great, but a RH Web site is missing the point.

It may depend on which point you wish to make, however. I happen to agree with the idea that the brands of trade publishers have little [End Page 424] meaning to consumers, despite the handful of exceptions (such as Penguin, Dover, and branded reference works such as Frommer's, the For Dummies series, and Merriam-Webster). And I am all for aggregations. But RH may be looking beyond this. This is because on the Web, aggregation can take place in real time, and what appears to be a would-be stand-alone destination site may really be a starting point for syndication, not to mention an important element of an intellectual property strategy.

To begin with the easy point: the RH Web site is a natural outcome of the proposed legal settlement between various publishers and Google. That settlement marks a significant change in the publishing landscape, from a time when the key split was between works under copyright and works in the public domain, to the settlement terms, where the split is between what is in print and what is out of print. By building an extensive POD site, RH is now asserting that more and more of its titles are in print, thus keeping them under RH's direct control and away from Google's agreed-upon right to exploit titles that are out of print. So score one for RH in terms of intellectual property: What was out of print is now in print, and the POD Web site is proof positive.

Once RH asserts its rights, it can then exploit them. One way of doing this is to create a Web site that is search-engine friendly, which will drive traffic to the RH site. But the traffic need not come to the RH home page; the links can be deep inside the site, on the granular level of individual titles (or keywords associated with individual titles). This is real-time aggregation: the Google search-engine results page is the new B&N, the new Amazon, an aggregation created dynamically every time somebody does a search. In the ecology of the Web, a publisher's own site is simply a loose assembly of parts, each of which is indexable by Google—thus findable and potentially leading to purchases, whether on the RH site or at the site of any other designated storefront. Offline, few publishers' brands mean much of anything; online, only one brand matters, and that is Google. All the rest of the Web is a basket of keywords, woven together by the act of search.

If all that matters is keywords and the individual products they support, why not build a Web site for each book? Not a bad idea, costs aside, but this raises the question of climbing high in search-engine rankings. Now, the algorithms of search engines can change at any time, but at this time a collection of pieces (books, book descriptions, [End Page 425] articles, etc.) has a higher ranking on search engines than would an individual item. The individual book, that is, benefits from the combined search rank of the rest of the site. This is seen clearly with Wikipedia. Test it. Go to Google and search on an obscure item. You will find a link high in the rankings for Wikipedia. You may be the only person who has ever searched Wikipedia for that item, but still the link to Wikipedia is usually among the top four or five on Google. This is because search ranking is cumulative: your search for an obscure item is raised up by the billions of Wikipedia searches on such popular terms as 'Obama,' 'Britney Spears,' and 'George Bush.'

We should not assume that RH does not know how search engines work. RH's Web site will give a higher ranking to all its books simply by putting them in one place and playing to Google's current search algorithms. The RH brand may have little meaning to consumers, but it will develop a huge significance for Google. It's simply wrong to think that the RH Web site is built for people: it's built for search engines, who then direct people to the ranked sites.

Another reason for a publisher to have its own site is simply to assert control of the information about its products. For all the merits of reader reviews, comments, and the like, few marketers of any product like to have others determine what is said about their products. The RH site gives RH an opportunity to create metadata (including abstracts, summaries, reviews, etc.) about each book, content that may then be syndicated across the Web even if no one ever reads it on RH's own site. If a particular title is available from Amazon as well as RH, Amazon may choose to use the RH metadata to sell books at its own (that is, Amazon's) site. This is true of any venue for books, which benefits from free access to the information RH has developed. In this scenario, the RH site is not a Web destination but a toolkit for other sites—not an aggregation in the conventional sense but a repository for others to draw on.

Having taken great pains to assert that the RH brand means little or nothing to consumers—but that having a RH-branded site is valuable regardless—it's probably worth asking if RH may be undertaking a long-term effort to give meaning to its brand. It couldn't do this in bricks and mortar; it couldn't do this when it sold one book at a time. But online, many things change. RH may begin to market subscriptions to certain categories—The Mystery Subscription or The [End Page 426] American Politics Subscription. In effect, RH may be taking the earliest steps toward a new kind of consumer publishing, one in which publishers' brands will matter. Offline, this was impossible; online, anything is possible.

Fundamentally, it's time to stop thinking of the Web as a universe parallel to bricks and mortar. Offline, there are stores; online, there are evolving dynamic relationships. Offline, aggregation is critical; online, aggregation takes place in real time and sweeps up virtual objects wherever an IP address can be found. Offline, B2B brands matter little to consumers; online, such brands can cleverly insinuate themselves into the value chain. We should not assume that the people at RH are stupid, despite the fact that they are, ugh, book publishers. [End Page 427]

Reprinted with permission. A slightly different version first appeared on the Web site of the project's sponsor: http://montereyinstitute.org/.