Annual Reviews (AR) recently announced that over the next 18 months it aims to make its entire portfolio of 51 academic journals freely available under a new journal publication model known as Subscribe to Open (S2O).
Annual Reviews is a pioneer of S2O, having first trialled it in 2017 with its journal Annual Review of Public Health. A number of AR’s other journals have subsequently been converted to S2O and the publisher is now hoping to migrate its entire journal portfolio to the new model.
What is S2O? The S2O Community of Practice web site describes it in this way:
“S2O allows publishers to convert journals from subscriptions to OA, one year at a time. Using S2O, a publisher offers a journal’s current subscribers continued access. If all current subscribers participate in the S2O offer (simply by not opting out) the publisher opens the content covered by that year’s subscription. If participation is not sufficient – for example, if some subscribers delay renewing in the expectation that they can gain access without participating – then that year’s content remains gated.”
The web site adds, “The offer is repeated every year, with the opening of each year’s content contingent on sufficient participation. In some cases, access to backfile content may be used to enhance the offer.”
In light of AR’s announcement, I emailed a number of questions to the President & Editor-in-Chief of AR, Richard Gallagher. Those questions, and Gallagher’s replies, are published below.
The Q&A begins
RP: Annual Reviews (AR) is the pioneer of Subscribe to Open (S2O). It began, in 2017, with a pilot funded by the Robert Wood Johnson Foundation in which the Annual Review of Public Health journal was made open access, including all the backfiles from 1980-2016. Further AR journals were later released under the S2O model and Annual Reviews has announced that over the next 18 months it hopes to migrate all 51 of its journals to S2O.
Can you talk me through the journey AR has undergone, what has been learned, and why it now wants to convert all its journals to S2O?
RG: When I joined Annual Reviews in 2015, review articles didn’t figure on the to-do list of the OA movement. A senior figure at one OA-pioneering research funding agency told me that he regarded writing review articles to be an out-of-hours activity that did not qualify for APC support.
The lack of recognition of reviews was understandable given the relatively small numbers published each year, but it was also frustrating because I felt that, from a practical point of view, reviews should be among the top priorities for open access.
“Usage increased in that one year by a factor of 4, and these were readers, not bots”
Our big break came with the 2017 Robert Wood Johnson Foundation grant that you mentioned. It had two components: to cover the cost of the Annual Review of Public Health for one year, allowing us to assess the impact of removing the paywall; and to fund the development of a viable OA model for reviews, since APCs and Read and Publish couldn’t be adapted to our needs.
Usage increased in that one year by a factor of 4, and these were readers, not bots. We assessed how far through articles users were scrolling and there was no difference between the open journal and our most similar paywalled journals; if bots had had an impact, the scroll patterns would have been different.
To help develop a business model we hired Raym Crow as a consultant at the suggestion of Kamran Naim, my co-grantee and then Annual Reviews’ Director of Partnerships and Initiatives. Raym had a track record in collective funding logic, and in the very first meeting with him I realized that he was laying out a practical approach that we could try. That was in June 2017.
It took time to generate a proposal that seemed robust enough to take to our Board, Editorial Committees, select customers (and our staff!) but for the 2020 sales cycle (i.e., by mid-2019), we were ready for an S2O pilot project.
In addition to the Annual Review of Public Health (which had remained freely accessible in the meantime), a mix of established and newer titles in the physical, biomedical, biological and social sciences were selected.
We discussed the project openly, outlined our thinking in Learned Publishing and opted not to copyright the Subscribe to Open name in the hope that others pursuing similar programs would also use it.
Subscription levels were maintained for the five S2O journals in 2020 and this, combined with usage that ranged between double and quadruple that seen when the journals were paywalled, encouraged us to expand the pilot program to eight journals for 2021 and 2022.
To date, we have results from offering 15 volumes of Annual Reviews journals under S2O, and all have been successful (five in 2020, eight in 2021, and both of the 2023 volumes published so far; the remaining six are on track).
It’s promising but it isn’t a full proof-of-concept because most of our customers take multi-journal packages, and it would not have been in their financial interests to exclude the S2O titles.
Nonetheless, I believe that now is the right time to offer the entire Annual Reviews portfolio under S2O. This is partly because of an action taken in the early stages of the Covid-19 pandemic.
In March 2020, we temporarily removed the paywall on all 51 journals to ensure that subscribers would have seamless access, and usage increased by a minimum of three-fold and a maximum (for the Annual Review of Virology) of more than 20-fold. That experience generated urgency to move forward.
Whether all 51 titles are published OA in 2023 is, of course, in the hands of the customers. We are taking nothing for granted.
The model and its benefits
RP: A number of other publishers have started experimenting with S2O too, but I think the model used tends to vary from publisher to publisher. Can you say how AR’s model works, how it differs from some of the other models being tested, and what you see as the benefits of S2O, to readers, to authors, and to publishers?
RG: The S2O Community of Practice includes all the publishers that currently offer S2O journals. The community has agreed a set of criteria that defines S2O, which walks the line between a tight, meaningful business model and giving space for innovation to allow publishers to address issues particular to their portfolio.
Of course, this agreement is completely voluntary, but it has worked well so far and the differences between offerings are minor.
“For authors, the appeal is simplicity and equity. Simplicity because there is no additional labor or payment from authors, and equity because under S2O the social sciences and humanities are on a par with biomedical research, and the work of researchers in low- and middle-income countries is treated equally to that of researchers in wealthy countries”
From what I gather, S2O is a much simpler and less variable proposition for librarians than a Read and Publish deal. Having said that, I acknowledge that it does add another item to the librarians’ workload.
The Annual Reviews S2O model is straight-forward, and the contracts will be very recognizable to customers. There is no “OA premium” associated with S2O, and the overall finances of our journal publishing program will continue to be made publicly available.
We are making some tweaks for 2023 to incentivize/reward customers. In total, 10 years of content will be openly accessible for journals that adopt S2O. For the five initial pilot titles, that means three fully OA issues (with CC-BY licenses) plus removal of the paywall on a further seven years of content.
On the 43 titles that we hope to publish using S2O for the first time, 2023 will be fully OA and the nine previous years’ content opened up.
Other publishers may offer five years back content, more restrictive CC BY licenses, and in one case a blended offering that combines Read and Publish and S2O components.
In some fields, APC payments from research funders are a crucial component in journal income. Currently, funders, librarians and publishers are exploring how best to incorporate funder support into the S2O business model.
Regarding the benefits of S2O, I’d like to quote a librarian, Curtis Brundy, who is the Associate University Librarian for Scholarly Communication and Collections at Iowa State University.
He said, “A model like S2O is something that librarians have been asking for, for many years. We need to move our mindset from procurement – where we focus on spending out our collection budgets – to really investing in the open future that we envision. Libraries are already stepping forward to support S2O and their institutional values. It is inspiring them.”
For authors, the appeal is simplicity and equity. Simplicity because there is no additional labor or payment from authors, and equity because under S2O the social sciences and humanities are on a par with biomedical research, and the work of researchers in low- and middle-income countries is treated equally to that of researchers in wealthy countries. These are issues that have been problematic for OA.
The appeal for publishers is that, if it takes hold, S2O offers a rapid, predictable and stable way to convert subscriber-supported journals to OA.
For readers, opening up entire journals is an advantage over the institution-by-institution process that characterizes Read and Publish deals. Again, this is an equity issue.
What is the final destination?
RP: Once a backfile has been released will both it and that year’s content remain OA forever? If so, presumably the longer the program continues, the less incentive there will be for libraries to continue subscribing? Can you say something about that and share your views on the free rider issue that S2O introduces? Could it prove a serious impediment in the future?
RG: Once a volume is published OA, it will always be OA. Back volumes published before S2O was introduced will be behind the paywall with the exception of those volumes that make up the previous 10 years of content, which will be free to access.
I agree that as the program continues the accumulation of open content may be a disincentive, but a minor one. We have a package of incentives for subscribers that we hope will keep them engaged with S2O in the long term.
Also, as S2O becomes an established model for other publishers, our subscribers gain access to titles for which they are not part of the subscriber base. The motivation for institutions to contribute to an open environment, which is what librarians, authors, funders and publishers all want, by maintaining subscriptions will be strong.
There is no free-rider option. If even a small proportion of our subscribers took a “calculated gamble” to free ride, the S2O offering would fail, and access would require a subscription to paywalled content.
This is one of the essential points to get across to the library decision-makers. It has not been an issue in the pilot project.
RP: I have seen a lot of discussion about the threshold required before AR decides that a sufficient number of libraries have subscribed to justify freeing an S2O journal’s content for that year. Can you clarify exactly how AR makes this decision and what percentage of previous subscribers need to sign up to allow AR to make the year’s content free?
RG: The costs to publish the journals under S2O are the same as under paywalled access. To meet these costs, we need all existing customers to renew (setting aside the small amount of churn in our subscriber base every year that is due to occasional changes in research and teaching focus at institutions). Every subscription is vital.
We will also be seeking to add new subscribers, something that we will pursue more actively when we have usage information to back up the proposal.
We are also exploring multi-year contracts, consortia-wide deals and countrywide licenses, all of which would provide additional financial stability.
RP: You have I think indicated that there may be need for additional revenue streams if AR subscriber numbers start to wane. Can you say something about this? Also, do AR journals currently have sources of revenue in addition to subscriptions.
RG: We are looking at projects that support our mission and that could provide additional revenue streams, but it is too early to discuss them.
RP: What has been AR’s experience in attracting new subscribers to S2O journals to date? Have you seen an increase in subscriptions since you started opening the content or have the numbers fallen off? Do you expect this to change over time?
RG: We have limited experience so far. We have added new subscribers, which is quite thrilling to me, but we haven’t yet engaged in a serious campaign to identify major users who are not subscribers.
We will make a strong case for such institutions, academic and non-academic, to become subscribers.
RP: As you make clear, S2O is only sustainable if existing libraries continue to subscribe and/or new subscribers join each year. This suggests it will always be vulnerable to discontinuance.
Even if it continues, presumably there may be gaps in the historical record so far as OA is concerned (although I realise the complete record will always be available behind a paywall). Presumably, this means that, although more and more content will become OA over time, S2O journals will always remain subscription journals. In a sense, perhaps, one could argue that they are neither fish nor fowl.
Given this, do you envisage AR at some point making a full commitment to OA and changing the default status of its journals from subscription to open access? If you did, how would they be funded if there was no subscription revenue anymore, and no APCs? Could Diamond OA be an option? What do you see as the final destination here?
RG: Yes, S2O journals are viable only if supported by subscriptions. Likewise, the APC route to OA is viable only if the APC is paid, Read and Publish deals are viable only if the fee is paid, and Diamond OA projects will only exist so long as they are backed financially. Incidentally, if diamond OA is simply “articles/journals that neither charge users to read nor authors to publish”, per Wikipedia, then S2O is a diamond approach, but there do seem to be other, to me enigmatic, criteria.
Getting S2O off the ground is challenging because we need to maintain every subscription rather than doing a one-off deal. If all our titles are published OA in the first couple of years, along with the 80-plus titles from 11 other publishers, S2O will become established.
In that event, I predict that the S2O program will deliver highly stable OA publishing that is not vulnerable to discontinuance. This stability can be supported by multi-year contracts with subscribers, by countrywide deals with central agencies, or through contributions from funding agencies that currently underwrite APCs.
With apologies for the corny allusion, (S2O) journal publishing is a journey, not a destination. We must earn the right to publish S2O, by generating subscription revenue from academic and research institutions. They are our preferred fellow travellers, as we share their commitment to learning, research and public service.
RP: The Brief has said, “S2O may prove to be the kind of model that works well as long as it is not widely adopted”. Is there something to that?
RG: On the contrary, the more widely S2O is implemented, the more likely it is to be successful.
RP: As I understand it, AR commissions papers rather than inviting researchers to submit them. I guess that is why you say the APC approach could not be adopted by Annual Reviews and why it therefore decided to go with S2O?
RG: Yes, all of our content is commissioned (proposals from potential authors are welcome). And yes, this excluded APCs as an option. We felt we could not invite researchers to spend many weeks writing a review, and then present them with a bill to publish OA.
RP: Can you explain how AR deals with the licensing of papers in its S2O journals. I believe the standard licence used for new articles is CC BY, but authors are still asked to assign copyright to AR. Why is that?
RG: As of 2023, authors of new articles will retain copyright, the license will remain CC BY. We retained copyright during the pilot program just to allow us to get on with things – authors had already signed forms and we didn’t have the bandwidth to revisit them.
RP: What about backfiles: what licence is attached to them? Is this likely to change in the future?
RG: We retain the copyright on the backfiles, and they are not licensed CC BY. They are available through purchase of our Electronic Back Volume Collection.
RP: I think you sell copies of papers via Copyright Clearance Center (CCC). What level of revenue does that bring in? Is that revenue source likely to become more important over time or is it incompatible with the concept of open access, and so hard to justify?
RG: CCC revenue provides 3.5 percent of our income at present. The backfiles of our journals are not switching to CC BY licences, so almost all of CCC revenues should be preserved in the near- to mid-term. A gradual decline in this revenue stream is manageable.
RP: Unless I am misunderstanding something, CCC is also selling rights to CC BY-licensed papers published by Annual Reviews. Is this AR’s policy or a glitch in the AR/CCC websites?
RG: Thanks for bringing this to my attention. We had not been aware that the S2O journal content was being charged for by CCC. This is a legacy issue – there has been a single feed from AR to CCC that didn’t discriminate between the paywalled and open titles.
We will get this fixed and will invite CCC to attend an S2O Community of Practice meeting so that others are aware of the issue.
Work to be done
RP: AR has also launched a publication called Knowable Magazine to provide journalistic coverage of the real-world significance of science. You have described this as “a complementary approach to opening up our content.” What would you say to those who might argue this is really all that the public needs in terms of free access to scientific knowledge?
RG: Knowable Magazine is great, but it is no replacement for the open availability of expert-written summaries of research progress.
We carried out a small IP address analysis of users from Australia of the Annual Review of Public Health and identified downloads from government departments, health care organizations, banks, rural hospitals, grammar schools, police departments, pharmaceutical companies, fire and rescue services, family planning clinics, local government family services, an HIV resource center, a prison and many departments of public health.
Scientific reviews are not a casual read. Having them available, however, means that we and others can extract relevant information to present to different audiences to meet their needs.
We are pursuing such a project aimed at policy makers. Having the source review article available OA guarantees the credibility of these adjunct products and provides further information for those wishing to take a deep dive into a topic.
About one percent of Knowable readers take advantage of access to relevant reviews.
“cOAlition S is still reviewing how it will engage with S2O and is in discussions with S2O publishers. All of us in the S2O Community of Practice look forward to the results of their deliberation”
RP: S2O has been endorsed by both OA2020 and cOAlition S. AR is understandably using this to promote its S2O activities. I note, however, that The Brief has suggested that S2O does not provide a good way of achieving the goals of Plan S.
As The Brief puts it, “Let’s say an author submits a manuscript in July or August. The paper may not be published until January of the subsequent year (or later). At the time of submission, the journal may or may not have met its target threshold for the next subscription year. This puts the author’s paper in a kind of Schrödinger’s cat state of superposition where the paper may or may not be in compliance with funder mandates. What if the journal never meets the threshold and the paper is ultimately published on a subscription basis?” Do you see this as an issue?
RG: Plan S does not mandate review articles to be OA, so this is not an issue for Annual Reviews. Of course, it would be disappointing for an author and for us as publishers if we were forced to introduce the paywall, and we will do everything we can to avoid that from happening. For primary research, where there is a Plan S mandate for OA, APCs could provide a backup in the event of an unsuccessful S2O offering, albeit a less than perfect one.
cOAlition S is still reviewing how it will engage with S2O and is in discussions with S2O publishers. All of us in the S2O Community of Practice look forward to the results of their deliberation.
RP: I found it hard to locate OA content on the Annual Reviews’ site. Individual papers seem to be signalled as being OA (with the OA logo attached), but I do not think the OA logo is placed on the browse journal page or on the paper titles in the volume list (although it is mentioned in the “about journal” information). Could this be better signalled? Also, is it possible to search only OA content on the site? How do users find Annual Reviews’ OA content?
RG: We are currently preparing for an OA future, there is certainly work to be done to make it easier for users to find OA content.
RP: The benefits of open access tend to be expressed in terms of the number of times a paper is downloaded, and I note you said that AR has experienced a significant increase in downloads for its S2O content. But do you think this is an adequate measurement tool? Could it even be a deceptive one – I have seen suggestions that much of the increased usage when content is made OA is a result of activity by bots and crawlers rather than new human readers. While you say this is not the case with AR (which publishes review articles rather than research papers), I wonder if there are other, better ways in which OA’s success/benefits can be measured, or tools that might be developed to do this?
RG: Total downloads provide a useful shorthand metric, but they are not sufficient. Ideally, and without undermining the confidentiality of the user, we would like qualitative and quantitative usage information.
This includes where is the content being used – in what type of institution, in what country and by what type of user – and how the user rated the content in terms of how understandable and how helpful it was, and what impact it had on their task.
This would create a feedback loop that could help us improve our products. Other useful criteria are citations numbers and patterns, and alt metrics of all kinds.
Regarding the validity of usage data, as you say, we have evidence (mentioned above) that the accesses are by humans. Also, the patterns of usage of individual articles are more consistent with selection by people than by bots. For example, articles on topics of broad general interest get more non-subscriber usage than articles on specialized topics.
“I see S2O playing an important role in achieving a fully open scholarly literature, alongside existing and emerging routes to OA”
RP: As you noted earlier, one of the problems APCs create is that those without the money to pay to publish (notably researchers located in the Global South) discover that a paywall has been replaced by a publish wall. One of the attractions of S2O is that it avoids this but is there any evidence that it leads to higher submissions from those in the Global South and other unfunded researchers. Does S2O assist them share their own research in a more equitable and inclusive way? If so, how?
RG: As noted, all Annual Reviews content is commissioned, so I can’t answer that question from our experience. (As an aside, I will say that our editorial committees are taking steps to increase authorship from the Global South and from other groups that are underrepresented in our journals.)
Informal feedback from other S2O publishers indicates that they have seen an uptick in submissions from the Global South, but it’s early days and I haven’t seen any data yet.
S2O is an equitable and inclusive approach to OA funding, since it doesn’t require a payment from the author or that the author’s institution has a Read and Publish deal with the journal that the author wants to publish in.
RP: In the early days, it was assumed that open access would solve three long-standing problems with scholarly communication – i.e., the problems of accessibility, affordability, and equity. Aside from the continuing problem of equity, I think the biggest disappointment for OA advocates today is that open access has not (yet at least) turned out to be any less expensive than subscription publishing. S2O certainly seems unlikely to reduce costs. Can you share your thoughts on this?
RG: OA does not lower Annual Reviews’ costs. Staff salaries account for a large fraction of our budget, and we won’t reduce the staff by a single position by converting to S2O.
The other side of the coin is that it won’t add any new staff either. It’s cost neutral. I don’t see how this can be viewed as a disappointment, given the impact of moving to open.
For reducing costs, price transparency may have some minor impact but real savings will require fundamental changes to the nature of publishing.
RP: Do you think OA advocates have been somewhat naïve about the potential for OA to lower costs, or increase inclusiveness?
RG: For a traditional publisher like Annual Reviews to be fully embracing OA is a testament to the vision and credibility of OA advocates.
In the small corner of OA that I’ve been active in, namely the S2O Community of Practice, the approach is pragmatic rather than naïve. I see librarians, funders, and publishers working together to achieve a shared ambition of equitable and transparent open publishing. S2O does increase inclusiveness.
Regarding lowering costs, we should all maximize efficiency, but major savings will need a different approach to publishing.
In that regard, it will be interesting to see how preprints develop, and I think that eLife is doing very interesting things (though not necessarily inexpensively, at the moment).
RP: How do you see the future for open access, and the part that S2O will play in that future?
RG: Open access is tremendously important. The Covid-19 pandemic proved this and also demonstrated the ability of all publishers to make information quickly and widely available.
We need to make that the norm because climate change issues and many other societal challenges and opportunities will be similarly impacted by an open science environment, including open access.
I see S2O playing an important role in achieving a fully open scholarly literature, alongside existing and emerging routes to OA. As experience accumulates, some consolidation of OA models would be welcome, otherwise it will get very complicated.
RP: With the current growth in populism, the prospect of a collapse of globalisation, a new cold war (with perhaps both Russia and China), and what seems to be a rush to achieve national resource security and increased protection of national IP, is it possible that when the history books are written open access will come to be seen as a good and high-minded, but ultimately unachievable, objective?
RG: I sincerely hope not, as open access is essential, not just good and high-minded, given the challenges we face.
At the global level, it’s hard not to be appalled and discouraged by current events, but scientific and technical exchanges persisted through the Cold War, and the research community should insist on maintaining those now. Open access is something that could assist in this. In addition (at least in democracies), action on climate change or on public health emergencies requires broad-based support.
Open access to relevant research can help counter misinformation and disinformation and help create bridges between research and our social, financial, political and cultural institutions. The content can be adapted for different audiences (as Knowable Magazine does), but the open availability of the research literature is important for building trust.
RP: Thank you very much for taking the time to answer my questions, and good luck with your S2O plans.