The role of academic publishers in shaping the development of Web 2.0 services for scholarly communication

Introduction

The emergence over the last decade of a rapidly expanding series of innovations in technologies, standards and practices collectively known as Web 2.0 appears to hold out the promise of new forms of scholarly communication. However, a recent survey suggests that a majority of scholars are not yet willing to embrace these innovations (Procter et al., 2010a). While some general-purpose Web 2.0 tools and services have been widely adopted, where there is a need for more specialized tools these will have to be developed, incorporated into institutions of scholarship, and new practices will have to be recognized and valued by scholarly community members. Our interest is in how this work is done and by whom.

Among the many interesting uses of Web 2.0 in scholarly communication (Procter et al., 2010b) are the possibilities it offers individual scholars, research groups and institutions to publish, annotate, review, discover and make links between research outputs. New tools and practices that facilitate this would seem to challenge the role of established academic publishers. However, despite suggestions that these are becoming irrelevant in the age of Web 2.0 and are defending an outdated and exploitative business model, we suggest that academic publishers may play an ongoing role as innovation intermediaries (Howells, 2006; Stewart and Hyysalo, 2008) through their access to a unique set of resources to shape innovation and use, providing the scholarly community with opportunities to use and learn about Web 2.0.

Apart from some major new players from outside publishing, no other types of players in the world of scholarly communication – scholarly societies, libraries, universities, funding bodies, etc. – stand out as being more motivated or better able to drive significant experimental innovation that actually gets in front of scholars in their everyday work. Academic publishers have incentives and can develop the expertise to develop and deploy new tools and services and bring them to the scholarly community. They are able to experiment over the longer term, integrating innovations with traditional services that remain in demand. However, despite their position as ‘obligatory points of passage’ (Latour, 1987) with privileged platforms to promote (or resist) change, their success has been rather limited to date: individual publishers may struggle to drive structural change in scholarly communication in an environment where researchers expect common standards and services to link all their bodies of knowledge, whatever the source.

In research sponsored by the Research Information Network (RIN) we studied the adoption of Web 2.0 and its implications for scholarly communication (Procter et al., 2010b). In a previous paper (Procter et al., 2010a), we reported on findings of scholars’ use of Web 2.0. This paper presents complementary findings, comparing the efforts of two contrasting academic publishers to provide Web 2.0 services for scholars. Though not necessarily representative of the activities of the sector as a whole, the experiences of these two publishers are, we argue, similar to those faced by many others.

We explore how these publishers have tried to build a bridge between scholarly communication technologies and practices by putting experimental services in front of scholars as they access articles online. The research questions we sought to answer were: how are publishers seeking to maintain or reinvent their role in scholarly communication; and how are they going about identifying and creating new ways to deliver value? We find that while they largely failed in early attempts to create scholarly versions of Web 2.0 applications, these two publishers are putting in place key parts of the technical infrastructure that may eventually knit scholarly articles more closely into a network of other resources, and they are playing a key role in animating debates and discussions.

Our findings highlight the need to take seriously the role of academic publishers in the move towards a vision of more rapid and open scholarly communication and to understand the factors that shape their role as intermediaries in the innovation pathways that may be needed to achieve it.

Web 2.0, scholarly communication and academic publishing

Web 2.0 is a term that originated as a business concept for internet companies after the ‘dot.com-boom’. It can be best characterized by forms of particular services, publishing formats or user practices, rather than simply referring to particular functionalities or configurations of technology. Hence, our definition is centred around the changing practices of communication and production of information by individuals and groups (O’Reilly, 2007) and brings together four ideas typically associated with Web 2.0:

increased emphasis on user-generated content, data and content sharing and collaborative effort, together with the use of various kinds of social software, […] and the use of the web as a platform for generating, re-purposing and consuming content. (Franklin and van Harmelen, 2007)

Publishers, libraries, professional associations and conference organizers have spent the last 10 years attempting to incorporate the capabilities that Web 2.0 promises into their activities and services. The dynamism and diversity of Web 2.0 offerings and the hitherto lack of established standards and practices have favoured an experimental approach, which balances the possibilities of ‘open scholarship’ with the often conservative practices of scholars and the limitations of existing formal scholarly communication (Harley et al., 2010). Publishers have also had to meet challenges from new intermediaries outside the traditional world of scholarly communication, be it major players in search such as Google, not-for-profit organisations or commercial start-ups.

Traditionally, one of the key elements of scholarly communication is the publication of peer-reviewed research in journals, which is not only central to scholarship, but also the core currency in building careers and reputations (Arms and Larsen, 2007; Harley et al., 2010). The mediators of this process – journal publishers and their networks of editors and reviewers – are key actors, providing not only practical resources, but also legitimacy, quality control and branding. Journal publishing is a diverse industry, involving scholarly societies, university presses, large-scale publishing houses, medium-sized and small independent publishers. In the past, a journal may have served a defined scholarly community, providing a publisher with a captive market. However, contemporary scholars publish and draw on research articles from many journals and the boundaries of disciplines are under constant challenge.

The majority of studies on scholarly publishing focus on users: scholars, libraries, and systemic costs to ‘Science’ as a system (e.g. Björk, 2007; Houghton et al., 2009). There is little literature on the business of academic publishers (Lorimer et al., 2011), let alone on innovation practices of publishers as teams of people with ideas, business goals, and (limited) resources, as they react to the pressures of their users/suppliers and the opportunities of new technology (Perciali and Aaron, 2008). The academic publishing market has undergone considerable change in the last 20 years. First has been the shift to online publishing, requiring investment and learning to provide online access, computerize the entire publishing process, digitize archives and provide new tools to subscribers. For many publishers, it has not been a matter of developing technology, but buying in software and services, adapting packages, and trying to develop and improve information standards. Second, the relationship with the primary customers – academic libraries – has been revolutionized, shifting from subscriptions to individual journals to bulk subscriptions for portfolios of journals provided only online. This has put enormous financial pressure on small publishers unable to benefit from economies of scale. Third, the financial pressure on libraries and the technical ease of online publishing awoke the open access movement: why should scholars and their funders pay to access digital copies of articles they had written, edited, reviewed and could make available on a personal or university website, or disciplinary repository (Houghton et al., 2009)? It has taken over 10 years to make open access publishing mainstream, involving changes in funding, individual publishing practice and adoption of partial open access by mainstream publishers, and it still remains a dominant issue in scholarly communication.

This environment has been interpreted by some as constituting a ‘crisis’ in academic publishing; others, such as the two organizations profiled in this paper, see this as an opportunity for innovation (Perciali and Aaron, 2008). For publishers to survive, brands have to be strengthened, and the quality and breadth of their services improved, justifying their costs, and value to scholars and their communities. Failure by entrenched players to respond to the demands of scholars and the potentialities of new technology also creates opportunities for new entrants.

Conceptualizing the socio-technological innovation processes

We follow a Social Shaping of Technology (SST) approach (Williams and Edge, 1996) to explore the role Web 2.0 can play in a reshaped model of scholarly communication. SST sees innovation as involving a seamless interplay between ‘technical’ and ‘social’ factors, with choices between different options available at every stage, shaped by interactions between diverse actors with differing commitments, knowledges and perceived interests. From an STS perspective, Web 2.0 is not a given: it is driven by people creating, investing in and experimenting with ideas. A finite number of small teams do this work, influenced by access to resources and a host of other constraints, and informed by visions of the potential of the technology. Technological innovations only come to fruition if stakeholders adopt and embed them in their settings and practices (Hyysalo, 2006). This involves often extended processes of ‘social learning’: experimentation, interaction with potential users, failures and reinvention (Sørensen, 1996; Williams et al., 2005).

The frame of reference for scholarly communication has been under enormous strain in recent years and, while there may be some broad agreement on the potential of Web 2.0, there is clear disagreement about how this will be reached and the role of existing stakeholders. Alongside the visionaries, there is the mass of more conservative scholars, whose responses will eventually play a decisive role in shaping innovation pathways. This demands that analysis focuses not only on powerful stakeholders, but on detailed biographies of design and development, on the evolving practices of scholars and interactions between the two. It is also important to identify the role various intermediaries may play in both stimulating and gatekeeping innovation processes. Such ‘innovation intermediaries’ (Howells, 2006), situated ‘between’ end-users (i.e. scholars) and producers of technologies and other stakeholders (i.e. institutions of scholarship, providers of Web 2.0 technologies and services) have been shown to play a key role in social learning. They configure technical systems, broker knowledge transfer and stimulate debate and create spaces where users can experiment (Stewart and Hyysalo, 2008). In this study, the concept of intermediaries helps us examine the intermediating activities of publishers and draw conclusions about how successful they have been in playing this role.

Methodology

Our overall research design involved an online survey of scholars’ communication practices, their uses of and attitudes to Web 2.0 resources, complemented by 50 interviews with selected survey respondents, reported in Procter et al. (2010a). We then conducted case studies of five scholarly communication service providers to investigate how publishers are exploring the possibilities offered by Web 2.0. A full account of the methodology is available in Procter et al. (2010b).

In each case study, we interviewed several key staff involved in Web 2.0-based service innovations and scholars identified from survey responses. Interviews were conducted face-to-face or by telephone, followed a semi-structured format that was adapted to the activities of each organization and interviewee. Informed by the social learning agenda, questions probed the motivations driving Web 2.0 development, organizational interests and constraints, representations of potential users, and interactions with stakeholders and end-users. A further round of interviews conducted two years later provide an important longitudinal dimension to the analysis, highlighting change or lack of change, and longer-term trends in Web 2.0 development and use. Interviews were recorded and transcribed and then analyzed using a ‘grounded-theory approach’ (Glaser and Strauss, 1967), inductively identifying and iteratively developing themes as they emerge through reading and re-reading the data.

This paper focuses on the case studies of two leading publishers of conventional peer-reviewed research articles. Though publishing in similar fields, they come from opposite sides of the publishing industry: Nature Publishing Group (NPG) and Public Library of Science (PLoS).

NPG is a medium-sized academic publisher, a subsidiary of Macmillan Group, specializing in science and medical academic publishing. Its flagship journal, Nature, founded 140 years ago, is one of the oldest scholarly publications and has perhaps the strongest brand in academic publishing. It is a very successful, for-profit business and operates with professional editors. In recent years it has invested considerably in innovation, including Web 2.0, creating Nature.com and Nature Blogs, and as is seen as a leading player in the field.

We interviewed five members of the Web development team, led by Timo Hannay. He was responsible for the strategic direction of the Nature.com platform and for the more experimental services that may get incorporated into Nature.com. He had been with NPG for over 10 years and took on this role responsibilities five years ago. He reported directly to the managing director of NPG. The other respondents develop and run the Web services, and joined in the last three to four years. Two were former research scientists, recruited for their experience of setting up Web 2.0 services, one was ‘new media’ developer recruited from the BBC and the other was a specialist in text mining.

PLoS was a start-up 12 years ago, an experiment in open access publishing. It is a non-profit organization publishing a range of leading journals. It is constantly experimenting with new ideas, such as continual publishing in PLoS ONE and article-level metrics. We interviewed the Managing Editor, Peter Binfield, the Head of Marketing and the Head of Technology. The Managing Editor had 15 years experience in academic publishing and had overall responsibility for both the local and outsourced development teams. The Head of Marketing had 10 years experience in editing and marketing in academic publishing. Her role was to promote the journal and individual articles. The Head of Technology, an open source enthusiast, led the development team. A follow-up interview was conducted with the new Head of Products and Publications.

Scholars and new forms of scholarly communication

Though scholars regard subscription journals as the most important source of research information, our survey showed that open access journals are not far behind. We find a similar picture with respect to publication practices; however, only 20% of respondents expected open access to become the predominant publishing model within the next five years (Procter et al., 2010b).

We found rapid uptake of scholarly communication innovations that are generic, intuitive and easy to use, build incrementally upon existing practices, are available free, offer near-zero adoption costs and clear advantages to scholars. More specialized services, arising, for example, from efforts of publishers and other knowledge intermediaries have made more uneven progress, despite offering more scope for novel forms of scholarly communication. We now turn to our two case studies to examine why.

Case 1: From Nature Publishing Group (NPG) to digital science

In the mid-2000s, NPG took the strategic decision to actively embrace innovation in scholarly publishing following a successful transition to Web access to their publications. The small department responsible was enlarged and given company-wide responsibility for innovation in internet services. The firm started to define itself explicitly as a ‘Science Communications Company’, recognizing its business could not remain as a publisher of printed, peer-reviewed journals. One interviewee defined the motivation to invest and innovate: ‘Our future is now in the hands of other people not ourselves and we need to take control of this’ (Publishing Director, NPG). The aim was therefore not only to provide new services to its user base, but also act as an R&D lab to explore how data in and generated around key assets, including user-generated content, might drive users to publish and read more articles in NPG journals and, ultimately, sustain the NPG business: ‘We’re trying to do our job as a scientific communication facilitator’ (Publishing Director, NPG).

The primary publishing platform, Nature.com, provides the gateway to the group’s publications. In the last five years, it has been augmented with a range of new services, including Connotea social bookmarking, Nature Blogs, Nature Network and online databases. Experiments were conducted with open peer review, with facilities for commenting on articles, wiki versions of review articles and a preprint service. While many of these services continue, a core part of the development team left NPG in 2011 to set up a sister company, Digital Science, to focus on a growing market for software to support research practice. Digital Science develops data management, metrics, search services for research organizations and tools for individual scholars, and supports start-up companies working in this domain such as Altmetric. ¹

NPG has other new publishing activities focusing on data sharing and repurposing of traditional content. Many articles published in Nature journals also have their data published in repositories that can be used alone and linked to other components of Nature.com from within the article. Magazine and journal articles are now supplemented by audio and video material aimed at communicating both to scholars and the public, and promoting the Nature business.

Case Study 2: Public Library of Science (PLoS)

The Public Library of Science describes itself not as a publisher, but as a ‘non-profit organization of scientists and physicians committed to making the world’s scientific and medical literature a public resource’ and, more prosaically, ‘a non-profit, a publisher and software house’. ² It was founded in 2000 by three leading biomedical researchers frustrated with the publishing industry’s failure to move to open access, they launched two online, open access journals, PLoS Biology (2003) and PLoS Medicine (2004). Subsequently, four other subject-specific journals and a general science journal, PLoS ONE (2006), were launched, all of them publishing in the biomedical field. These journals have become successful, attracting authors by offering fast turnaround times and providing a way for them to comply easily with the open access publishing requirements of funders.

One of the criticisms the founders had of existing journals is that editors, given limited space in print journals, and the need to maintain a journal brand and its impact factor, select articles not only on scientific merit, but also on potential impact. The general-purpose journal PLoS ONE was created specifically to break out of the constraints of print-based publications – developing a process of continuous publishing of peer-reviewed articles without editorial selection. Instead of selecting articles likely to have ‘high impact’, PLoS selects certain articles post-publication to promote to the more general media: ‘Scientists like that, they enjoy it, they can get lots of people to look at their work, there is nothing to stop them looking at it’ (Head of Marketing, PLOS). PLoS ONE operates a more subtle form of filtering. The pay-to-publish model (although there are certain waivers) acts as a filter to publishing where authors are unable to pay the publishing fee. ³

This publishing process required development and redevelopment of a model of filtering and distributing articles to editors and reviewers. These demands put the initial system, an open source software product not designed to be used in this manner, under considerable strain as the number of submissions and popularity of the journal grew.

Since its launch, PLoS has become more focused on publishing. Innovation has continued in a rather public experimental mode, creating new services, and subsequently refining, evaluating and developing them with its community of users. At the same time, PLoS hopes that these innovations will attract scholars to publish in its journals. One example has been PLoS Currents, launched in 2009 in partnership with Google. This is a ‘pre-publication’ system aimed at providing a space for timely – but so far un-reviewed – work to be posted in a moderated forum.

PLoS is a much smaller organization than NPG with modest budgets and much less software/service development resource, so the scale of experimentation is much more limited. However, they have a high profile, and have focused on a small number of innovations, notably PLoS ONE and open access, that have been highly successful.

Early experiments: Versions of Web 2.0 for scholars

The case studies illustrate NPG and PLoS first investing in experimentation directly inspired by some of the successful (at that time) and emblematic Web 2.0 services. They were looking for ways to reproduce and tap into the practices of blogging and online media, including supporting open commenting on articles, as if they were YouTube videos or blog posts (NPG and PLoS), providing ‘wiki’ versions of review articles (NPG) and building a social networking site for academics (NPG – Nature Network). Hanney (NPG) observed of Connotea: ‘I looked at Delicious, social bookmarking site… and I said we should do something like this but tailor it for academic scientists’. Though mainstream Web 2.0 offerings were not adapted to scholarly needs, such as formal citation standards, the opportunity they offered to ‘share’ or make public comments on articles was somehow consistent with the spirit of scholarship and new ideas of Open Science (Neylon and Wu, 2008).

These developments were clearly seen as experiments that could only be conducted successfully with close engagement with the scholarly community. In the spirit of perpetual beta (Morris, 2006), NPG ‘set up a publishing department that has a remit not just to experiment with new things internally but to set up new user facing products and services that have a kind of experimental remit, an experimental approach’ (Publishing Director, NPG). However, more was needed than just putting the service online. Both NPG and PLoS published articles and editorials, and set up blogs to enable a two-way dialogue with the scholarly and publishing communities as part of a public innovation process. However, these experiments largely failed to reach a significant level of use, as confirmed by our survey findings. Despite this, these initial experiments led to new ideas, new resources, a more focused view of the role of publishers, and the potential for and limits to developing and profiting from innovation that both supported scholars and their own mission. This can be illustrated by exploring a selection of the services developed.

Web 2.5: Towards article-level metrics

For NPG and PLoS, creating Web 2.0-based services for scholars was an interesting experiment, but not necessarily a situation where a single publisher could either lead or make an impact. Clearly, more focus was needed to find what elements of Web 2.0 were feasible to develop for a publisher with software capabilities and would be valuable to a wide audience. Reflecting the findings of our survey, a PLoS user interviewee remarked:

Ten per cent [of scientists] care or know about Web 2.0 and only 1% of which would understand the conversation we’re having… regular academics are not really interested in all these wonderful bells and whistles. All they want is to be published in a journal with the highest impact factor and then they can retire happy.

Over 60% of our survey respondents used Google Scholar (Procter et al., 2010b) – way above any other common tool and this is a problem for a publisher trying to maintain a brand: ‘at the end of the day does anyone use it or do they just come in via Google and leave again and then come back to a different paper via Google?’ (Head of Marketing, PLoS).

PLoS set out to explore systematically the possibilities of article-level metrics (ALM), the first version being launched in September 2009. ⁵ The aim has been to continue the innovation process started in PLoS ONE, allowing articles to be judged on individual impact, rather than on the impact factor of the journal. PLoS ALM publishes page views, download figures, citation rates from citation databases such as PubMed and Scopus, social bookmarking data, including data from Connotea, and links to blog aggregators (Nature Blogs, Bloglines and ResearchBlogging) and other online media to try and present a broader picture of impact. ⁶ Developing the technology was a considerable challenge involving over four years of work. Continuing efforts are being made to find ways to generate this data. Links to enable bookmarking and linking from mainstream and scholarly social media sites are included on article pages to encourage readers to make the external link and thus generate new meta-data items. However, even prior to launch, the PLoS team recognized the cultural challenge would be a much greater obstacle than technical ones. Though our survey revealed that nearly half (47%) of respondents expect peer review will be complemented by article-level metrics of some sort in the future, many regard blogs and informal media as not holding any importance (Procter et al., 2010b). To attempt to overcome scepticism, PLoS put in place a media and communications strategy to work with different communities of scholars to explain and get feedback on ALM. The data were made available for outsiders to analyze and comment on publicly. ⁷ However, by December 2011, PLoS admitted the slow progress: ‘we plan to improve our Article Level Metrics to a point where they will provide genuinely valuable context about individual articles and hopefully be more widely used and understood by decision makers such as tenure committees and funding bodies’. ⁸

NPG embraced ALM, not so visibly as PLoS, but clearly saw the market in advanced metrics for institutions. By setting up Digital Science, this could be developed cross-publisher, and services and products created that would meet institutional metric needs. In addition to this, NPG identified an opportunity to focus more on data within articles, to use this to improve the value of the data, and build new services to support discovery and thus increase use of NPG journals: ‘How do we integrate the data that’s associated with our papers?’ (Developer, NPG). The key to this was seen as data mining:

What you really want to be doing is data mining… based on all the free information that you are getting from this. Doing the front end point without the backend part isn’t really Web 2.0, it is just the appearance of Web 2.0 without Web 2.0. (Developer, NPG).

Shaping Web 2.0: Scholarly publishing as a media communication business

We can situate these developments in Web 2.0 in the context of the changing scholarly publishing process. The development of ALM highlights two developing roles for the scholarly publisher. First, the publisher not only conducts the behind-the-scenes work of managing, editing, reviewing and producing a journal, but is expected to actively promote the work it publishes, not only for the good of facilitating discovery of scholarship to its readers and the general public, but also to reinforce its attractiveness for scholars who will then try to publish in its journals and – critically to this process – to raise their impact factor, their key currency. To do this, both PLoS and NPG have a ‘full media strategy’, selecting articles to promote in mainstream and social media: ‘At some level we’re becoming like a broadcaster’ (Publishing Director, NPG). Social media becomes part of general engagement with media in promoting the work of scholars, and thus journal brands: ‘We embrace platforms from other people constantly – we use Twitter, Facebook, Wordpress’ (Head of Marketing, PLoS).

Second, and this is one of the aims of ALM, the publishers are now trying to help authors, their employers, funders, etc. to understand the impact of their work, encapsulated in its most valued output, the article:

In an online world, the article is what it is all about, and it is the article that generates so many different impacts – the article will generate citations, blog posts, mainstream media coverage, social media bookmarking, downloads. There are so many new ways to see the reach of your research. Who would not want to see their article on the front page of Time magazine? (Head of Marketing, PLoS).

Not only are individual scholars interested in who is using their research and how, but their institutions and funders are also turning their attention to non-traditional ‘use’. However, as PLoS identified, and consistent with the negative attitudes expressed in our survey, the social acceptance and valuing of these metrics is a slow process. Metrics are used to judge quality and impact, and a career is based on these judgements. Publishers are not just onlookers in these processes; their business models stand or fall on metrics too. The introduction of new metrics helps them reinforce their own business but, unless they actually do the work of making these numbers available, there can be no debate about their value.

Innovation practices

Our case studies also allow us to explore in more depth the processes of innovation that were used to create these Web 2.0 services. Broadly speaking, they follow the ethos of the ‘perpetual beta’: making early, simple versions available to a user community which experiments with and ‘co-produces’ new tools, services and content in a more or less continuous process (Morris, 2006). This provides the opportunity for rapid feedback before investment is made to develop innovations fully (Procter et al., 2011). It facilitates management of risk because it allows early evaluation and service change. The publishers made decisions to pursue some paths of development and to sacrifice others where the evidence of uptake and returns on investment is lacking.

NPG initially followed innovations that arose in the broader internet community. However, it is seen as a leader in academic publishing and it went on to invest considerable resources in service development, focusing on what could be achieved, and the reaction of an enthusiastic community of users and informal testers, rather than working out a business model in advance.

If we’re going to fail, hopefully we can fail quickly and cheaply, but try different things and see what works and see what will be useful and also see what we can make money out of and what we can’t. (Publishing Director, NPG);

[in] their experiments you can’t see with any of them what their business model is, how they’re going to make money out of it. I think they’re just in this mode of massive experimentation. (Managing Editor, PLoS, on NPG).

Like NPG, PLoS has followed innovations elsewhere in online publishing. With a small technical department, PLoS admits that they have few resources to do more than modify the open source or licensed software systems they use. Like NPG, PLoS has attempted a number of innovations that have required new tools to bring together diverse information sources, and has had to engage in a programme of outreach and dialogue with users and policymakers to promote its vision of shaking up traditional scholarly publishing,

Ongoing user engagement is essential for driving forward innovation processes (Williams et al., 2005). Development of Web 2.0 services by both NPG and PLoS was characterized by a diverse pattern of interaction with potential users through conventional and new media platforms, where they attempted to justify their activities and gain constructive feedback: ‘Innovation comes from having two-way communication channels to allow them [the users] to shape what we do’ (Head of Marketing, PLoS). However, for PLoS and NPG, readers were not the only users to be engaged with; directors, editors, authors, scholarly societies and outspoken scholars were part of a range of influential stakeholders, and competitors with whom they needed to develop common standards and practices.

NPG promoted the company’s vision at conferences, hosted and stimulated discussions on Nature Network, published articles and journal editorials on Web 2.0, their experiments ⁹ and the future of scholarly communication. ¹⁰ PLoS and NPG stimulated and encouraged debate in scholarly blogs. NPG have long experience in engaging with their specialist market and a good knowledge of its concerns, keeping up to date through public debate, informal communications and everyday working relationships. As newcomers, PLoS were at a relative disadvantage and so employed someone from NPG to develop this capability for them.

To understand their dynamics we also examined the fine structure of these innovation processes. The two case studies highlight the importance of enthusiastic users and co-innovators of the Web 2.0 services they develop:

We want to be in tune with the early adopter and working with them to develop services that hopefully the mainstream will eventually come along and find a use for. We don’t want to be too far ahead ’cause that’s as much a risk as being too far behind. (Publishing Director, NPG).

For NPG and PLoS, these early adopters tend to come from particular disciplinary communities, notably bioinformatics and chemistry.

Managing uncertainty and developing future prospects

The services highlighted in our two case studies are at a relatively early stage, not so much in terms of their technical implementation, but in terms of establishing stable user communities and patterns of usage. It will be clear from the discussion that it is not always possible to predict the trajectory of these innovations. Both publishers highlight the role of specific groups of users, noting that particular disciplines have embraced what they are doing, while others reject. There are individual, cultural and technical factors at work. Chemistry uses information objects amenable to early incorporation in linked data and in social software tools, which are not yet available in biology and genetics. Different disciplines were perceived to be more or less open to the ideas of ‘sharing’ data and community interaction: for example, PLoS considers bioinformatics to be very open, and molecular biology very closed: ‘Genetics and genomics are very used to sharing data. Anyone used to sharing very readily takes to this. When there is a disincentive to sharing, they will not take to it so readily’ (Developer, NPG).

Providers of innovative scholarly communication services need to develop an understanding of disciplinary practices and concerns and may need to provide advice to potential users, for example, on attribution mechanisms or data curation issues. Many of the innovations have yet to be embedded in the scholarly community. While for some this may be seen as just a matter of time, institutional and cultural barriers, such as established publication models and issues of reputation, are hurdles yet to be overcome (Harley et al., 2010).

Despite the potential of Web 2.0 services to disrupt existing practices of scholarly communication, the role of traditional publishers and peer-reviewed journals remains strong. Service innovations by traditional publishers often serve to maintain this position in the face of potential challenges by new players such as PLoS. By creating new platforms, such as Nature Network or Blogs, NPG is able to increase the status and visibility of its products and of those who contribute to them.

The NPG case study illustrates tensions that established publishers may experience. On the one hand, NPG is developing a core platform that is the main point of entry to its products and allows users to discover and discuss research. On the other, NPG realizes that Nature is just a node in a network of scholarly communication, and users are just as likely to find synergies between communication and information sources outside Nature as within, with the implication that everything has to be done to facilitate users in linking Nature to other sources and services. With this in mind, NPG is embracing open data standards to ensure that Nature.com remains a key point of passage for researchers.

The NPG case study also provides an illustration of how attempts to integrate Web 2.0 services can put stresses on established divisions of labour and working practices. Despite top-level management support, service innovations exposed differences of interest and practice, reflected in a complex pattern of responses. An important influence appeared to be relations with users. Scholarly communities are taking to Web 2.0 in different ways and this is reflected in the attitudes of NPG frontline editors, some of whom see no relevant user demand for journals to introduce features such as reader comments. Although willing to experiment, there are strong concerns that innovations such as this and ‘open’ reviewing will undermine quality control. Though publishers might delegate this to users to a certain degree using a ‘wisdom of the crowds’ approach (Surowieck, 2004), it is not clear that it functions effectively without careful intervention of publishers, such as white listing of bloggers and filtering of posts – publishers must still be able to exercise some editorial control. Nor is it clear that that ‘crowds’ are enthusiastic to take part. Some editorials asking for input received no feedback and comments facilities have been seen to be of little use. ¹¹ And though the NPG Web 2.0 development team report receiving a good deal of feedback, this is from a quite narrow community of enthusiasts. Despite low initial uptake, development of metrics of journal use and Web 2.0 services made available by moving online is starting to provide important information. Developing a strategy that ensures support from within the organization as well as effective and sustainable exploitation of the potential of Web 2.0 services remains crucially important.

PLoS, on the other hand, started out wishing to shake up academic publishing, but soon had to face the reality of running a publishing business. They still attempt to innovate, but face considerable inertia from the scholarly community. The development of ALM is one example. Considerable technical work was needed to bring together a range of metrics that might measure the impact of an individual article. This was a necessary step to promote their goal of reducing the importance of a journal’s brand on the evaluation of the impact of an article. However, launching ALM was only the first stage in stimulating a conversation with the scholarly community over the usefulness of ALM, which could only be established by actually having it used. Without other publishers doing the same, however, the success of the venture must be in question.

Returning to the idea of publishers as innovation intermediaries (Howells, 2006; Stewart and Hyysalo, 2008), it is clear from our case studies that, though NPG and PLoS developed little original technology, they were able to play an important role in stimulating experimentation and debate at the heart of scholarly communications and in the everyday practices of early users. Their rapid integration of innovations with existing platforms facilitated joint experiments between users, developers, editors and other stakeholders. As high profile organizations, they were able to bring these to the attention of stakeholders, and to act as legitimate actors and as brokers of ideas. In this process, they were up against some powerful, alternative voices and embedded institutional constraints.

Conclusions

Our findings shed light on the attempts of two academic publishers to take on board new forms of scholarly communication and find ways to appropriate Web 2.0 technologies that not only satisfy their audiences, but fit the practicalities of running peer-reviewed journals and make commercial sense. They illustrate tensions that publishers can experience in their efforts to innovate services and develop sustainable business models. Web 2.0-based services can not only improve information management, discovery and sharing, but also provide an important source of knowledge for providers. However, while some may use this for advertising or other revenue generation, it is not yet clear whether or how academic publishers can take advantage.

Our study suggests that, despite the focus often given to Web 2.0’s ‘bottom-up’, user-led dynamics, the capacity of established players to harness such innovations looks certain to have a major bearing on progress towards new forms of scholarly communication. A key issue will be the strategy and innovative capacity of scholarly publishers, which varies widely, rather than just size and resources. An experimental approach – developing various prototype services and seeing what works – is relatively cheap, quick and easy to learn from: it can be done by the smallest scholarly society and by major corporations, the former benefiting from institutional flexibility and the latter from deeper pockets. For experimentation to have an impact, however, it needs to be delivered in ways that mesh with users’ everyday practices. It is also clear that considerably more investment is required to move beyond experiments to services that really take advantage of what Web 2.0 can offer.

We have seen how NPG and PLoS have tried to play the role of innovation intermediaries: they are able to act as brokers by controlling some key assets, and having high visibility and presence. However, individual publishers have only partial control, which limits their ability to drive change. They may play a role as configurers of technologies and service innovators; in making new services available they may facilitate experimentation by the scientific and publishing community. However, just providing services does not change practices overnight. When initial experiments in commenting and rating failed to take off, NPG and PLoS turned to a new model of scholarly publishing service centred on discovering and linking to other sources. This is geared to helping scholars to begin to understand what happens to their work when it is published, to track how articles are being ‘consumed’ and by whom. It remains far from clear how such metrics will be adopted and used. To develop this model, publishers will need to work together and with other stakeholders to promote their ideas and develop open data standards.

Online fora, editorial and reviewing systems provide publishers with a platform to present innovations to users who, in turn, get to try out novel features. However, we have seen how publishers may face differences of view from within their own organizations, and apathy and resistance from users. Publishers such as PLoS and NPG may be able to design services that build their brand and utility for users, making it easier for them to discover and reference articles and data within their publications. However, the benefits they might reap are constrained by being just one node in a heterogeneous network of scholarly communication and the increased ease with which users may follow links to external sources. The challenge for publishers as they embrace new forms of scholarly communication is to ensure that they remain a key point of passage in the scholarly communication cycle when scholars may be just as likely to find synergies between communication and information sources beyond their own service offerings as within them.

More generally, we draw attention to the emergent character of the new knowledge infrastructures of scholarly research. We note the uneven dynamism that surrounds development of new services, their adaptation to the needs of particular groups of scholars, the responses of the wider scholarly community and its institutions – and the widely dispersed processes of social learning involving trial and error experimentation with innovative tools and practices. Another kind of intermediation may be important in capturing local innovation experiences and exploring how they may be applied in other settings – what Sørensen (1996) describes as social learning ‘by interaction’. Some publishers are in the vanguard of this process.

This study has broadened our understanding of the range of pertinent actors, drawing attention to innovation intermediaries – and to publishers, in particular – who constitute the infrastructures of scholarly communication and whose roles have hitherto been little addressed. It has demonstrated the value of detailed, longitudinal research in explicating the biography of technological artefacts and practices as they evolve through protracted processes of experimentation. We hope our findings will help in mapping out pathways for the emergence of innovations that are geared more closely to the exigencies of the communities they aspire to serve.

To this end, we suggest a number of areas where further research is needed. First, on continuing efforts to innovate quality control, including both conventional and novel scholarly outputs. Second, how new types of metrics are being appropriated and how they are shaping practices at the level of individual scholars and institutions of scholarship. Third, how (if at all) new socio-technical configurations support new discovery practices and improved scholarship. Finally, how these are being shaped by major trends such as the shift to open access and the globalization of scholarship.