The characteristics of the journals business

Traditionally, journals have been a highly profitable sector of publishing. Whereas there is often considerable uncertainty as to how many copies of a book will be sold, journals sold on subscription exhibit a high degree of reliability for sales forecasting as well as predictable cash flows. The Economist declared: ‘If a company owns the must-read title in, say, vibrational spectroscopy, it has a nice little captive market’ (5 December 2001). Fixed costs tend to be higher than in other publishing sectors and, given that it takes time to establish a journal’s reputation, this means that the period necessary for a journal to break even can extend up to five to seven years. However, the feature of high fixed costs also underpins the sector’s profitability since once those costs are exceeded the profit margins can be extremely healthy (up to 30 per cent).

Journals have tended to be less price sensitive than some other forms of publishing. The Morgan Stanley research report on the industry in 2002 said: ‘The scientific journal business is characterised by relatively inelastic demand, with individual journals generally having a strong following within their particular niche … since 1986 the average price of a journal has risen by 215% while the number of journals purchased has fallen by only 5.1%’ (Morgan Stanley Equity Research, 2002: 2). This was also noted by Page et al. (1997: 281), who wrote: ‘it appears that the price of a journal has only a very small effect on the sales of that journal’. A reflection of the sector’s overall profitability is that journal publishers tend to be bought and sold for relatively high multiples of the sales revenues. As Eric de Bellaigue commented on the purchase of the academic publishing assets (with a multiple of four times sales) of Wolters Kluwer by Cinven and Candover in 2002: ‘it is a vindication of the attractions of a business that is cash generative, judged to be relatively immune to economic fluctuations’ (de Bellaigue, 2004: 222). However, tighter institutional budgets and greater resistance to automatic price rises mean that the sector has become more price sensitive in recent years, and large increases may attract ‘name and shame’ campaigns on library listservs or in social media.

Another characteristic of the journal business is that many companies have moved towards a service model as opposed to a product model. Subscribers have 24/7 access, remote access, sophisticated search facilities and access to aggregated content. Online publishing means that publishers know who their customers are and can measure levels of satisfaction with the service offered, whereas previously with print publications they dealt with key intermediaries such as the subscription agents. Many journals also use online submission systems to increase efficiency.

Journal publishers also have to offer an efficient service to their authors. Typically, no payments are made to authors for their contributions to journals, or to the reviewers of individual papers. The editors of many journals are paid an honorarium and receive a contribution to their office costs. Traditionally, page charges were asked of authors (in the expectation that the fee would be covered by their institution or from a research grant), but this practice is much less common than previously. Learned societies may publish journal titles themselves (for example, the American Chemical Society) or subcontract to a publisher (for example, the Political Quarterly is published by Wiley).

It is estimated that over 90 per cent of English-language journals are now available online: 87 per cent of journals in the humanities and social sciences (HSS); and 96 per cent of STM titles (Cox and Cox, 2008). Many libraries still prefer to have print copies as well as online access, partly for security of archiving, but over time there has been movement towards online-only subscriptions. There are also differences by subject field and in the area of HSS there is a much stronger print tradition. Considerable effort has been expended on digitizing the backfile of journals, enabling librarians to clear their shelves of print copies. Users now expect digital access, and with libraries keen to save on shelf space there are benefits to all parties. It was estimated that ‘the profitability of a customer improves by 15% as they transfer from paper and on-line subscriptions (most pay for both currently) and opt for just on-line access’ (Morgan Stanley Equity Research, 2002: 2). In the UK and other EU countries the issue is clouded by the addition of VAT to online products and print/online combined products.

The life cycle of a journal

There are risks to the launch of a new journal and it may take several years for a new journal to break even. At present, the area of HSS is seen as attractive for the launch of new journals, with some parts of the sciences seen as relatively saturated. A new journal concentrating on a smaller subject area may spin off from a title with broader coverage. A strong brand (Nature is a good example) may help the launch of other titles. Ulrich’s data for the period 2000 to 2004 shows that 1925 new STM journals were launched, of which 1850 were still active in the Spring of 2007 (Amin, 2008). If data from 2005 to May 2007 is examined, a total of 1256 STM journals were launched. The top ten publishers accounted for 56 per cent of the journals started in this period; in total, 120 publishers were involved in journal launches.

Typically, it may take five years for a journal to become profitable – cover its costs – and longer to provide a return on the initial investment. The break even point is taking longer to achieve than previously, with the average period having increased from three to five years to five to seven years, as it takes longer to establish new titles in the market (Bannerman, 2008). ‘The third year is often a critical milestone in terms of both sufficient subscriber growth and high quality article submissions to enable publishers to assess the title’s likely success or failure’ (Electronic Publishing Services, 2006: 56).

There is considerable concentration in the publishing of journals: the ten largest companies publish 35 per cent of journals, and the top 100 companies publish 67 per cent of titles. More than half of journal subscriptions are now bundled up within a collection of titles (Ware and Mabe, 2012). Edlin and Rubinfeld comment on this concentration of ownership in the industry and the development of the ‘Big Deal’, whereby libraries enter into a long-term contract in return for a large collection of journals: ‘Bundling can be seen as a device that erects a strategic barrier to entry. At a simple level of analysis, the Big Deal contracts leave libraries few budgetary dollars with which to purchase journals from new entrants’ (Edlin and Rubinfeld, 2005: 441). There are definite advantages to bundling in terms of the range of journals offered, but in the view of librarians it then becomes difficult to vary or withdraw titles (Fowler, 2008). In universities, courses may change or may even be dropped from the course portfolio, but under the terms of their agreement with publishers the journals portfolio cannot be altered easily to match these changes. Withdrawing the whole bundle would be deeply unpopular, as the journals have entered the catalogue and users have expectations that these titles are available. In any case, libraries will not be able to adopt new journals immediately as it will take time to alter their budgets. Taking on a new journal often means cancelling an existing subscription. The open access model does provide a way to break into a market with a new title, and this advantage can be mimicked by publishers offering new titles for free within a larger bundle for the first two years. But librarians are faced with another difficulty regarding open access journals – which ones should they catalogue? How much time and expense should be invested in this process? In 2013, nearly 10,000 open access journals were listed in the Directory of Open Access Journals (DOAJ).

Typically, journals are reviewed five years after launch to assess their long-term viability. They will be assessed on their intellectual health – e.g., usage rates, impact factors, rejection rates – as well as on their financial health. Those journals which are made available as part of bundled deals, in addition to separate subscriptions, may be able to survive on a circulation of a few hundred subscribers.

Pricing

The price of an individual journal when newly published will be set in line with the market, for example with competing titles. The publisher has to adopt the most effective strategy to get the title established and build market share. Over time, however, what happens to journals pricing? If demand is price inelastic – less sensitive to price changes – then publishers can simply push up prices in response to cost pressures and maintain their profitability.

The 2008 Research Information Network report suggested that:

As institutions have opted to buy bundles of journals, they have become interested in the value gained from a subscription, such as the number of titles included and the terms of access. Publishers would argue that increased value has been provided as titles are bundled together, page extents have increased and search functions have been greatly improved. On average, the price paid by a library in the UK for a journal is only one-tenth of the list price (see Chapter 13), and, overall, journals are much more readily available. A journal that would have been bought by about 500 libraries in the early 1990s is now available in about 7000 libraries (Campbell, 2008).

The journals industry saw large price increases up to the turn of the last century: ‘throughout the 1980s and 1990s, the prices of journals were increased year on year well above the rate of inflation. The increases were particularly dramatic in what is broadly described as the STM (scientific, technical and medical) field, but the cost of journal subscriptions has increased significantly in the social sciences and humanities as well’ (Thompson, 2005: 99). Tenopir and King (2000) reported that the average price of a journal was US$284 in 1995 compared to US$39 in 1975 – over seven times higher; when adjusted for inflation, the price increase was a factor of 2.6.

Looking at more recent pricing changes, in the UK between 2000 and 2006 journal prices rose by 39 per cent, while inflation during this period was 16 per cent. ‘On average, periodical prices have risen faster than inflation in the UK; however this is a very crude measure, taking no account of the relative value in terms of journal size or quality’ (White and Creaser, 2007: 4). Niche journals tend to have higher prices (arising from a lower subscriber base); higher-circulation journals will have lower prices and are more likely to have advertising income.

There are pricing variations between disciplines, reflecting differences in funding and the necessity for scientific researchers to have the latest research. Table 5.1 shows price differences between a selection of academic disciplines, and suggests that the rate of price increases has been slowing in many subjects.

Consolidation among publishers in the journals market, and the price inelasticity of the market, did give the publishers considerable power to raise their prices. By way of response, the development of library consortia has strengthened the hands of the purchaser in any negotiation over pricing. Price increases on bundles may be capped for a certain period. The open access debate has also cast a spotlight on journals pricing and the value offered. Librarians have a greater set of metrics at their disposal, including usage rates, when considering the value to their institutions of bundles or site licences. The variation in pricing of journals between commercial publishers (higher) and non-profit publishers (lower) has been documented, but as King and Alvarado-Albertorio note (2008: 262), this does not take account of the value given by the bundled deals put together by the large commercial publishers. Examining the price of an individual journal has become less relevant if it is a title that is commonly available as part of a bundle. Publishers would prefer to concentrate on other measures – such as the price of each user download, which has been falling as usage has risen sharply. At Elsevier the average price paid per article download in 2008 was around US$2 (Amin, 2008).

Budgetary difficulties for libraries, exacerbated by the generally poor economic conditions during the financial crisis of the late 2000s, have cast a further spotlight on the price rises in journals publishing. Bosch et al. commented in their 2011 survey of periodicals pricing: ‘There’s no way to sugar-coat the impact higher serials prices have on the information marketplace, or the dire state of funding for libraries. Libraries are no longer in a position of having to cut low-use journals in order to make room for high-use ones; instead, they are now being forced to cancel heavily used, even essential subscriptions, much to the dismay of their patrons. The economy still drives any discussion of serials pricing, and it remains a very ugly story’ (Bosch et al., 2011).

Cost structure

The costs of publishing a journal can be divided up into fixed and variable elements. The fixed elements, which will not vary according to the number of subscribers, include editing, the costs of peer review, typesetting and the costs of an online platform. At the Canadian not-for-profit publisher, NRC Research Press, the costs of copy-editing rose by 160 per cent between 1986 and 2001, and the costs of peer review by 117 per cent (Holmes, 2004). At Taylor & Francis, their fastest-growing cost line is payments back to the academic community, to support office and other costs for journal editors. This has risen faster with the increased use of full economic costing by universities (Bannerman, 2008). Some publishers have achieved cost savings by outsourcing the production route to external suppliers, perhaps in India, who carry out a variety of activities including copy-editing, typesetting and project management. Variable costs, which are dependent on the number of subscriptions to a journal, cover the printing and mailing of print copies and any extra costs associated with additional electronic subscriptions. With online delivery, variable costs tend to be lower (partly because the cost of printing, if any, is shifted to the end-user), but the fixed costs are significant. An online environment requires skilled IT staff, an enhanced sales function and customer service that supports 24/7 delivery. Again at NRC Research Press, five support staff were taken on over time as the publisher moved to electronic delivery; additional to this comes the purchase (and replacement) cost of hardware and software. It is reported that ScienceDirect, the online platform from Elsevier, cost around £200 million to establish (Clarke, 2007).

In journals publishing, the proportion of fixed costs is high, with estimates of between 60 and 80 per cent. Over the long term this is likely to lead to higher profits, once a title has become established. But if the number of subscribers starts to fall there is a greater risk of a loss being made, and prices have to rise. Fixed costs are independent of the business model.

Fixed costs are also called first copy costs and there has been much discussion over their level per article. This has become important to the debate over open access and the level of fee that needs be charged in the author-pays model of publication. Any fee charged to an author is not necessarily the same as the true cost of publishing an article. There are differing estimates for the level of first copy costs – from between US$250 to US$2000 per publisher article, up to around US$4000 or even US$8000 (King, 2007; Wellcome Trust, 2004). The variations can be explained by a number of factors. For example, the costs of peer review will be higher in a high-quality journal with a high rejection rate (up to and exceeding 90 per cent). Again, high-quality journals will have higher editorial and administrative costs, and each issue may include more editorial copy, comment and news. A Research Information Network report suggested a figure of £1136 for the average first copy publishing cost (if the non-cash costs of peer review are included this rises to £2331) (Research Information Network, 2008: 35). A subsequent JISC report concluded that the per-article costs could be broken down as follows: writing, £5300; peer review per published article, £1400; library costs, £1.11; and publisher related costs, £2900. The total of £9600 compares to a figure of £88,600 for a research monograph (JISC, 2009: xiv). A recent estimate suggests a per-article cost to a commercial publisher of between US$3500 and US$4000 (Van Noorden, 2013).

Moving to purely online delivery, without the costs of print publishing, will reduce the costs of publication. The JISC report suggests ‘an average publisher cost of around £3247 per article for dual-mode production [print and online], £2278 per article for print production and £2337 for e-only production’ (JISC, 2009: xv). For some journals struggling to remain viable, this is a possibility, and certainly the market has moved on in its expectations. Largely left behind in Western markets are the previous doubts about online-only delivery, summarized by John Thompson in his 2005 book on academic publishing:

Some markets, for example the Australian market, are more receptive to online-only publication; the situation in the UK is complicated by the VAT applicable to digital, but not print, products (Bannerman, 2008).

Subscription model

The predominant business model in journal publishing has been one of subscription. The majority of subscriptions are taken out by institutions rather than individual subscribers, and there has been a steady decline in the take-up of individual subscriptions, which are price-sensitive (Tenopir and King, 2000). The major development in recent years has been the bundling of subscriptions together – in order to offer digital access to a number of titles – and the availability of site licences to institutions. This so-called ‘Big Deal’ has provided subscribers with a range of content, and publishers have sought to add value to these bundles by increasing the number of titles and pages available. Academic libraries acquire more than half of their journals in bundles of more than 50 titles (Primary Research Group, 2008), and bundles have been described as ‘a near ubiquitous feature of the research library collection’ (Hahn, 2006). In 2006, 93 per cent of US research libraries held bundles with at least one of the top five publishers, and three publishers (Wiley, Elsevier and Springer) had over 70 per cent market penetration. In this market, 82 per cent of bundles were acquired through consortia. The subscription model has the considerable benefit for the publisher of predictable advance income. For the subscriber, working within a fixed budget, there is also a predictable cost – by comparison, for example, to a pay-per-view model.

Yet Toby Burrows (2006: 172) talks of bundling as a ‘Faustian bargain’ with a trade-off between access to a higher number of journals and greater rigidity in the system. He writes:

More recently, some libraries have resisted elements of the ‘Big Deal’; for example in March 2011 Cornell University Library announced that it would no longer sign contracts which included confidentiality clauses (Cornell Chronicle, 2011). Libraries have also looked for additional elements of flexibility within bundling agreements around the cancellation of titles.

Alternative business models

What alternatives are there to the traditional subscription model for journals? Some publishers already operate pay-per-view (PPV) services and this can provide useful additional income. In theory, increased online access should mean there is less need for users to purchase articles individually, but this market remains healthy. For those without institutional access, there are services such as those offered by DeepDyve, which provides ‘affordable access to millions of articles across thousands of peer-reviewed journals’ (http://www.deepdyve.com). Users searching on Google may come across articles not readily available to them and decide they are worth the one-off purchase fee. For most publishers this income will rarely exceed ten per cent of revenues, and Ian Bannerman of Taylor & Francis commented: ‘We are not quite in iTunes territory’ (Bannerman, 2008). For the Dutch publisher Brill, the income is less than two per cent, perhaps suggesting less demand in the area of HSS. But could a more systematic PPV model enable libraries and other institutions to monitor journal usage rates and assess better which journals should be kept and which should be weeded out? Would it be a welcome antidote to the continuous growth of journal collections? Or does the subscription model continue to win out for its predictability? JISC-sponsored trials in 2006 looked at the possibilities of PPV. The final report concluded:

The model of patron-driven acquisition is a development of PPV, and is used in the area of ebooks in particular. Titles are made available through the catalogue – but only those books that are used are triggered for purchase, ensuring that the library collection meets the needs of its stakeholders. This system could be used with a journal collection with certain caps on expenditure set by the library – by period or subject field.

Advertising is another source of income for some journals. A prominent scientific journal such as Nature has the ability to command substantial revenues, but for most publications the opportunities are more limited. None of the leading publishers has advertising revenues that exceed nine per cent of total income (Outsell, 2007: 4). This has the advantage of shielding journals from the dips in advertising income that affect the fortunes of the magazine market, for example. On the issue of advertising, Morris et al. (2013: 229) state that ‘periodicals that deliver a highly targeted audience to advertisers will rely on in-house generated or non-peer-reviewed content, although there are some peer-reviewed journals that are totally supported by advertising’.

As online advertising revenues grow, are there new opportunities for journal publishers? Potentially, subject portals could attract advertising revenues (although the level of income will vary according to the discipline). PracticeUpdate is an online resource from Elsevier in the area of healthcare. Access is free once a user has registered, and revenues come from online advertising, sponsorship and education grants (http://www.practiceupdate.com). Some are doubtful about the possibility of simply transferring print advertising online.

With journals moving steadily towards online-only publication, the option of print on demand remains to provide additional revenues or to aid start-up journals. Themed collections of journal articles may be given ISBNs and published in print or as PDF downloads. Publishers have also used print on demand to lower the investment risk of starting up new journals (Wilson-Higgins and Bernhardt, 2008).

Among other innovations was the use of Second Life by Nature, the aim being to make use of the virtual world to hold events and market virtual meeting rooms and conference facilities. From 2006 to 2009, Nature Publishing Group ran more than 50 events in Second Life, including public lectures, conferences and film premieres, but the experiment was concluded in 2010.

The society model of publishing operates among learned societies, with organizations either self-publishing or subcontracting to a publisher. Publishers have to prepare pitches to societies to secure their business, with a new bidding process at each renewal. Some societies rely on journals for a high proportion of their income.

Open access

The arrival of open access has formed a significant challenge to the status quo in journals publishing, posing a threat to the reliable business models of journals publishers. Open access can be viewed as a profound philosophical challenge or as just another business model. Experimentation with open access continues by both new entrants and established players, and support from research bodies and policy makers is encouraging rapid growth. The UK’s Finch Report (2012) recommended the adoption of the gold route of open access, which was subsequently adopted by the government and Research Councils UK (RCUK). The EU published a Communication in 2012 mandating all research output from their billion-euro 2020 Horizon research programme to be published in either green or gold open access; and, in the US, all government agencies with a research budget of over US$100 million were instructed to design an open access policy which seems to be heading for the green route.

Viewing open access from a philosophical perspective would lead naturally to favouring a not-for-profit model, which should mean lower costs. There are lower transaction costs involved, even if under the author-pays model payments have to be collected from authors (with the risk of bad debts). Although a surplus may be required to supply investment funding, there is not the necessity to provide a return for the owners of the company. In terms of the income that open access journals attract, this can come from the authors, their institutions or the relevant research funding bodies. The attraction of author payments as a business model is that break even should come much sooner, rather than having to wait for a title to command sufficient subscriptions. However, an author-pays model may put off some contributors, and as Claire Bird comments: ‘The journal business model is therefore turned on its head, from “reader-side payment” to “author-side payment”. It seems immediately clear that this kind of model may only be viable in certain disciplines where authors have access to funds for publication’ (Bird, 2008: 200). Funding streams are likely to provide more support for this model of open access in the sciences rather than in the humanities. There may be cultural differences between disciplines, as Mary Waltham comments: ‘Within certain disciplines there may be some resistance to shifting to a producer pays model because of enduring scholarly traditions and/or questions of quality’ (Waltham, 2006: 94). The overall rejection rate in HSS journals is also much higher than in STM journals, which increases the costs of peer review: studies have found an acceptance rate of only 11 per cent for HSS journals, compared to 42 per cent for STM journals (Waltham, 2010).

There are a number of variations on open access from the gold route (funding through article processing charges [APCs] from a research body or from the author’s institution) to the green route (where the author self-archives with the article, as pre-print or post-print, appearing in a subject-based or institutional repository). Some journals may be fully open access; commercial publishers may operate a hybrid model, through which the author can pay for open access on publication, with their article appearing alongside content that is not open access. There may be a discount on the payment due from the author if he or she is based at an institution with a subscription to the journal. The subscription price of the journal may also be reduced as open access content increases. The current implementation of the gold model in the UK requires a system to be established to process considerable numbers of APCs, and raises challenges for research institutions around workflows and the decisions as to who is awarded funding.

The number of open access articles published is growing fast, with an estimated 12 per cent of the 1.7 million articles published worldwide in 2011 published in full, immediate open access journals (Laakso and Björk, 2012). Within this percentage, hybrid publication only added around one per cent, suggesting its limited importance. After a period of six months, a further five per cent of the total articles published had become open access. By contrast, estimates of the income from open access journal publication in 2011 put the figure at US$128 million – only 2.2 per cent of the overall market (Outsell, 2013). The average APC was US$660 – compare this to the higher per-article costs cited earlier in this chapter.

Most commentators think it is too early to say how effective the open access model will become. As open access journals become more successful, it is entirely possible that their costs will rise accordingly with the costs of peer review. The rise of the mega journal and cascade models (see Chapter 9), however, suggests otherwise, with efficiencies being found in the peer-review process. These might be through the acceptance of papers simply if their methodology is sound, or through allowing publication in another journal within the same stable. For the moment, there is evidence of falls in income for publications that go fully open access, and this has persuaded many publishers to opt for the sponsorship model, where authors pay for open access at the point of publication. Membership schemes, such as the one run by BioMed Central, involve institutions paying funds upfront into a deposit account – then APCs are covered from this account with suitable discounts. Launched in 2012 for journals focused on the biological and medical sciences, PeerJ provides membership to individuals which gives them lifetime rights to publish – and the most basic form of membership costs only US$99 (https://peerj.com). Another sponsorship model is to gather support from institutions – and having a number of partners will spread the cost and minimize the potential competition – which offer financial contributions and editorial support from their staff. As an example, the Journal of Archaeology in the Low Countries was started in 2009 with nine partners contributing €2000 each, plus their support for peer reviewing and the editorial board.

The research company Outsell commented on the open access business model:

Traditional publishers would respond that in the face of this flood of free content, it is ever more important for branded, peer-reviewed journals to offer quality assurance to their authors and readers.

Future of business models

For the moment, the most sustainable business model for journals remains subscription. Advertising does not form a significant source of revenue and it poses difficulties for publishers who wish to move to online-only delivery, since the business may not simply transfer online. PPV does not offer the budgetary reliability of the subscription model, neither to the publisher nor the customer. Costs can be reduced by moving to electronic-only publication, and this has been a growing trend, as can be seen with the new journal start-ups. Johnson and Luther predict that: ‘As the opportunity cost of continuing to invest in print becomes too great, online will be the growing focus of publishing processes. Except for top-tier, broad circulation titles – which sometimes are used more like magazines – surviving printed editions may become mere add-ons available via print on demand’ (Johnson and Luther, 2007: 2).

Open access models have a number of variants. There remain questions about the financial sustainability of these models, especially outside the relatively well-funded disciplines of science and medicine. Commercial publishers have experimented with hybrid models but their success remains limited. However, there is no doubt that open access can provide marketing benefits, for example to aid market penetration for a new journal or to encourage usage of subscription journals. Open access publishing also serves to support the ‘long tail’ of journals publishing – those niche journals which cannot survive as commercial publications.

A possible future trend is a reversal of the bundling model, or at least the offer of greater flexibility by publishers. When licensing ebooks, librarians are keen to avoid a similar structure, preferring to go with a particular platform and then choosing which titles are relevant.

Theodore Bergstrom asked why profits remain high in the journals industry: ‘There is free entry to the journal publishing industry. Libraries are not compelled to subscribe to expensive journals, and scholars are not compelled to write for them, referee for them, or edit for them. Why has competition not driven profits to zero?’ (Bergstrom, 2001: 189). Part of the answer, as he himself acknowledged, is the strength of the reputation of the established journals. Authors would prefer to publish in them, and scholars and professionals want their institutions to subscribe. This is reflected in the time it takes to establish a new journal and for it to become profitable. Bergstrom argued that the ‘academic community is stuck in an equilibrium where it will continue to pay huge rents to owners of commercial journals’ (ibid.: 192).

How will open access affect business models? An increase in self-archiving, for example, might lead to a diminished demand for subscriptions if students and researchers can find articles from other sources; or there could be a mandate to ensure journals make articles freely available after a certain period. A study in 2012 for the Association of Learned and Professional Society Publishers and the UK Publishers Association, among senior librarians internationally, suggested that an overarching mandate with an embargo period of only six months would have a significant impact on subscriptions. Libraries would expect to scale back their level of subscriptions, and ‘STM publishers could expect to retain full subscriptions from 56% of libraries; AHSS publishers could expect to retain full subscriptions from 35% of libraries’ (Bennett, 2012: 4). The journal Annals of Mathematics went green open access and discovered that one-third of its subscriptions disappeared over a five-year period; it retreated to a subscription model in 2008 (http://annals.math.princeton.edu/). So far, gold models see higher APC charges in hybrid journals rather than in pure open access journals, and among journals with strong brands. A market is developing, and a study of pricing in 2011 found a variation between US$8 and US$3900, with the lowest prices charged by journals from developing countries and the highest prices charged by journals with high impact factors from the major publishers (Solomon and Björk, 2012).

The open access model has certainly provided a challenge to the equilibrium in the industry, but so far has not fundamentally shaken its structure. It is certainly less of an option in subject areas that are not so well funded, for example in the arts and humanities. Many publishers assert that the health of the journals industry is a reflection of the commercial discipline and entrepreneurship within the industry. The continued consolidation of the journals market suggests that this remains an attractive sector of publishing, and that larger players can benefit from growing even bigger. Further growth can be anticipated with the strength of newer markets such as China and India. The relatively high fixed costs of journals publishing mean that if companies can spread those costs over a larger number of titles they will increase their profitability – this is a driver for further acquisitions. In the area of pricing, the industry has moved away from pricing and selling a physical product (in print form) towards pricing a service on the basis of the value offered to journal users. Customers also have a fuller set of metrics with which to negotiate. With better data available to customers, commercial publishers are finding it harder simply to increase prices to increase profits – they have to prove that they are offering greater value for money.

In conclusion, a variety of business models and continued experimentation are desirable for the health of the sector. As Brown et al. (2007: 4) wrote about academic publishing: