Discoverability, Demand and Access: the Role of Intermediaries in the UK Supply Chain for Academic Books
Richard Fisher and Michael Jubb
Outline list of Contents
Section 1) The roles of intermediaries
- The distribution of physical books
- E-book distribution
- Digital printing and print-on-demand
- Market consolidation
Section 2) The demand for academic books: a known unknown?
- Some challenges in investigating demand
- The changing role of intermediaries
- Demand, price and elasticities
- Institutional demand, collection and acquisition
- Individual demand, Amazon and exposure
- Changing geographies of demand
Section 3) Discovery, discoverability and the creation of audience
- The centrality of metadata
- Access to e-books: platforms and standards
Discoverability, Demand and Access: the Role of Intermediaries in the UK Supply Chain for Academic Books
Discoverability, Demand and Access is the part of the AHRC/BL Academic Book of the Future project specifically investigating current trends in intermediary behaviour, i.e. the gamut of organisations and institutions, commercial and non-profit, public and private, that link together book-length research outputs (primarily monographs) and their readers, whether within or without the academy and whether in print or on-line. These intermediary agents include not only booksellers and library suppliers but also sales agents, wholesalers, exporters, distributors, aggregators (private or public), metadata organisations and a host of other agencies, each with a significant financial stake in the economy of scholarly communication and collectively in receipt of anything up to 50% of the overall sales revenues of given titles.
This report is by definition international in scope, reflecting the international nature of its subject matter, but, in keeping with the AHRC/BL Academic Book of the Future project as a whole, its primary focus is on publishing developments as they impact scholarly and research behaviours in the United Kingdom. In addition, a lot of the emphasis in this report is (unfashionably) on print books, for two reasons: firstly, academic books in print (howsoever sourced) still constitute anything from 75% to 95% of the total sales profile in the arts and social sciences of all of the major transatlantic publishers active in the sector; and secondly, academic books in print still continue to be the stated preferred format of scholars in the arts and social sciences (and again by much the same percentage proportion). The fact that this preference is not shared by university libraries to anything like the same degree is an interesting and important tension with some significant consequences.
- The roles of intermediaries: an introduction
The task of delivering either a physical or a digital copy of a book from a publisher into the hands of a reader is a complex one, and the complexity has grown rather than diminished as a result of new technologies, not least the advent of e-books, and the continuing quest for greater efficiency and speed, with reduced costs, throughout the supply chain. Some of the agents in that chain specialise in the requirements of the academic book market; but for the most part academic books are treated as but one part of the outputs of the book industry as a whole. These complexities have been compounded, especially in the UK, by a remorseless increase in publisher outputs, both of academic and of consumer/retail titles.
The chain from publisher to reader – for both printed and e-books – starts with the supply and transmission of bibliographic data (metadata created and transmitted to prescribed standards and formats) from publishers to a range of agents in the chain; and specialist database providers such as the British Library, Nielsen, and Bowker play a key role here, alongside standards organisations which bring together representatives of the key players in the industry such as Book Industry Communication (BIC) and Editeur. Getting the data right, and transmitting it speedily and effectively, is critical to ensuring that all the key players in the chain, as well as potential end-users, are aware of all the books that are currently available (both front and back lists). For the retail sector of the market in particular, it is critically important that accurate and comprehensive information about books and their current availability is linked into bookshop ordering, stock management, and electronic point of sale (EPOS) systems. Wholesalers, distributors and library suppliers need similarly accurate and comprehensive information if they are to fulfil their roles in the supply chain.
The distribution of physical books
The largest trade publishers tend to run their own in-house services to distribute physical stock to booksellers and libraries; and some now have web-based e-commerce systems which allow individual consumers to purchase direct from the publisher (though this usually represents but a small part of their sales) . But smaller publishers rely on specialist distributors and wholesalers – in some cases themselves subsidiaries of the larger publishers – with their own warehouses and logistics services. The precise nature of the services offered by such organisations, and the commercial agreements between them and individual publishers, vary according to circumstances. But for the retail trade, services will typically cover warehousing, distribution, credit control, the fulfilment of bulk and individual orders, invoicing and payments, and the handling of returns. Wholesalers meet the costs of such services by purchasing books from publishers at a discount to the published retail price; distributors hold publishers’ stock on a consignment basis, and invoice against sales actually achieved, with agreed discounts and against agreed credit terms. In both cases, agreed discounts are usually shared with booksellers. Again, it is important to note that only the very largest publishers can handle sales and distribution overseas, and that most publishers must therefore negotiate separate agreements with distributors in their key overseas markets.
Wholesalers and distributors will ensure that the stock they hold is listed on their websites and through the databases and links they provide to booksellers (distributors such as Bertrams and Gardners supply their own ordering and EPOS systems to booksellers, as well as making their databases of available stock accessible via Nielsen’s and other systems). They may also promote individual titles through newsletters and the like. But marketing and promotions remain the responsibility of the publisher; this may be achieved in-house, or by outsourcing to specialist sales and marketing agents (again usually separately in the UK and overseas).
Actual sales to consumers are made, of course, through retail booksellers on the high street, on campuses, and online. In the UK, Waterstone’s, followed by Blackwell’s, are by far the most important chains of physical bookshops (both also operate online), although independents retain an important place in the market. Online sales are dominated by Amazon, but there are many other smaller online services such as Hive and Wordery. Both physical and online retailers fund the services they provide through the discounts they receive from wholesalers, distributors or (in the case of online retailers) direct from publishers. And for some popular books they may promote sales by passing on some of the discount to purchasers, though few booksellers now devote much effort to promoting academic books.
Distribution to libraries – particularly academic libraries, tends to be handled by specialist library suppliers, which are in some cases subsidiaries or associates of larger retail and trade distributors (Dawson, for example, is part of the Bertram group; and Askews and Holts part of the Little Group which also includes Gardners). They provide specialist facilities for librarians to search and order books, streamlining library acquisition workflows and integrating with library management systems, catalogues and discovery services. Some library suppliers provide approval plans, under which libraries receive supplies of new titles selected according to a profile of their collection interests, with the right to return what they decide not to buy. Library suppliers usually provide books in shelf-ready form, with covers, subject classification labels, barcodes and security tags, and accompanied by catalogue records in customised form. Like the retail distributors, library suppliers fund such services out of the discounts they negotiate with publishers; and again a proportion of that discount is shared with their library customers. A UK universities procurement consortium has framework agreements with eleven library suppliers, the majority of whom provide physical books as described here, as well as e-books.
As with physical books, publishers can sell e-books direct to consumers online, either in the form of outright sales, or as subscriptions for a limited time. But most sales and subscriptions are made by e-book vendors which aggregate content from a range of publishers. In the retail sector, Amazon is by far the dominant player, but other vendors include Google and Apple, along with smaller players such as Hoopla and Kobo.
Again as with physical books, library suppliers offer specialist services. Major academic publishers such as the university presses of Oxford and Cambridge, or Taylor and Francis, have developed their own platforms through which they make collections of their titles available to libraries; and OUP and CUP make their platforms available to other publishers. There is some evidence that libraries are becoming more interested in acquisitions direct from publishers, since they provide access free of digital rights management (DRM) technologies which restrict the usage of e-books in various ways; but most sales to libraries are currently made via specialist suppliers who aggregate titles from a range of publishers, and provide their own platforms, along with (as for physical books), links to library management and discovery systems. Major suppliers for UK academic libraries include ProQuest, EBSCO, and Dawson. The terms on which books are supplied range from subscription for a limited period, to outright sale, and there has been considerable experimentation with different sales models in recent years, as discussed in Section 2 below.
Digital printing and print-on-demand
A key change that has had a major impact on the relationships between publishers, library suppliers, distributors and wholesalers (and eliding the differences between the latter two) has been the rise of digital printing, and in particular of print-on-demand (PoD), greatly facilitating short-run impressions of specialist work. The rise in the quality of digital printing, the fall in its costs, and the development of systems for speedy fulfilment of PoD orders, are together of particular importance for titles – like many academic books – with low unit sales; and they are profoundly changing the economics of publishing such books. It is arguable that the impact to date has been greater than the advent of e-books, and Amazon (in particular) has been a hugely important transatlantic channel in this context, working as both retailer and as producer.
For wholesalers and distributors, these developments have opened up the possibility of providing for publishers a digital virtual warehouse, with no physical books, and fulfilling orders (which may come from booksellers, libraries, other distributors, or from publishers themselves) either as e-books or physical printed copies.
As noted above, funding the work of the different players in the supply chain from publisher to reader takes up to 50% of the revenues from end-purchasers (either retail purchasers or libraries). But competition is fierce and margins tend to be low; and the demands for investment in logistics and a wide range of online services, e-book platforms and so on are relentless. Hence a key feature of the market in the last few years – from bibliographic services through wholesaling and distribution, to retail bookselling – has been increasing consolidation. In the UK, companies such as Nielsen (in bibliographic and market intelligence services), Gardners and Bertrams (in wholesaling and distribution for the retail trade), ProQuest and EBSCO (in library supply), and of course Amazon in retail sales, have become increasingly powerful, partly through the acquisition of smaller companies and successful start-ups. Publishers, retailers and librarians have all expressed concerns about the rate of consolidation, and as of 2016 it looks rather unlikely that the situation will be ameliorated by new entrants to the market: indeed, further consolidation looks much the more likely outcome.
- The demand for academic books: a known unknown?
Some challenges in investigating demand
This Discoverability, Demand and Access project has included an extensive interview programme with senior representatives of selected key international intermediaries (see below). We are nonetheless (very) conscious of the limitations, both temporal and geographical, of the evidence-based we have used, but this investigation is at least a start in what remains in general a very under-examined and statistically deficient nexus of relationships. As others (prominently Geoffrey Crossick in his 2014 HEFCE Report on Monographs and Open Access) have also noted: there is simply no robust transatlantic, let alone global data, available on the unit sales of academic books in the arts and social sciences (valuations of ‘overall market size’ are emphatically not the same thing).
Indeed, it is arguable that there are few areas of scholarly communication where more is confidently predicted or stated on the basis of less concrete evidence or hard data than the demand for academic books. Rather too often, whether within legacy publishing practice or in the behaviours of new, sometimes Open Access start-ups, the nature of the demand for academic books has been taken as a given, from which each sector draws its own conclusions. Leaving aside student textbooks (in keeping with the Academic Book of the Future project as a whole), there is an acknowledged meta-narrative of long-term transatlantic sales decline at the title level, very significant downward institutional pressures (driven, not least, by the competing demands of escalating journal subscriptions), reduced external exposure to individual consumers (not least with the massive changes in retail bookselling of the past two decades) and steeply rising unit prices, which in turn drive down further already weakening levels of demand.
Like all such generalities, this narrative contains within it much (particularly at the title level) that is unquestionably true, and supported by extensive publisher (and indeed author) experience. It also, however, throws up as many questions as perhaps it answers, particularly when (as this project has endeavoured to) some effort is made to distinguish between sales and demand (which may not of course necessarily be satisfied), institutions and individuals, and access; and these complexities are further compounded by the transitions wrought by the digital revolution of the past two decades which, whilst its impact on the academic book has unarguably been less profound (thus far) than on the academic journal, has nonetheless wrought very significant changes indeed to modes of discovery, access and production (whether of physical or electronic books).
The changing role of intermediaries
The most important single conclusion of this ‘Demand’ part of the Discoverability, Demand and Access study is perhaps this. Over the past two decades, the role of intermediaries in driving or enhancing demand for individual academic books has fallen very sharply, leaving publishers as the only agents actively driving this critical agenda. Given that those publishers have been simultaneously responding to the concurrent fall in unit demand by actively increasing title output (especially in the UK), the actual marketing attention given to each title has fallen, both within publishers (given that their overall levels of specific marketing expenditure have remained largely constant as a percentage of title revenues) and, much more dramatically, on the part of other agents in the supply chain. In a publisher context where the cost and complexity of running, effectively, a dual ‘print-and-electronic’ book publishing operation (in terms not just of core published product, but of almost all marketing ancillaries) has compromised still further already fragile publishing margins, such shrinkage is understandable. Nonetheless, it needs to be recognised more than is sometimes the case that external pressures on book expenditures have been compounded by actions from within academic book publishing itself, across its various agencies; and the levels of demand-driving attention given to individual titles has thus been actively reduced just, arguably, when the need for it was actually growing. This core conclusion may well be contentious to some and merits expansion.
A quarter of a century ago, a transatlantic network of wholesalers, distributors, agents, library suppliers and (perhaps above all) booksellers all saw the active promotion of academic books, and especially new academic books, at the title level as an explicit part of their remit, and part of their justificatory value-add when negotiating discount terms with originating publishers. Whether promoting through their own catalogues, advertisements, co-operative exhibits at Book Fairs and academic conferences, (especially in overseas markets outside the UK and US) or other mechanisms, these agents all assumed significant responsibility for generating demand for individual titles as they were published. The digital transition, and perhaps most fundamentally the impact of Amazon, has transformed this picture fundamentally, as have the related sequence of consolidations, take-overs, collapses and market disappearances that have reduced a once vigorously-diverse print-selling sector (both at the institutional and individual level) to one now dominated by a very small number of (largely) international players, with many of whom we have spoken during the course of this project.
The core priority for many of these agents in 2016, and one to which most of their inward investment has been directed, is no longer the expansion of and subsequent satisfaction of new demand, but rather the immediate or instantaneous satisfaction of pre-existing demand, responding to the changes in both individual customer and institutional behaviours wrought above all by the impact of Amazon, and its culture of more-or-less-immediate gratification. Now clearly Amazon are (inter alia) the world’s most important academic bookseller, especially of monographic work in paperback whose availability as noted elsewhere has been transformed by the impact of Print on Demand technologies (see RKF in The Scholarly Kitchen etc), and any analysis of the demand for academic books that does not take Amazon’s transformative role very seriously will be sadly deficient. Nonetheless it is important to recognise that Amazon’s transformational impact has been as much in this indirect capacity, affecting the behaviour of legacy intermediaries, as in its own bookselling activities.
Demand, price and elasticities
To those working within traditional or legacy book publishing, the (very) price-inelastic nature of the demand for most academic books is a core belief. Amongst the most telling pieces of hard evidence in support of this assumption has always been the recurrent performance of ‘first-time’ or ‘new in’ paperbacks, invariably monographs, routinely and remorselessly selling perhaps a half or even a third of their original £55 hardback numbers than when released perhaps eighteen months later in their new £20 paperback guise. As has been widely discussed elsewhere, the advent of short-run printing and on-line bibliographic discovery and market availability greatly liberated and de-risked this strand of publishing in the first decade of the present century, a development that has been particularly resonant within the monographic world of the arts and humanities.
However, this has emphatically been a development within a traditional publisher nexus, and one largely serving traditional academic publishing audiences, whether institutional or individual (and in this context it is important not to underestimate the related significance, not least in Asian markets, of the Amazon proposition for institutional customers, as much as for individual academic consumers). What, in a world of potentially universal global on-line access, are legitimate levels of demand for academic books? Is the poverty of the aspirations of legacy publishers, informed by their experience of the past generation of the shrinking print market for individual academic books, circumscribing their sense of potential demand, and reinforcing a (very) conservative and cautious view of market possibility? Fundamentally, is there a large non-specialist and much more geographically dispersed audience for academic books that existing print-based publishing models are failing, utterly, to reach?
The perspective of massive demand-underarticulation by legacy publishers, especially but not exclusively in the Global South, is certainly intrinsic to the publishing model of the new Open Access academic book initiatives that have launched over the past decade, and Dr Rupert Gatti of Cambridge-based Open Book Publishers has provided one of the best-known and cogent articulations of this view on http://blogs.openbookpublishers.com/introducing-some-data-to-the-open-access-debate-obps-business-model-part-one/
Such presentations, of course, invite other, much more wide-ranging questions about the democratisation of knowledge and the provision of open learning resources that go far beyond this brief study of publishing intermediaries. There is also a real danger that this debate about the global demand for scholarship is reduced to a simple closed-open binary, when the reality is altogether more nuanced, and especially when the important distinctions between availability and access, and product-knowledge and effective demand, are drawn out. In a context of very significant increases in scholarly outputs of all kinds, differentiation and effective articulation of those outputs is arguably even more important in the online context than is the case in the legacy environment of print, and it seems unarguable that one of the biggest barriers to the more rapid penetration of Open Access publishing models into academic books (leaving aside the still widespread and recurrent preference for print (as opposed to e-books) of most scholars working within the arts and humanities) has been a lack of investment in title-marketing, and in the intermediary workflows discussed below, and corresponding lack of public visibility, despite their open status. Whether simple website ‘hits’ on freely-available open material constitute evidence of sustained demand remains unproven, and as yet there is no conclusive evidence of the extent of the conversion, in new and emergent markets, of website hits into active usage and genuine impact. Indeed, achieving effective institutional recognition of new Open Access academic books within a library context has proved a major challenge, as has (and this is absolutely not an issue restricted to OA publications) title recognition within those institutions that actually enjoy access. Current processes mean that libraries make more efforts to improve the metadata for books they have purchased, than for OA books, where there is no trigger to stimulate such work. Achieving positive individual demand for the publications, whether paid-for or Open, held by institutions has been something that all respondents to our enquiries have stressed as a core current priority, for both publishers and libraries, and it is fair to say that no parties are particularly happy with the current status quo, as the next section makes clear.
It is also worth emphasising, given that the financial margins of most academic book publishers are very significantly lower than amongst the large majority of academic journal publishers, and that very significant players in the academic book market are both in practice and by statute non-profit University Presses (especially in North America), book pricing in response to this agreed demand inelasticity has nonetheless varied greatly, with North American university presses recurrently stressing perceived unit-sale-maximisation and in general much less robust and more ‘author-acquisition-friendly’ pricing models than their European peers, whether amongst University Presses or in the commercial sector. American University Presses also emphasise domestic demand to a far greater extent than their UK-based counterparts, almost all of whom have been export-driven for (at least) a quarter of a century (see below). Such a pricing strategy has, of course, depended on a very different financial relationship between parent universities and their university presses than that which pertains in the United Kingdom, and has accordingly been exposed to significant strain in recent years as many North American Universities, especially Public Universities, have recalibrated their Press operating models in response to recessional pressures. The reorientation of parts of the American university press sector, and the establishment in certain institutions of a much closer reporting relationship between university libraries and university presses, is emphatically a consequence of both financial readjustment and of the wider operational trends in academic book markets that we have observed in the course of this study. As certain North American institutions emphasise a more locally-focussed (in terms of its author base) University Press, operating on some form of Open Access model or models and providing, above all, faculty-centric digital publishing services (rather than publication in the traditional sense), a potential tension is emerging between these internal-facing institutional priorities and the external emphasis intrinsic to the act of publication per se, and indeed intrinsic to the behaviours of almost all University Presses over the past half-century.
Institutional demand, collection and acquisition
Until about ten years ago, the great majority of university libraries built up collections of printed books, and regarded the development of those collections – based on their knowledge of the university, its courses, and the scholarly interests of its academics – as a key part of their mission. But the rise of e-books, combined with increasing pressure on library space for physical books, and on library budgets (particularly the budgets for book acquisitions, given increasing expenditure on journal subscriptions), have brought fundamental changes over the past few years.
Librarians in large research-driven universities were already becoming concerned about the high proportions of the books in their collections that were never borrowed, since given the huge numbers of titles available in any particular subject, it was impossible for library acquisitions staff to know precisely which titles would be relevant and used by the undergraduate and postgraduate students and staff working any specific subject area. Hence there had been for some years talk of the need to move from ‘just-in-case’ to ‘just-in-time’ collection-building: enabling users to connect to the content they need (not necessarily collecting it), and acquiring specific titles only when there was clear evidence from academics and students of demand for them.
The rise of e-books, and of readily-available metadata about them, has transformed libraries’ ability to achieve that goal. Library suppliers have created large collections of e-books, and new web-based discovery and access systems (see Section 3) enable them to provide information about those books direct to library users. This makes it possible for libraries to let their users find and identify books relevant to their needs through the library’s discovery service before any decision is made to purchase them. Since around 2010, ‘patron-driven acquisition’ (PDA) – now more commonly termed ‘demand-driven acquisition’ (DDA) – models have become increasingly common as one of the key ways in which academic libraries develop their collections. The models vary in detail, and in some cases, as with YBP (now part of EBSCO), DDA can be one of the purchase options associated with approval plans. DDA can also be used as a model for the acquisition of print books, though the mechanics are more complex, and it is neither as effective nor as widespread as DDA for e-books, not least because the time taken to supply physical books; some libraries report that significant proportions of print titles ordered via DDA are never retrieved by the requester..
Current DDA models require a complex technical infrastructure, though some booksellers suggest that it could be simplified if, for example, libraries were prepared to forego MARC records and make use of existing e-commerce routes. With all the variations in the practical aspects of the model agreed between aggregators and libraries, DDA essentially implies that metadata about the book appears in the library’s discovery service; and with access from the aggregator’s site turned on, once a certain number of pages have been read, or a number of short-term loans (STL) made, the library automatically purchases the title, without the user(s) ever knowing that it had not always been part of the library’s collection (unless the library chooses to alert them ). STL can itself prove an efficient means of meeting demand, and considerably more efficient than inter-library loans (ILL). Finally, purchase under DDA models may take a number of forms, from permanent acquisition to licenses for access with various restrictions on the number of uses, number of simultaneous users and so on.
Such arrangements have many attractions for libraries: their users have access to more titles, the costs of acquisitions typically fall, and the development of collections is directly related to the activities of students and staff. In a virtuous circle, all this can amount to a means for libraries to demonstrate to university managers that they are operating both efficiently and effectively in meeting the needs of the university, its staff and students, thus helping to sustain ( and in the best case to enhance) library budgets.
While many publishers and aggregators actively promote the comprehensiveness of their e-book packages, and many libraries still make a virtue of the size of their collections, there is a growing realisation that the core interest on both sides is to ensure that libraries acquire the right books that fit best with their university’s research, teaching and learning activities and strategies, and that library users get access to the titles they want or need. Nevertheless, DDA poses problems for publishers: purchases are delayed, or may never take place, as compared to the ‘just-in-case’ library acquisitions model (which thus creates a cash-flow problem for publishers); short-term rentals replace what might otherwise have been some outright purchases (although this may be ameliorated by the new “Access to Own” model introduced by ProQuest); and many sales that might otherwise have taken place are lost altogether. As a result, revenues fall.
As a response to these problems, many publishers have developed ‘evidence-based acquisition’ (EBA) models as a hybrid between DDA and the outright sale of a complete collection. Libraries pay an upfront fee at a discount to the full cost of a collection, and users have access to the collection for an agreed period (usually a year), at the end of which the library decides which titles to purchase, based on levels of usage or other criteria such as the fit with the library’s collection priorities. The number of titles in the collection may be negotiated between the publisher and the library, and again, precise arrangements vary; under some models, for example, libraries agree on an amount, in addition to the access fee, that they will spend on purchases at the end of the access period, with or without flexibility on carry-over to a subsequent period. The advantage to libraries is that as with DDA, they acquire only what their users need, and they may have more control over the selection criteria and expenditure than with DDA. Moreover, the technical infrastructure needs are not so complex. For publishers, the advantage is essentially the fee they receive upfront.
The balance of advantage as between DDA and EBA for libraries and for publishers varies according to individual circumstances. Even the most comprehensive of the aggregators’ packages for DDA typically do not cover all the titles of every individual publisher, and some publishers have withdrawn titles from DDA packages. Libraries are also concerned about aggregators’ use of digital rights management (DRM) mechanisms to protect publishers’ content; but there are obvious advantages for libraries in dealing with the major aggregators, and the support they can offer in enhancing acquisition workflows, rather than dealing separately with a large number of publishers. Nevertheless, it can make sense for libraries to buy titles direct from publishers (where DRM is typically much more relaxed, if present at all), title by title, package by package, or via EBA, particularly where the library requires a significant volume or proportion of the titles from that publisher. There is thus a growing literature on the relative advantages of different acquisition models.
Taken together, the advent of the widespread availability of e-books, along with the development of large-scale aggregations, of the platforms to gain access to them, and of web-scale knowledge bases and discovery systems, have brought great changes to the relationships between publishers, libraries and library suppliers. First, while there is strong evidence from readers – despite the convenience of e-books for some purposes – of a continuing preference for print when reading books (unlike journal articles), there has been a significant shift in library purchasing in favour of e-books. Second, this has had a major financial impact on publishers, libraries and suppliers, with pressures on revenues for publishers as e-book sales have cannibalised print sales; and pressures on suppliers too, since margins and discounts for e-books tend to be smaller than for print, while the need for investment in platforms and IT infrastructure is considerable. This has contributed to consolidation in the numbers of suppliers and aggregators, as smaller companies have been acquired by bigger ones such as EBSCO and ProQuest. Third, the development of DDA, STL and EBA models, along with approval plans for e-books as well as print, has brought major changes in the patterns of acquisition across the academic library sector, with libraries seeking the most cost-effective means of acquiring the books they need. Fourth, the many different variations in terms and conditions associated with those models that have been introduced, and the frequency with which they have been modified in the light of experience on all sides, have brought difficulties for all parties in judging what works best for them and their partners.
Some publishers criticise libraries for having abandoned responsibility for collection development, passing it to a combination of the major aggregators and library users. Some suppliers also argue that libraries are in increasing danger of disintermediation, as large companies take responsibility for all the key systems to meet the content needs of students and academics. For at least some of the larger and better-resourced libraries, however, the changes outlined above, together with the use of online reading list and research management systems, are beginning to enable them to become more strategic in the development of their collections, and to link collection development more closely to institutional strategies for teaching and learning and for research. Such libraries are seeking to analyse data from a range of sources – institutional research, teaching and learning systems, bibliographic databases and knowledge bases, usage data and so on – to build profiles of their collections and of how to develop them to meet demands and needs across the institution. They can then use a variety of the acquisition models outlined above to try to ensure that their collections do indeed meet institutional needs, as efficiently and cost-effectively as possible. Achieving all this demands new skills from librarians, not least in data analysis. Whether it also means acquiring more books, print and/or digital, is not yet clear.
Individual demand, Amazon and exposure
The much-increased technological and data sophistication of university and research libraries in recent years has meant that, whatever the acknowledged lack of global sector statistics, there has been at least some exploration of patterns of institutional behaviour in the acquisition of academic books. By contrast, general studies of the patterns of demand for academic books by individual consumers, whether faculty, early career researchers or non-specialist readers, have been very few and far between and largely restricted to the North American and, especially, University Press experience. The publishing profile and pricing policies of many North American University Presses has generated guesstimates of individual demand levels for academic books of up to 50%, for certain Presses (e.g. the University of Chicago Press) in certain disciplines. Such figures will include significant elements of upper-level adoption sales (i.e. academic books adopted for use in course teaching, very often at graduate level), and for both commercial and non-profit publishers in the arts and social sciences, the North American taught graduate course was long a primary publishing focus, especially of those ‘new-in-paperback’ monographs referred to above (and more recently revalorised and de-risked by the advent of short-run printing technologies). The impact of Amazon on this publishing sector, and upon transatlantic campus bookselling at this graduate level, has been huge, and arguably rather greater (thus far) than on undergraduate textbook behaviours. As discussed elsewhere in this report, the implications for major high-street bookselling chains in the UK like Blackwell’s and Waterstones (which began life as a retailer of upper-level philosophy books on the Charing Cross Road in London) and Barnes and Noble in the USA have likewise been profound; and Waterstones and Barnes and Noble have both experienced a significant ‘retreat from range’ in their academic stockholding, with obvious consequences for title visibility and physical exposure of a certain kind, both compounding and a consequence of declining marketing at the individual title level. Publishers are having to consider how they might fill the gap by promoting and marketing individual titles.
Moving away from the academic purchasing environment, it is a complaint often made by scholars in the humanities and social sciences that the number of subjects for which any kind of non-specialist demand is well catered are actually very few, and that ‘academic impact’ titles aimed at a broader audience, and published by publishers that include major commercial imprints like Penguin Books and Harper Collins as well as much more overtly scholarly publishers like Oxford University Press, often inhabit a very narrow disciplinary spectrum, in which history and biography (above all) predominate, but from which works of ideas, let alone theory (!) and quantitative work are very often excluded. This presumption probably underestimates the impact in the past decade of important American University Presses like Princeton, as well as other independent publishers like Profile or Polity Press, but what seems unquestionably true is that in recent years non-fiction book retailers have been forced to make much harder discriminatory purchasing decisions, in which ‘crossover’ or ‘academic impact’ titles are designed as such from the outset, rather than emerging as a result of a series of serendipitous punts at various points along the supply chain (something which of course could be seen as simply symptomatic of the inefficiencies of the sector). In that sense, one consequence of the enormous expansion in UK academic book outputs in recent years has actually, and ironically, been a narrowing of promotional possibility, and whilst Amazon has been an enormously liberating force for very significant elements within academic book publishing, but for whom (it is arguable) several major academic imprints would have ceased to exist, it is likewise much more difficult to argue that its impact in this particular (albeit restricted and/or elite) sector has been entirely benign.
When discussing Amazon and the individual demand for academic books, one important consequence of the changes in production and operational workflows wrought by the digital transition, associated printing technologies and the growth of e-tailing has been, for many publishers, a revalorisation of the backlist, something that has continued with the related expansion (see below) of publisher e-book offerings. The extent to which university libraries or individual consumers and faculty have been driving the (unexpectedly significant) global demand for the numerous programmes of paperback backlist revivals launched by every major humanities publisher is not well understood, but one generally accepted consequence has been re-balancing of the sales profile of many academic monograph-led publishers from a relatively standard 40%/60% split, as between the sales of new books and backlist titles, to something often closer to 25%/75%. One possible interpretation of this shifting ratio is that it demonstrates not so much the strength of the backlist, important though those sales are but rather the weakness of frontlist sales, and the consequences of the diminished new-title-marketing cited elsewhere in the report.
Changing geographies of demand
It is central to the global political economy of academic book publishing in the arts and social sciences that it remains (happily) a multilingual undertaking, in marked contrast with academic scientific serials which are fast becoming outside Asia an almost exclusively English-language publishing format. The existence of strong and distinctive academic book publishing traditions in (e.g.) French, German, Italian, Spanish, Portuguese, Hebrew, Arabic and a multitude of oriental languages, whilst each faces perhaps even more severe versions of the pressures to which English-language academic book publishing has been subjected, is central to many disciplines within the arts and humanities, with special resonance for UK-based scholars in subjects like history, classics, philosophy and theology, as well as (most obviously) modern foreign languages. A long-standing tradition of high-level translation both from and into the English language, with historic financial support from (notably) both the French and German governments, has been characteristic of both the humanities and the social sciences, although the extent of this activity has undeniably shrunk in recent years, and indeed academic translation is arguably the only major book publishing sector in which output has actually fallen over the past generation.
This linguistic variety is nonetheless one basic reason why geographies of demand for academic books in the arts and humanities diverge significantly from those for scholarly work in the STEM subjects and the harder social sciences. Again, hard statistical data is elusive, but as a working generalisation most major UK-based arts and humanities book publishers report continued levels of significant demand in (especially) northern Europe; in Germany, Scandinavia and the Netherlands, but with rather lower (and declining) levels of demand in Italy, France and the Iberian peninsula. These trends vary somewhat by subject, but it is worth emphasising that for many of the larger UK-based publishers, commercial or non-profit, reported European demand for academic books outside the UK comfortably exceeds that within the UK itself, and this of course has been one reason why this publishing sector, almost above all others, was vocal in its support for the losing Remain side during the recent Brexit referendum, and the medium-term implications of that decision are potentially very significant indeed for the sector as a whole.
Globally, the larger UK-based academic book publishers active in the humanities and softer social sciences report overall sales profiles that conform, very roughly, to a pattern of two parts European and two parts American (north and south) to one part Asian-Pacific. Now, as we have emphasised throughout this report, that 40%/40%/20% ratio is not (necessarily) indicative of potential demand, but it does emphasise the internationalism at the core of the proposition, and central to the sustainability of the legacy publishing model. Our project interviewees all emphasised that higher education systems in Asia and in emergent markets, including the BRICs, were investing far more heavily in STEM subjects and the applied social sciences (especially business, management and finance) than in the arts and humanities, and that important Asian territories of traditional high demand for English-language academic books (most notably Japan) had experienced stagnant, if not declining sales for nearly two decades, and that the anticipated growth sectors like China had (in this arts and soft social science context) disappointed: India, which possesses one of the most vibrant English-language academic publishing cultures in the world, was widely cited by respondents as the territory with the greatest potential going forward, although the fiduciary and regulatory challenge of meeting Indian demand (even on this legacy publishing model) was not to be underestimated.
Fundamentally, it looks very unlikely that what had once, however speculatively, been seen as the salvation of traditional monographic publishing models, namely the significant expansion of sales into new and emerging centres of demand beyond the traditional transatlantic core, is going to materialise. UK-based publishers reoriented themselves successfully into highly-export-driven businesses during the 1970s and 1980s, with important implications for their editorial acquisitions policies. Few of our respondents, certainly, see similar transformative, let alone salvatory potential in emerging markets now, certainly on any kind of legacy publishing model.
- Discovery, discoverability and the creation of audience
The centrality of metadata
Latent demand can be turned into effective demand only if potential purchasers (retail and library acquisition staff) and library users become aware of a book’s existence, and where and how they can acquire it, to purchase or to read. To state the obvious, what potential purchasers and readers do not know about they can neither purchase nor read. Hence the critical importance, particularly in the context of the reductions in marketing efforts outlined above, of ensuring that comprehensive information both about new and forthcoming titles and about what is still available on back-lists, flows effectively through the supply chain we have described. Marketing efforts at conferences, festivals and bookfairs, the distribution of catalogues and flyers, advertisements and reviews in scholarly journals (and where possible, in more widely read magazines, newsletters and the like), the provision of inspection copies, and the traditional work of sales staff, still remain important; but such efforts will be limited for most titles, and their impact on individual purchasers and readers may well be similarly limited and indirect.
In such a context, it is of particular important that all possible steps are taken to ensure that accurate, comprehensive, high-quality standardised metadata – information about books and their availability in different formats and from different sources – flows through the supply chain to academics and other end-users . But that is a complex business; and the systems do not always work as effectively as they might.
The creation of metadata starts with publishers, and for new titles it starts several months before publication. The largest of them publish thousands of new titles a year, the middle-sized ones a few hundred, and a long tail of small publishers release only a few titles each year. The digitisation of back-lists, making them available as e-books or via PoD, adds to the volume flow, along with the need to produce metadata records for each of the different formats in which a title is available. The larger publishers are also seeking to produce metadata for book chapters as well as whole books. The huge – and increasing – volumes of metadata mean that machine-to-machine interfaces which require minimal manual intervention are essential. That in turn implies the use of formats and standards to structure the large amounts of descriptive and administrative metadata that is transmitted through the supply chain.
ONIX is the standard format used by publishers for this purpose, but it can be challenging for large publishers and daunting for smaller ones. The creation of basic metadata starts in the editorial process, and since many smaller publishers do not employ metadata specialists, editorial staff may have full responsibility for creating the metadata that is transmitted through the supply chain. But small mistakes in metadata content and formats can have a significant impact on the discovery, sale and use of publishers’ titles. The metadata aggregators and vendors, such as Nielsen in the UK and Bowker in the USA, do some quality checks and standardisation, before delivering records to wholesalers, distributors, library suppliers and booksellers; ensuring that the data feeds from the metadata vendors to the full range of relevant distributors, wholesalers and booksellers work effectively is therefore of critical importance for the smaller publishers in particular.
But the interests and requirements of the different agents in the chain are not congruent. There is a particular tension between the needs of publishers, distributors and booksellers – especially for retail sales – on the one hand, and libraries on the other. The ONIX standard essentially meets the needs of the former; but libraries want MARC (Machine Readable Cataloguing) records – created according to an internationally-agreed set of standards and guidelines – to enable them to document and manage their collections and to allow for the creation of collection-wide finding aids for readers. MARC records are typically created not by publishers, but by metadata vendors and national libraries, making use of information supplied to them by publishers (though some larger publishers are now beginning to produce MARC records themselves). The ONIX and MARC standards result in very different kinds of records created for different purposes: the administrative data on marketing, production, price and availability for the fulfilment of sales orders that is of critical importance for ONIX records is not included in MARC records, while MARC records themselves contain much more fine-grained subject classification codes than those needed in the ONIX records, so that national and global aggregations of bibliographic data covering all the titles, in all the formats, held in various collections can be created and searched. Given the different interests, roles and responsibilities of the different agents involved in creating, amending, enhancing and transmitting metadata through the supply chain to purchasers and readers, it is not surprising that the quality and comprehensiveness of the records varies too; and there is considerable duplication of effort. From a library perspective, these complexities can affect the display, discoverability and de-duplication of records once they are transferred to MARC environments; and may even mean that book reviews become more prominent in searches than the titles themselves.
Despite these difficulties, it is generally thought that even though – as noted in our discussion of demand – there have been marked reductions in the marketing efforts to make potential retail purchasers aware of new titles in which they might be interested, once their interest has been piqued, they are reasonably well-served with means – not least through the online catalogues and search facilities of booksellers such as Blackwell’s Waterstones and , of course, Amazon – of finding out precise details of a book, where and how it might be purchased, in what formats and at what price. None of the metadata vendors’ databases on which such services are based is fully-comprehensive, with particular gaps in the coverage of the smallest publishers’ lists and of foreign language titles. And it is worth noting that publishers, wholesalers and distributors who make titles available via PoD have learnt – mainly because of the reputation for low-quality production that remains from only a few years ago – to avoid any mention of PoD in their ONIX feeds and the databases used by booksellers; for how long this will remain a problem is not clear. For most purposes, however, search and query services covering front-list and back-list titles for booksellers and end-users work reasonably well.
Nevertheless, the reduction in marketing efforts at the title level – particularly those aimed at the key target audience of professional academics – does bring problems, even if there are some complaints about academics being overloaded with publishers’ catalogues. For academics are not only potential readers and purchasers in their own right; they also act as key intermediaries, recommending books to undergraduate and postgraduate students by including them on course reading lists. Such lists – now increasingly created and managed on online systems such as Talis Aspire – then act as information sources for acquisitions librarians: in the best-resourced universities, libraries typically treat the appearance of a title on a reading list as a direction to acquire it.
In a world where online search via Google, Amazon and a host of other services is so pervasive, however, it is easy to forget that search is of limited use in alerting end-users to titles of which they are unaware. And the online recommender systems currently provided by Amazon and others are not especially well-suited to dealing with the highly-specialist world of academic books. There is work to be done before publishers and online booksellers can fulfil the job that the best bricks and mortar academic bookshops – like the best online retailers of consumer goods – can do in shaping readers’ choices by making them aware of – and stimulating their interest in – titles that they did not know about. Achieving such an outcome is made the more difficult by the relatively coarse-grained subject classifications in the form of BIC codes used in ONIX metadata, and by the lack, for most books, of chapter or section-level metadata.
On the library side of the academic book business, the issues are somewhat different. As we have noted, sales to academic libraries tend to be dominated by the large library suppliers .The GOBI (Global Online Bibliographic Information) system run by YBP, now part of the EBSCO group and operating mainly in North America, provides an interface for searching, selecting and ordering titles, from a range of publishers and suppliers, and seeks to support the complete acquisition and collection development workflow for libraries. Other suppliers operate similar services based on their own databases and ordering systems. But none of these databases is comprehensive, and libraries may have to make special efforts to find books from specialist sources. As noted earlier, once books are purchased, they come with MARC records which libraries can import into their catalogues and library management systems.
When it comes to e-books, the position becomes more complex. Companies such as ProQuest, EBSCO, Dawson (in the UK) and Baker and Taylor (in North America) aggregate titles from hundreds (or more) of publishers, and create collections to which they provide access on their own platforms. Some e-books are also sold to libraries direct from publishers. As with printed books, both aggregators and publishers seek to ensure that once libraries acquire e-books, their catalogues and discovery systems record the fact that the e-book is indeed accessible.
Most academic libraries have in recent years adopted web-based discovery services founded on Open URL knowledge bases that seek to offer comprehensive coverage of the content that is available from a wide range of sources. Since not every institution has purchased access to all the content that is in theory available to them, the knowledge bases provide tools to customise them so that they reflect institutions’ individual collections: the content that can be accessed electronically by their users, as well as what is owned by the library in print format. For e-books (like e-journals) link resolvers are then used to find the service(s) through which the content is available, and a link is created to the online text. But this system works only if the data in the knowledge base is accurate, comprehensive and up-to date, with constant updates of both descriptive and administrative metadata from publishers, aggregators, libraries, link resolvers and others throughout the supply chain. Both libraries and the discovery service providers have found it problematic to meet these challenges, as both the volumes of e-content and the complexity of the models for acquiring access to that content have increased rapidly.
Libraries may acquire e-books under a subscription model, where access is purchased for a limited period, or under one or more of the PDA, DDA, EBA, STL and approval plan models discussed earlier (Section 2). Under any of those models, the discovery system must turn on access to large numbers of e-books for the period of the agreement, modify the access period if the title is actually purchased, and turn off access to those titles that have not been purchased at the end of the agreement period. There is considerable scope for things going wrong at any stage; publishers and libraries both complain that this happens frequently, and that they have limited ability to put things right. Libraries find difficulty in monitoring whether they have sufficient copies, when the same titles are held under different purchasing models. Publishers also have concerns about lack of visibility and understanding of how data are used, the risk of dilution of their brands which may be lost or hidden in research results, and lack of transparency as to how relevance ranking systems work. These problems are of particular importance as both libraries and publishers, partly because of the PDA and other acquisition models mentioned above, are paying greater attention data on the usage of e-books. And some smaller aggregators such as JSTOR claim that usage of their content has fallen when libraries have implemented new discovery systems of the kind outlined above, rather than making use of the system provided by JSTOR itself. This may reflect ways in which the simple search interfaces in library discovery systems are geared more to the needs of undergraduates than the more specialised scholarly needs of postgraduates and academics.
Access to e-books: platforms and standards
The different supply models for e-books in the library and retail markets have implications for modes of access. The retail market is dominated by sales of e-book downloads in various formats appropriate either for e-book readers, such as Amazon’s Kindle, or for laptops and tablets. Downloads are sold outright for users to keep, although Digital Rights Management (DRM) systems may restrict usage in various ways.
In the library market by contrast acquisitions usually take the form of purchasing access to e-books which are hosted on a third-party website. Both aggregators and the larger publishers have established web-based platforms for this purpose. Libraries must perforce deal with many different platforms; and they vary in many different ways.
First, and obviously, the platforms provide access to different titles (which is why libraries have to purchase access to content on different platforms). But there is often considerable overlap between their contents: publishers release different overlapping but not congruent sets of titles to different aggregators; and sometimes titles are accessible only on the publisher’s platform. Hence a library will have a mix of titles some of which are accessible only on one platform, while others are accessible on several. Even worse, from a library and a user perspective, an aggregator’s contract with a publisher may change, or come to an end, which may mean that titles disappear from the platform, or that the functionality associated with a particular publisher’s titles is withdrawn.
Second, different platforms have different levels of DRMs which may affect, for example, the amount of downloading, or printing, or the number of simultaneous users. Publishers tend to have fewer DRM restrictions than aggregators, or even none at all.
Third, the features provided for users – and how they operate – can vary significantly. Some platforms provide full-text searching, and the ability to highlight and save search terms; some allow users to create a bookshelf where they can save the books they are interested in, and bookmarks and notes within books (which may or may not include facilities to download or print the notes); some have a citation facility which provides a one-click citation to a book or chapter (and there may be facilities to export citations and references). But not all platforms provide all these kinds of features, and even when they do, they operate in different ways. Features to enhance accessibility for readers who, for example, are partially sighted or dyslexic, also vary.
Fourth, the content on different platforms varies in how it displays in different browsers, and libraries need to take this into account in organising their support for users. Accessibility on mobile devices varies even more.
Fifth, the administrative interfaces vary hugely along with the facilities that allow libraries to control authentication and authorisation methods; or to set loan periods, or change default landing pages, search settings and the like; or to ensure that the content is discoverable via the library catalogue or otherwise(such as tagging titles within web-based discovery services), and to configure link resolvers; or to get COUNTER-compliant usage data (along with the formats in which reports come).
All these variations make for potentially-daunting problems for libraries, and difficulties for users as they get access to content from different platforms. More standardisation of the facilities provided by aggregator and publisher platforms would be very welcome to most users. The perception that, rightly or wrongly, institutional access to the major journal platforms is more straightforward is one shared by many, it would seem, and helps to reinforce patterns of usage behaviour which in this context can only be to the long-term detriment of the continued publication health of long-form academic research, howsoever sourced and howsoever supplied.
One of the most widely-cited comments in scholarly book publishing in recent years has been the pithy remark of the senior US-based industry consultant Joe Esposito, that ‘the problem with the academic monograph is not lack of access but lack of demand ‘. Much of this brief report has been devoted to showing how ‘access’ and ‘demand’ are in fact entwined in all sorts of ways of which most scholars and indeed most university administrations are blissfully unaware, and that the relationship of discovery, access and demand has been made more complex, not less, by the advent of digital publishing technologies. The complexity has been further compounded by rising levels of new-title output, especially from the major UK-based publishers active in the arts and social sciences: the role of successive UK Research Assessment Exercises, and its successor Research Excellence Framework, in incentivising the increase in such outputs is unquestioned, and achieving any kind of equilibrium between the supply of new long-form research and the capacity of individuals or institutions around the world to consume same (whether published conventionally or in Open Access forms) has proved, thus far, impossible.
In the institutional context, the demand amongst librarians for book access workflows that are simple, flexible and accessible, and yet at the same time sufficiently technologically advanced to reap the potential rewards presented by new publishing formats, has proved (again) very hard indeed to meet effectively, not least because of the simple multiplicity of publishers, intermediaries and titles, certainly as compared to the serials world. Many of those publishers and players in the humanities and social sciences are, as we have seen, small and often undercapitalised to meet the challenges arising, with consequent frustration on both sides. Publishers, for their part, find frustrating the perceived lack of technological commitment by some institutions and some of their library staff, and see a failure within some libraries to generate demand for and facilitate access to materials the former have themselves obtained, whether in traditional or Open Access forms. These tensions show no immediate sign of resolution, and probably compound those existing pressures on the successful publication of long-form research of which all parties in this particular research sector are well aware. The intermediary actors we have encountered in the course of drafting this report see a complex and fragmented global picture, and one in which (as often in the academic domain) some of the most important and determinant players (notably Amazon) no longer see the academic sector per se as a determinant business imperative, with important disruptive consequences.
In sum this report is not, its authors would hasten to note, an investigation into a dying area of scholarly communications, although it is emphatically one confronting very different, and in some ways starker pressures than those at play in (say) scientific serials. Articulating, understanding, clarifying and then (perhaps) simplifying some of the relationships and some of the networks that we have described is an important agenda for all of those involved in long-form scholarly communications, and we hope that this brief summary provides evidence and context for all of those involved, and particularly perhaps the research community itself.
Ken Rhodes, Executive Director, Rowman & Littlefield International/NBN International (29th February 2016)
Frank Smith, Director, Books at JSTOR (29th February 2016)
David Taylor, Senior Vice President, Ingram (7th March 2016)
Peter Lake, Group Business Development Director, John Smith (7th March 2016)
David Prescott, Chief Executive, Blackwell’s (7th March 2016)
Ben Ashcroft, Vice President, Sales and Marketing, de Gruyter (9th March 2016)
Michael Zeoli, Vice President, Content Strategy & Publisher Relations
YBP Library Services
(9th March 2016)
Christoph Chesher, Group Sales Director, Taylor and Francis (14th April 2016)
Kari Paulson, Vice President Market Development, ProQuest and Anna Bullard, Senior Director, Content Strategy, ProQuest (14th April 2016)
Lara Speicher, Chief Executive, UCL Press (14th April 2016)
Frances Pinter, Chief Executive Manchester University Press and Founder, Knowledge Unlatched and Christina Emery, Partnership Manager, Knowledge Unlatched (17th May 2016)