Throttled

Throttled

1

The World Wide Web was created as a means of sharing documents and information among academic researchers. While much of what we think of as the internet has followed this basic information-sharing principle (e.g., search engines providing links to information sources), over time the advertising-based business model of most internet services has led to a subversion of the Web’s basic purpose. For many sites, “engagement” – keeping a viewer on one’s own platform where one can show that viewer more and more advertisements – has become the end goal, rather than meeting the needs of the user. 

Elon Musk’s recent admission that yes, X does deliberately throttle the spread of any tweets that include links, made obvious to many what has long been the practice of nearly every social media platform. While platforms such as Instagram and TikTok were never really about sharing links, the former Twitter certainly was. LinkedIn, which has become a more popular venue for sharing professional information and encourages the writing of posts and newsletters, engages in the same throttling practice as its peers. 

It is not just social media sites that have become allergic to links. Search engines, which were designed from the beginning to index other sites and to link to them, are increasingly focused on providing AI (artificial intelligence) summaries instead of sending users to linked information. This is particularly disappointing for those of us old enough to remember when Google was a champion of the “open web” in opposition to the walled garden of AOL or even the curated experience that was Yahoo!.

This is obviously problematic for anyone in the content business. For example, publishers. And it is why publishers and others in the content business are (cautiously) embracing Bluesky. The mantle of open web champion has passed to this upstart social media platform. Bluesky, unlike nearly every other social media platform, does not deliberately suppress the performance of posts with links. It also enables users to tailor their experience by “choosing their own algorithms” – in sharp contrast to Meta, LinkedIn, and X and their rigid, centrally controlled algorithmic feeds that determine what users see (at least until Bluesky potentially succumbs to an advertising-based business model). 

While we don’t expect Bluesky to challenge the dominance of platforms such as Instagram, Facebook, TikTok, or YouTube, it is possible that audiences seeking serious content, from sources they trust, will gravitate to channels with prominent outbound links and where they can choose their algorithms.

Given how little traffic other social media sites pass through to publishers, Bluesky doesn’t need to become anywhere near as large (from a daily active user volume perspective) as its peers to have a meaningful impact. Indeed, it may already be happening. Bluesky reports driving 3x more traffic to publishers such as The Boston Globe than X and Threads provide.

The concentration of one’s users also matters. Altmetric data show that use of Bluesky is growing rapidly in the academic community and, for some papers, traffic from Bluesky is larger than from any other social media platform. Over the past few months, many publishers have joined Bluesky’s #MedSky, #AcademicSky, and #PhDSky communities (The Brief has joined this bandwagon, as has the team at C&E). If the people you want to reach are there, the overall size of the platform is less relevant. 

That said, discovery of scholarly and professional journals is very different from that of news and other kinds of consumer media. For most journals, the number of social media referrals remains tiny (often less than 6% in aggregate) as compared to the number of referrals from specialized search and discovery services, such as Google Scholar (which does not yet attempt to provide AI summaries), PubMed, Scopus, Web of Science (WoS), Dimensions and so on. Certain fields have even more specialized discovery services (e.g., APA PsycInfo for psychology). The presence of these services has created less reliance among journal publishers on social media platforms for referrals. (That said, as we noted in the October 2023 issue of The Brief, a social media strategy is about much more than just referrals.)

While a social media platform without link throttling could change how marketers use social media, we are not seeing a material impact from Bluesky on journal referrals, yet (there may be exceptions for specific communities that are more active on the platform). We are, however, seeing increased traffic to journals from ResearchGate, LinkedIn (despite link throttling), and interestingly, ChatGPT and Perplexity. This latter LLM (large language model) places links to sources at the top of its summaries. While AI discovery is still a small source of referrals, Perplexity and ChatGPT combined provide more than any other single social media platform for most publishers. 

Publishers have long structured web pages for search – not just general SEO (search engine optimization) but also industry-specific formatting for Google Scholar. While still in its early days, growth in AI services (as a replacement for search) points to an emerging need to shape how content is understood and surfaced by AI systems. Unlike the well-established practices of SEO, the rules, impact, and MarTech (marketing technology) around AI optimization approaches are still unfolding. In fact, no one can even settle on one name for it – with LLMO (large language model optimization) and GEO (generative engine optimization) being the two most commonly used terms.

Finding the right balance between supporting discovery through AI (increasing traffic) and licensing content to AI (essentially reducing traffic) is a challenge. However, proper citations, brand recognition, and outbound links in AI responses can drive traffic, making it less of a zero-sum game between licensing and discovery. The challenge is maintaining content value and authority in an ecosystem where AI increasingly mediates information discovery and access.

Legacy

2

Market consolidation has been a long-running trend in the scholarly publishing space since the advent of the Big Deal, when journals first moved online. Scale has become an essential component of success, originally fueled by the benefits of building large packages of subscription journals and now turbocharged by open access (OA) business models that reward ever-higher publication volumes. December saw yet another consolidation as the independent publisher Mary Ann Liebert (MAL), Inc. was purchased by Sage. 

The deal makes sense for both sides. For Liebert, the acquisition ensures the imprint’s ongoing legacy. Legacy was, based on the language of the press release, top of mind for MAL’s founder and namesake, and understandably so – building a company of MAL’s stature is a life’s work. Liebert clearly saw in Sage’s founder, Sara Miller McCune, a kindred spirit. Sage is now operated under a trust that protects the company from being sold to another publisher. Under the terms of the acquisition, MAL will continue to operate as a separate imprint, maintaining Liebert’s legacy. 

For Sage, a key benefit of the acquisition is undoubtedly publication output. Sage was, until around 2013, one of the “big five” commercial publishers in the market, by publication volume. By 2023, Sage had fallen to the ninth spot by article output as competitors built out their portfolios. The purchase of MAL, which published around 8,300 articles in 2023, will significantly boost Sage’s market share. Combined with MAL, Sage would have slotted three spots higher as the sixth largest in the market in 2023. MAL also brings expertise in medical and life science publishing – areas of strong funding and consistent growth.

Still, the merger of two independent publishers means one fewer option for society-owned journals. Should further open/public access policies be enacted over the next year, we expect the market to continue to narrow.

ONOS

3

Much ado was made last month around the announcement of India’s US$710 million, three-year, “One Nation, One Subscription” (ONOS) deal with 30 major journal publishers. The deal includes access to 13,000 journals for some 18 million students, faculty, and researchers at 6,300 government higher education and central government research institutions. One report notes that an earlier intention to include access at public libraries was left out of the deal, which also excludes private universities, research organizations, and other institutions. 

India is an area of significant research (and publication growth), and this largely cements the nation, along with China (the other region of greatest publication growth), outside the realm of required OA. It is notable that the ONOS deal lands just as the US federal agencies finalize plans for their OA policies (see Item 4 below), potentially shifting the US toward a gold OA orientation. The ongoing “pluralization” of the market continues, and as we noted back in 2022, “Publishers (as well as other organizations) are increasingly forced to navigate very different regulatory regimes with, at times, conflicting requirements.”

Final Drafts

4

The end of the year brought with it the deadline for US federal agencies to publish their final public access policies. While the future of the Nelson Memo and its associated zero embargo policies remains up in the air(presumably we’ll learn more in the coming months as the new administration begins), a few points are worth noting.

First, both the National Institutes of Health (NIH) and the National Science Foundation (NSF) have changed the language around the rights they are requiring from funded authors, both organizations removing requirements that authors grant them the rights to create derivative works (and the rights to allow others to do so). This appears to be in response to concerns that allowing the generation of derivative works would compromise scientific integrity (e.g., an administration could alter results of a climate change study to better suit its policies). These changes may be problematic in that agencies generally do create derivative works from the Word documents or PDFs that are uploaded by authors, by formatting them, typesetting them, adding links to references, etc. Doing so could be subject to legal challenge, although it seems unlikely that anyone would file such a case. The NIH’s guidance does state that they will “continue using features, existing or to-be-developed, that ensure accessibility and usability.”

Much discussion has been happening online about wording from the NIH that resembles the Rights Retention Strategy (RRS) promulgated in Europe and the UK in recent years. Under the RRS, authors can claim particular rights (such as the right to post their accepted manuscripts in public repositories with zero embargo, suggested by NIH) at the behest of their funder in advance of submitting them to journals, hence (in theory anyway) pre-empting any journal requirements around payment for immediate release. 

The NIH language was likely included as a result of being (like all US federal agencies) in the difficult position of having to implement an expensive policy while not being offered funding to cover the costs. Each agency goes to great pains to repeatedly state that compliance is “free.” While technically accurate (there is no cost for an author to deposit their manuscript in the agency repository), the statement fails to recognize the reality of the essential role of publication in the academic career structure (deposit is free, but if you want to both deposit and publish your article in the journal of your choice, that may be quite expensive).

As has been the case with the RRS, regardless of any rights previously claimed by authors, journals can simply require payment of an article processing charge (APC) and gold OA publication for any authors unable to provide the full rights necessary for subscription publishing (or they can simply reject the article or offer transfer of it to a fully gold OA title in their portfolio). 

The NIH further attempts to circumvent this by declaring that “Costs for publishing services that are charged differentially because an Author Accepted Manuscript is subject to the NIH Public Access Policy or the work is the result of NIH funding are unallowable because charges must be levied impartially on all items published by the journal, whether or not under a federal award.” This language is sufficiently vague to be largely unenforceable – any journal that treats NIH-funded manuscripts the same as all other manuscripts requiring zero embargo release or demanding limited rights transfer, regardless of funding source, would seem exempt from the limitations.

Given that the Nelson Memo guidance requires policies to be “agnostic” toward journal business models and that the NIH itself states that it “promotes author choice in journal selection,” the Nelson Memo remains, for most authors, an unfunded gold OA policy. Any attempt by funding agencies to limit academic freedom by circumscribing an author’s ability to publish in the venue that best reaches their intended audience and provides their preferred career advancement rewards would be met with significant protest. This is why Plan S, for example, quickly added wave after wave of contingencies allowing funded authors to publish pretty much anywhere they chose, and primarily resulted in boosting the financial performance of hybrid journals. 

Actual enforcement of a “no-APC” policy by federal agencies would likely be made moot by universities instead using grant overhead funds in their libraries to pay for an increased number of transformative agreements. The same APCs would be paid from the same grants, just indirectly. And of course, this would bring about the unintended consequence of privileging authors at wealthier institutions and driving market consolidation through favoring the largest commercial publishers.

Briefly Noted

5

eLife has chosen pragmatism in the ongoing saga of trying to reconcile its idiosyncratic publishing model with existing industry indexes. Rather than disappearing altogether from WoS, eLife will provide the service with a feed of articles rated “solid” or better by its peer reviewers. While this partial indexing still precludes the journal from receiving an Impact Factor, it does at least allow the majority of eLife articles to remain discoverable through the service. Meanwhile Scopus has determined that eLife no longer qualifies for inclusion in its journal collection and will move the title to its index of preprints. 

eLife is not the only journal that has fallen under scrutiny from Clarivate. Last month, the analytics company de-listed the prominent Elsevier journal Chemosphere from WoS. The journal retracted eight articles and issued 60 expressions of concern last month and was in the news recently for publishing a now-corrected, flawed study on black plastic spatulas (resulting, we should add, in at least one of us here at The Briefunnecessarily replacing several kitchen spatulas).

The Gates Foundation has added yet another wrinkle to its increasingly incongruous publication policy activities. Gates has given Public Library of Science (PLOS) a $3.3 million grant to, among other things, pay APCs for Gates researchers (something the funder said it will no longer do) for publication in PLOS journals (something the funder said is no longer necessary, with preprints sufficing for all publication needs). The funds are meant to help PLOS on its long-sought journey away from the APC model, though we note it is being used (at least in part) for APCs. Securing a grant to pay APCs seems a circuitous  way to move away from them but perhaps having that money in the bank (along with recent awards from the Gordon and Betty Moore Foundation and the Robert Wood Johnson Foundation) will allow further experimentation from PLOS in the next three years. 

Ian Mulvaney, CTO for the BMJ Group, adds his thoughts as the idea of establishing identity-based signals of trust in scholarly publishing builds steam. As Mulvaney notes, “The reality is that the scale of publishing has now gotten so large that we can no longer rely on informal trust networks.”

Ongoing research integrity issues are making systematic reviews harder and harder to accurately perform. Science reports that researchers are finding the literature so polluted with paper mill publications that the effort to filter out the fake from the real is drastically increasing the burden on review and guideline authors. 

Does it seem like every other ad on television (and yes, streaming services too) is from a pharmaceutical company? That may change in coming years, as a newly stated goal from the incoming US administration is to ban direct-to-consumer drug advertisements. While achieving this may be difficult, doing so would potentially be a boon to scholarly publishers as a small (but perhaps meaningful) portion of the marketing budgets of drug manufacturers would likely shift toward healthcare professionals (specifically, prescribing physicians) through journal ads, custom publishing, and events.

Last month brought a study in contrasts as authorities in Finland downgraded publications in a large number of “grey area journals” from MDPI and Frontiers to a level where they no longer count toward researcher assessment, while simultaneously a large German consortium of over 100 universities signed a new two-year publication deal with MDPI. Dorothy Bishop, FRS, has provided some thoughts on these two apparently conflicting developments. Frontiers, meanwhile, is of course not happy about the Finnish policy.

Science offers an informative overview of SciELO, its successes (growth in journals and increases in those indexed in Scopus and WoS), and its ongoing struggles (“I can’t say it is a sustainable model”). Well worth a read.

Do peer review awards work as motivation? A new study suggests just the opposite – those who receive recognition for peer review work become less likely to continue reviewing. This lines up well with David Crotty’s 2015 Scholarly Kitchen piece explaining the difference between extrinsic and intrinsic motivationthe 2015 Scholarly Kitchen piece explaining the difference between extrinsic and intrinsic motivation, which quotes educator Alfie Kohn explaining, “The more that people are rewarded for doing something, the more likely they are to lose interest in whatever they had to do to get the reward.” The BMJ and ResearchHub are going against this conclusion. The BMJ is offering £50 payments to patient and “public” reviewers, though not all reviewers, across its journals. ResearchHub, launched in 2020, takes a more radical approach, paying reviewers in a specially created cryptocurrency. To our eyes, incentivizing reviewers to work in bulk and get paid more, potentially by using AI to review papers that may have been generated using AI, to collect cryptocurrency, offers an end-to-end complete cycle of purposeless, computationally driven, environmental destruction.

El Pais reports that, “The great Saudi university farce is coming to an end.” To boost the international rankings of universities, Saudi Arabian schools were paying foreign researchers to list the schools among their affiliations, essentially gaming the system. International exposure of the practice appears to have resulted in Clarivate adding new filters to eliminate its benefits.

The Internet Archive (IA) has dropped its appeals and ended the four-year-old copyright case that arose from its practices during the pandemic. The answer is definitive now – the controlled digital lending practices of IA (scanning copies of in-copyright print books and distributing them online in a one-owned to one-loaned manner) is not Fair Use and is indeed copyright infringement.

Springer Nature’s ninth annual State of Open Data global survey results suggest progress, but there is still a long way to go toward achieving Open Science goals. 

Although on a smaller scale than Sage’s purchase of Mary Ann Liebert, Inc., the market consolidated a little more with Emerald’s acquisition of Information Age Publishing, responsible for some 2,600 books and six journals.

While investors continue to pour in billions of dollars (and AI companies suggest they will need “unimaginable sums of money” to succeed), it’s worth asking the question whether being the company that builds AI infrastructure or the one that builds services around a variety of interchangeable AI systems is likely to end up the better business. 

While January 1 has in the past been a day of celebration as more works enter the public domain, we at The Brief are beginning to dread it, or at least the seemingly obligatory horror movies spawning from newly available intellectual property. If Winnie the Pooh on a bloody rampage wasn’t enough for you, say hello to this year’s Popeye the Slayer Man. Sigh.

***
You come right over here and explain why they are having another year. – Dorothy Parker, in a telegram sent to Robert Benchley, December 31, 1929