Open access (OA) at Taylor & Francis Group
Taylor & Francis Group is part of Informa PLC, a multinational publishing and conference company2. Taylor & Francis (T&F) publish some 1,850 journals, two thirds of which are in social sciences, arts and humanities (SSAH) subjects (published under the Routledge imprint). It also has thriving books divisions operating globally. Over 500 journals are published on behalf of societies and institutions who own their titles, and there are a number of regional co-publishing partnerships as well. Finally, we must also acknowledge the shareholders that we serve, as a commercial publisher. We are responsible to all of these groups for ensuring that our own and our partners' journals remain sustainable. Sustainability has become somewhat of a hackneyed phrase in recent months, but perhaps this is because it best expresses the main preoccupation of those publishers who are part of the current ecosystem, ensuring that we adapt to this new age without jeopardizing the future of the journals that we publish.
We have had a formal Author Rights charter in place since 20033, clarifying polices around the archiving of author versions of manuscripts (‘green’ OA). There are currently embargo periods on archiving the peer-reviewed ‘author accepted manuscript’ (AAM), usually of 12 months for journals in science and technology fields, and 18 months for SSAH.
To contextualize some of the subsequent discussion, our ‘gold’ OA content comprises less than 1% of our total content. Gold OA here refers to publication of the ‘version of record’ (VoR) within a journal as OA, often but not always through payment of an article publishing charge (APC). This is probably due to the subject mix of the T&F journals portfolio. We currently have no titles in the clinical medical sciences, which are published by our sister company, Informa Health Care4; and as noted above, the majority of journals are in SSAH fields where funding for APCs is limited, if it even exists. To a certain extent then, we are behind the curve of certain other large commercial publishers, who have seen a greater influx of content being published on a gold OA basis. Despite this, T&F has seen its paid gold OA increase by 375% in 2013, compared to the previous year.
Developments around OA, especially within the UK, happened more rapidly than we had anticipated. We have referred to this as a ‘seismic shift’. In the months following the release of the ‘Finch’ Report5 and the Research Councils UK mandate6, we agreed upon a number of common-sense philosophies around OA. First and foremost is that OA is something to embrace. It has presented some significant challenges, but these have encouraged the staff within T&F to review long-maintained systems and processes and reflect on how best to improve, refine or even redesign these systems and processes.
“We have referred to this as a ‘seismic shift’.”
Another core philosophy is that the choice of publishing behind the subscription paywall (with green OA postings available after designated embargo periods), or paying for gold OA, so that the VoR is immediately freely available online for public access, should remain with authors. We have invested time in improving our communications with authors to clarify their options around gold and green OA (the latter being allied with the traditional subscription route). After consultation with editors and society partners, we enrolled the majority of titles in our hybrid OA programme, Open Select7, in early 2013 to facilitate this choice. Over 90% of journals are now compliant with the gold and green OA mandates of major policymakers and funders, including Research Councils UK (RCUK). Out of those journals which are not part of Open Select, some are review journals that contain commissioned papers, and others are owned by society partners who have chosen to opt out of Open Select due to concerns about the gold OA model in their subject communities.
Operational and strategic challenges
The changes are manifold: a shift to the author as customer (who is buying a product and services around publication of their article) and a concomitant change in legal status of the author; the need for micro-payment systems that also offer record-keeping and compliance reporting back to institutions and funders; new processes of article production and marketing services; and, not least, the introduction of new licensing and article publishing agreements that are legally fit for purpose in an open publishing world.
In the subscription world, about 85% of transactions are administered by subscription agents, or mediated through institutional, consortial or national sales deals. OA is driving an increase in the number of complex micro-payments, usually between the publisher and a funder or institution, but increasingly between publishers and individuals. There is thus a shift in trade occurring – from a commercial relationship with an institution to one with an author for the supply of services.
This shift from macro to micro is reflected in the impact OA has on our production workflows. In particular, hybrid OA, where the VoR is published on an OA basis within a subscription journal, creates increased complexity. If an author chooses gold OA for his/her final published article then metadata must be updated, copyright information altered and information on funding and data captured, and so on. Particularly important is heightened clarity for and recording of rights clearance of third-party copyrighted material within the work. This demands a significant level of manual intervention at the level of the individual article.
“OA is driving an increase in the number of complex micro-payments …”
This increased granularity creates more cost, as there is less scope to introduce journal-level workflows which allow for economies of scale. Headway is being made on this front, particularly with the advent of industry standards such as ORCID8 and FundRef9, but it will take time for the benefits to be felt. The obvious solution may seem to be to convert more journals to the pure OA model. Although in theory this seems sensible, it would make many smaller journals unsustainable, unless either they introduced submission charges or dramatically increased their output, potentially impacting quality as a result. Some ‘small’ journals in SSAH in fact have rejection rates of over 95%.
The processing of APCs has introduced a level of complexity that has not been part of the subscription environment for a number of years. We have worked with RightsLink10 since we first began to offer a hybrid OA option (now Open Select) back in 2006. In late 2012, it became clear that we needed an additional service which could offer adjusted pricing, and could handle APCs for our pure OA titles (which have differing APCs). This was allied to our creation and rollout of OA membership options for funders and institutions. There are a number of intermediaries offering (or due to offer) an APC management platform, but we felt that we didn't yet have a clear enough grasp of what we and our customers would need from an APC service, and also that the intermediary market was yet to mature. Subsequent discussion led to the decision to create an internal APC team, managed by our Customer Service department. Being a manual process, it is easy for us to adapt to customers' requirements; however there is a high level of internal cost in terms of staffing and administration. Nonetheless, this has been a very informative exercise and will inform our decision making and investigations into more automated systems.
“… traditional copyright assignment grants publishers the right to reuse an author's work …”
There are a number of dichotomies that we have been dealing with. Licensing is one such example. The Creative Commons Attribution11 licence is the licence of choice for funders such as the Wellcome Trust and RCUK, as it facilitates maximum reusability of content. Many authors, especially in SSAH subjects, have reservations about this licence, particularly when it is mandated. One such reservation is around the loss of control for authors over how their work is reused, though it is worth noting that traditional copyright assignment grants publishers the right to reuse an author's work, often for commercial gain. Another concern is around the notion of intellectual property (IP). The IP of an SSAH scholar will be the ideas contained within an article, whereas a biomedical researcher may write an article to describe a product which has already been protected by patent (for example a new drug).
It is the case that no T&F author is required to assign copyright to the publisher when publishing an article with us. Although this is our preference, any author is able to retain his/her own copyright and simply grant us a Licence to Publish, or even just sign an author agreement giving us first publication rights if they wish the work to be licensed out on a Creative Commons basis when they pay an APC.
As outlined above, our own green OA policy has been in place since 200312 and, whilst we are supportive of this model, we are mindful of the risk that it could pose. We support the principle of evidence-based decision making and in November 2011 launched a ‘zero embargo’ pilot, allowing authors contributing to 37 library and information science journals to post the AAM in their institutional repository, subject repository or personal website immediately following publication. These authors also retain copyright of their article13.
Our aim was to monitor the effect on usage and subscriptions for these journals. As of November 2013, the effect has been nominal, but this is also due to low deposit rates amongst the author community. We are extending this pilot into 2014 and will be promoting it much more heavily, and actively encouraging authors to deposit their AAM. It seems that, left to their own devices, many authors do not wish to exercise the very liberal posting rights granted under this scheme.
We are also contributing data on article usage half-life to a study commissioned by the British Academy14 (usage half-life being the period at which a typical article has received half of its lifetime downloads). This study is gathering data from a number of different publishers on usage half-lives, where these measures are typically being used as a proxy to anticipate the likelihood of librarians renewing subscriptions to journals whose article usage half-lives are lower than an embargo period.
Our own data are showing some very interesting phenomena, that we aim to investigate further:
- usage half-lives vary dramatically within subject areas – different journals in the same subject have quite different profiles
- life sciences and materials science have on average the shortest usage half-life profiles in our STM portfolio, and philosophy(!) has the shortest average half-life among our SSAH titles
- physical sciences and maths have on average long usage half-lives, similar to many SSAH subjects
- quite different half-life profiles are obtained if you analyse the same set of titles at different months of the year, or over different years, leading us to want to investigate further the ‘stability’ of journal half-lives over time
- the existence of extensive online back-file archives will extend some journal half-lives dramatically.
“… collaboration has been very influential in shaping our understanding …”
There have been numerous meetings, phone calls, workshops and focus groups with editors, society partners, librarians, institutional representatives, researchers, funders and policymakers over the course of 2013. This collaboration has been very influential in shaping our understanding of the needs of these different groups, and we hope it will continue into 2014 and beyond. During 2013, we regularly circulated briefing documents for our editors and societies, to keep them abreast of the latest OA developments, and have since launched a dedicated website which contains the latest updates on OA and other areas of interest to these groups. Our marketing department have created a wealth of material around OA, with a focus on being informative, providing prospective authors with definitions of gold and green OA models – of which there is a diversity – as well as our traditional option of publication behind the subscription paywall (with 50 e-prints to circulate to colleagues, and the ability to follow a green path to author posting and archiving). The choice of publication model is up to them.
Our Open Access Author Survey15 was conducted towards the end of 2012 and into the beginning of 2013.
It has attracted a fair share of comment and incited some interesting debate, especially around OA licensing, reuse rights and peer review. The results of the survey have been made openly available under the Creative Commons Attribution (CC BY) licence and further analysis has also been made available through publication of supplements (also published under CC BY). This survey has guided our internal policy development, such as decisions around licensing for gold OA content, and we have brought these outcomes to the attention of policymakers. Our White Paper ‘Facilitating access to free online resources: challenges and opportunities for the library community’, focused on the issues that librarians face in cataloguing and quality assurance of free online content.
In early 2013, our membership of the Open Access Scholarly Publishers Association (OASPA)16 was approved. We had to implement some changes to the way that our OA content was displayed online to be granted membership, and our decisions around OA licensing were also influenced by our desire for membership.
“… Taylor & Francis Group launched a new, pure OA division, Cogent.”
Admittedly, our size and legacy subscription business has hindered rather than helped in some respects. Unlike new entrants to the market which work solely on the OA model (pure OA publishers), we must be mindful of our 1,850 existing journals, the majority of which have a subscription element, and many of which, again, are owned by societies or co-publishing partners. In view of this, in September 2013 Taylor & Francis Group launched a new, pure OA division, Cogent17. Cogent shares many of the same philosophies as T&F but is a pure OA publisher. Every publisher is aware of the importance of their relationship with their authors, but Cogent will be able to focus on this particular group because they have fewer other players to take into account. The staff at Cogent will be able (and encouraged) to take new risks and experiment with new models, pioneering and promoting OA without needing to think about possible impacts on the subscription model. Cogent anticipates launching a number of broad-based mega-journals during the course of 2014.
The to-do list
There has been a lot of progress within the industry over the past couple of years, but this has thrown up a number of challenges which need to be addressed. The major issue is around transparency. T&F has a subscription adjustment policy18, but we are still considering how best to address the ‘double dipping’ issue to ensure not only that institutions and funders that have paved the way for OA don't pay twice for content, but that we also provide them with an equitable return in view of their investment in gold OA.
We are also reflecting on a number of related points. Some of these are technical, such as making OA content (especially hybrid OA) more discoverable on our platform, expanding our current framework agreements19 around deposit with the World Bank and National Institutes of Health to other bodies, contributing to industry standards around permitting text and data mining, and so on. Others are less tangible, focused for example on increased collaboration with the other participants in the publishing ecosystem, reflecting on the future for grey literature, and considering public engagement around research.
We advocate creation of a consistent OA terminology. Our Production and Author Services departments receive hundreds of e-mails a week from authors who find it hard to grasp the key concepts around OA, especially when they hear often contradictory definitions. We created a short motion graphic in late 2013 which we hope is informative. Whilst alluding to the many nuances around OA, it explains the ‘main’ OA definitions of gold and green OA in under two minutes20.
We are facing an environment that is ever changing, and both Taylor & Francis and Cogent will continue to cater to the needs of the academic community. Behind the scenes development has been rapid, thanks to the skills and effort of a large group of talented staff, who have been eager to engage with some of the challenges faced by an increasingly micro-level publishing model. However, thought still needs to be put into how smaller publishers and learned societies can adapt or make a transition to OA.
We end by challenging the idea that we are already in a state of transition to OA. Subject differentials, funding and attitudes around reuse make an entirely gold OA future seem somewhat unlikely. We are currently in the ‘third age’ of publishing, part of a complex and ever-changing ecosystem. Journals serve as a prime example of the network effect in a connected online publishing world. Coupled with technological advances of recent decades, more people are now connected and there is more material than ever before available through online media. This ecosystem incorporates the subscription, site licensing and OA models. Each of these models co-exists within this ecosystem, yet could exist without the other models (excluding green OA which, due to its current dependence on the subscription model, would not be able to operate independently). Combined, the strengths of each model become clear, and some of the weaknesses are mitigated.
“We are currently in the ‘third age’ of publishing, part of a complex and ever changing ecosystem.”
Publishing is of course a service industry, adapting to the needs of the (academic) community. At the end of 2013, 99% of T&F journal publishing is still dependent on institutional sales models or subscription publishing. Should the academic community wish OA to become the best means to communicate the outputs of their scholarly research, then OA will become the standard model. We need to make sure we have the tools, processes and suite of services available to support this option to the author, funder and policy communities globally.