Annual Reviews announced today that the 2020 volume of the Annual Review of Cancer Biology has been published open access and that the back volumes of this journal are also now available for free reading. As the pioneer of the Subscribe to Open model, congratulations are due on achieving their first open title. The 2020 articles are published copyright to Annual Reviews with a CC-BY license. The backfiles do not carry a CC license. Annual Reviews developed their Subscribe to Open model in partnership with Raym Crow, Managing Partner, Chain Bridge Group, with funding from the Robert Wood Johnson Foundation. As interest in Subscribe to Open grows based on the experiences of early innovations, publishers and libraries need to develop an understanding of the various approaches to Subscribe to Open and the benefits and limitations of the model.
The Model
Subscribe to Open is an example of an assurance approach to addressing a collective action challenge. In the Subscribe to Open model developed by Annual Reviews, each subscribing library is motivated to continue to subscribe (because they have been a subscriber and as such have already made a decision that the content is worth paying for) by a discount that is built into the Subscribe to Open offer. The model is two-fold. First, if all libraries continue to subscribe, then not only will those libraries have access to the content for their users, but Annual Reviews will also make the content openly available to non-subscribers as well and apply a CC-BY license to the articles. Second, if all libraries do not continue to subscribe, then those that do will still receive the discount — as well as access to the content — but the content will not be made available to non-subscribers. In either scenario, the subscribing libraries receive a discount and access to the content. Essentially, this is a no-risk opt-in for the subscribing institution. Martin Paul Eve has outlined a similar possible model for society publishers but with a three year rather than annual timeframe.
Ultimately, Subscribe to Open builds easily from the established subscription system, maintaining many of its affordances while also opening content. Curtis Brundy of Iowa State University shared with me his view that, “Since it is based on the current subscription system, it is relatively easy to adopt for publishers and libraries. The annual recurrence offers a way to control free-riding. And since it isn’t article based, it doesn’t erect a barrier in front of authors.”
Subscribe to Open also offers an additional benefit to publishers that may not be immediately apparent. As a model, it sidesteps the additional demands that libraries and funders are increasingly requiring of open access publishers, which they do not demand of their subscriptions. The most obvious of these currently in play is copyright retention by authors. Many libraries and funders require copyright retention as a principle for funding, for example, an APC. Some Subscribe to Open models include copyright retention by authors but not all. For example, Annual Reviews is still requiring copyright transfer at this point. Richard Gallagher, President and Editor-in-Chief of Annual Reviews, explained that, “During this pilot phase of S2O, we are asking authors to transfer copyright…because articles started the copyedit and finalization process months before we knew if S2O would be successful.” Though Gallagher said that Annual Reviews looks to develop a workflow that includes copyright retention, by having the copyright transferred during the pilot, Annual Reviews “can quickly alter the license to CC-BY” after the Subscribe to Open threshold is reached.
Growing Interest
Annual Reviews is not the only publisher exploring Subscribe to Open. Raym Crow shared with me that he “is in discussions with several society and university press publishers interested in exploring” the model. In January, Berghahn Journals announced the success of their Berghahn Open Anthro Project, which worked with Knowledge Unlatched to implement their Subscribe to Open model. 13 journals in anthropology will be published open access during the 3-year pilot and authors retain their copyright and choose among three CC licenses for their work. The backfiles of the Berghahn journals do not become freely available via Subscribe to Open. And, just last week, EDP Sciences announced Subscribe to Open for its journal, Mathematical Modelling of Natural Phenomena. In the EDP Subscribe to Open mode, for new articles, authors retain their copyright and publish their work with a CC-BY license and the backfile of articles is available for free reading.
Challenges and Limitations
Perhaps obviously, Subscribe to Open is only a model for funding open access publishing for those publications that have subscribers. This seemingly obvious statement is important to unpack because the number of publications that might be offered under this model is more limited than it might initially appear. First, of course, this model is not possible for native open access publishers or any publications that are currently “Diamond” journals that are subvented in some way. These publications do not have subscribers.
Second, many journals do not have subscribers in the way that Subscribe to Open requires. For example, Michael Clarke has detailed how society journals published in partnership with commercial publishers likely no longer have a subscriber list as those publishers focus on selling packages of content and not specific titles. Read and Publish agreements and other approaches that support hybrid approaches (open access and paywalled articles in the same journal) further limit the possibilities for specific titles to adopt Subscribe to Open because of the complexity of avoiding double dipping in general, much less in the context of a Subscribe to Open offer. In essence, if a journal is already offering APC-based publishing or is part of a transformative agreement, pursuing Subscribe to Open would require first reverting to being subscription only or the enterprise of pursuing Subscribe to Open will be derailed by administrative complexity and suspicions of double-dipping.
In addition, it must be acknowledged that some journals are more valued than others. High value titles that libraries are unlikely to cancel in challenging budgetary times are better candidates for long term success with Subscribe to Open. These titles are not likely to induce any initial response of libraries risking non-renewal to see if others renew sufficiently and they are also less likely to revert to being paywalled because of cancellations over time. A publisher may not want to risk a journal see-sawing between open and closed year to year (both a public relations and logistical challenge) and so will likely either not offer marginal titles in this model or only do so by bundling those with other more valued titles, which of course recreates one of the aspects of the Big Deal for subscriptions that libraries are increasingly rejecting.
Early innovators with Subscribe to Open may also be experiencing some first-mover advantage that will not be enjoyed long term. Libraries signing on to successful Subscribe to Open campaigns not only ensure they have continued access to the content but also benefit from the public relations benefit of being promoted as a supporter of open access. As such, libraries that were not previously subscribers may sign up in order to support open access even if they had not been subscribers previously. And, indeed, according to Gallagher, the Annual Review of Cancer Biology picked up some new subscribers, “solely in support of the program.” The draw of this signaling function will dissipate over time as Subscribe to Open grows and as libraries increasingly have open access investments that they can point to without signing up for the latest offer. As a consequence, in the long term, Subscribe to Open limits the potential for revenue for publishers, since after this initial period there is little further expectation of subscriber growth, and the only way to raise revenues will be to increase prices.
Finally, inducing renewals of Subscribe to Open subscriptions may prove challenging for even highly-valued titles. It is possible that a subscribing institution may see a drop in the usage of Subscribe to Open titles on their usage reports. This could occur because, even if the institution’s users are using the content at the same or even higher rate, that usage may not be able to be attributed to the institution. Users will no longer have to pass through an authentication system to get to the content and so any off-campus, mobile device, etc. usage may end up untraced to the institution. If libraries see a drop in their institutional usage, even if overall usage of the content goes up due to open access, they may reconsider their subscriptions. Though publishers may attempt to offset any attrition by signing on new subscribers, converting a non-subscriber to a subscriber is likely to be a heavy lift.
The Future of Subscribe to Open
Subscribe to Open is a model in its infancy and its future uncertain. The status of the remaining four titles in the Annual Reviews pilot program is not yet known and Annual Reviews has not yet determined if it will offer additional titles in 2021. And, regardless of the specifics of a particular publisher’s Subscribe to Open offer, a central component of them all is that, should libraries drop their subscriptions, the journal(s) will revert to their default status of closed access for new articles. Only in a few years will we know if 2020 marks the beginning of a growing program of success in sustainably funding open access or an experiment from which to learn.
Discussion
25 Thoughts on "Subscribe to Open: A Mutual Assurance Approach to Open Access "
As someone who has always kind of struggled to get his head around this model, I really appreciate this piece, Lisa. There are still some things that confuse me, though. For example:
First, if all libraries continue to subscribe, then not only will those libraries have access to the content for their users, but Annual Reviews will also make the content openly available to non-subscribers as well and apply a CC-BY license to the articles. Second, if all libraries do not continue to subscribe, then those that do will still receive the discount — as well as access to the content — but the content will not be made available to non-subscribers.
If I’m understanding this correctly–and I freely admit that I may not be–it means that “subscribe to open” only results in OA if all existing subscribers continue their subscriptions. If at least one existing subscriber cancels, then access goes back to being subscription-only.
Is that correct? Or is it less categorical than that — is there some threshold level of continuing subscriptions (something less than 100%) that would still allow the content to be made open? Or does this vary somewhat from deal to deal?
Hi Rick,
It is less categorical than that, at least in the case of Annual Reviews. If Subscribe to Open is to bring stable open access to our journals it requires some level of robustness and we are working to build this in. Of course, all publishing models deal with a certain level of churn in subscriptions/ subscription revenue.
Richard
Thanks, Richard. It would be interesting to examine the terms of a variety of subscribe-to-open deals and see what the threshold variability looks like across them.
I would think that keeping the exact threshold level a secret is part of the strategy for an S2O model — you want everyone who can to subscribe, and if you set some public limit and make clear that the limit has been reached, or that there’s ample room for cancellations before that threshold gets crossed in the other direction, it will only inspire reduced revenues.
I don’t think anyone is saying this is a no-risk strategy, FWIW. So …
Yes, but on the other hand, if you cancel and that puts the S2O below the threshold for opening the content, then you have the problem that you are no longer a subscriber and so have lost access because the content is now paywalled and you can’t get past the paywall.
Anticipating you are now going to say “ok, so just subscribe after that point” … yes, you can. But if that’s what you are going to do, they strategy wise you should just go ahead and renew under S2O and save yourself the labor of canceling, then adding, etc. That labor, etc. is costly, the process is then outside of standard workflows, and it would mean your users are still without the content for at least a couple of months. There is a reason Curtis said what I quoted in the piece!
What is a bigger challenge is that your usage reports are possibly going to show a drop and library’s often drop what isn’t being used as much. And, the publisher must have either built in a cushion to build up savings or possibly face having to increase prices substantially suddenly if there is an unexpected one-off expense (e.g., new servers).
Anticipating you are now going to say “ok, so just subscribe after that point” … yes, you can. But if that’s what you are going to do, they strategy wise you should just go ahead and renew under S2O and save yourself the labor of canceling, then adding, etc. That labor, etc. is costly, the process is then outside of standard workflows, and it would mean your users are still without the content for at least a couple of months.
But if you’re a librarian forced with the choice of cutting what you consider essential publications from your budget (and this is often a choice librarians must make), wouldn’t you favor cutting the one where you still might have access after making the cut? Isn’t the potential extra labor worth taking that gamble?
My concern is less the extra work than the disappearing budget. I would suspect that once a library decides to stop voluntarily paying for something they can get for free (discussed here: https://scholarlykitchen.sspnet.org/2019/10/09/roadblocks-to-better-open-access-models/), that money will be spent elsewhere and be gone from the budget, regardless of any extra effort that gets put in. Although one could just play a waiting game, and hope other libraries will pony up the money to get the journal back to OA eventually.
Hence my point about how this model is a better match for highly valued content (i.e., content a library wouldn’t want to risk not having).
I would suspect that once a library decides to stop voluntarily paying for something they can get for free, that money will be spent elsewhere and be gone from the budget, regardless of any extra effort that gets put in.
And here I think it’s worth pointing out that this decision could very easily be taken out the library’s hands. Any OA publishing model that relies on libraries continuing to pay voluntarily is endangered by the scenario in which a university administrator starts asking “What are we getting in return for this money that we’re giving to the library?”. Many of my library colleagues believe that administrators will find the answer “We’re contributing to the systemwide upkeep of an OA publishing environment” a compelling one, but I’m not sure how confident we should be about that.
Rick, I think the genius of S2O is that a library is actually paying for something – a guarantee that its users have access to the content, whether others also then do or not. Now, whether this can be explained well enough to keep what you are worried about from happening … that’s a different question. But, that it is processed as a subscription rather than a membership payment or the like also bodes well I think.
Rick, I think the genius of S2O is that a library is actually paying for something – a guarantee that its users have access to the content, whether others also then do or not.
Right, and this will be a compelling argument the closer the subscription-threshold level is to 100%. In other words, if no library can cancel without the toll gate coming down again, then my answer to the administrator is “I have to keep paying this subscription fee or we’ll lose access.” However, the lower that threshold gets, the harder it is for me to make that argument. A more Machiavellian administrator (or, to be fair, one who is faced with more difficult choices about allocating campus resources) is going to answer “If _we_ were to cancel, how likely is it that we’d lose access to the content?” My answer to her would have to vary depending on whether that threshold level is 99% or 95% or 80% or whatever.
Definitely a risk analysis moment … and then weighed against the labor/admin cost of re-establishing the subscription and all it entails to set up a “new” resource in our various authentication, etc. systems and the PR issues if it turns out access is lost. Why it matters as well if the content is highly valued/highly used.
Rick, FYI that EDP Sciences has stated a threshold “EDPS considers that the minimum amount of subscriptions necessary to sustain the publication of a journal is attained when at least 95% of subscriptions are renewed year-on-year.” I suspect that means that 95% of the previous year’s subscriptions rather than specifically 95% of the libraries that were subscribers (e.g., if your library drops and mine adds, there’s a substitution effect) but it doesn’t say so explicitly. See “When does EDPS consider that the minimum amount of subscriptions is attained?” on this FAQ: https://www.edpsciences.org/en/faqs-the-edp-sciences-subscribe-to-open-programme
Very helpful link, Lisa, thanks.
Rick, I’ve just started playing around with an idea that library collections can be understood conceptually to be built upon four layers: the isolated collection, the (resource) shared collection, the Big Deal collection, and the open access collection. Each layer exists alongside the other layers and are not mutually exclusive of one another. That said, if we can understand library collections in this way, each layer subsequently increases the access to information that libraries provide to their local communities.
Earlier, you said that library budget/funding decisions “could very easily be taken out the library’s hands. Any OA publishing model that relies on libraries continuing to pay voluntarily is endangered by the scenario in which a university administrator starts asking ‘What are we getting in return for this money that we’re giving to the library?’. Many of my library colleagues believe that administrators will find the answer ‘We’re contributing to the systemwide upkeep of an OA publishing environment’ a compelling one, but I’m not sure how confident we should be about that.” Perhaps our responses to university administrators could instead be something like, “As financial stewards, libraries have and will continue to support scholarly publishing by providing access to information for our users. However, the transition from information scarcity to information abundance requires that the way libraries go about providing that access has changed from that of supporting local subscriptions to content to that of supporting the global creation of content.” Our administrators must understand that in this age of information abundance, the way that libraries support local collections is by supporting global collections. Thus, perhaps librarians need to reconceive our collections budgets as “collections as creation” budgets.
However, this response doesn’t directly address your underlying question about what it will cost an institution that is considering dropping financial support for a publisher/publication. This risk, will need to be negotiated amongst collaborating/supporting libraries and the publishers. Perhaps libraries can draw parallels with the stock market. That is, libraries can invest directly with a publisher/publication (like stock investments), invest as part of a consortia (like mutual funds), and those consortia can invest collectively (like mutual funds of mutual funds). Hopefully I didn’t butcher that completely since I don’t understand the stock market very well, but perhaps this offers a kind of model that libraries can use. Anyway, I digress…
Sad to have missed your keynote at the recently cancelled SCELCapalooza!
Back in 1979 the Enquiry into Scholarly Communication, a reported funded by Ford, Rockefeller, the NEH, and other sources (and published as a book by The Johns Hopkins University Press), included among its key recommendations that more universities contribute to the sustenance of the academic publishing system beyond just purchasing books. That recommendation never was acted upon to any noticeable degree, so the system that had grown to that point, with some 80 universities providing the university press system for the benefit of all university faculty everywhere, continued on as it was. with a few smaller presses closing since then and a few newer ones added to the system. Your question could thus be asked of this system. Why do just some universities bear all the burden of making publishing outlets available to the faculty of universities that do not support a press (often at a loss)? Some administrators may cite the “free rider” problem in justifying why their universities do not want to support a press, but I think most that already have presses on campus would accept the argument that they are rightfully playing a role in supporting the entire system that benefits everyone.
One of the questions that I keep coming back to is whether research departments will be okay with the library administering funds for their publishing. One assumes a significant amount of this will be funded by the grants received by individual researchers, and in some countries and circumstances, by block grants to universities. Does the Chemistry Department want librarians making decisions about where and how its faculty can publish? Does the History Department want librarians making the call on whether their chosen book publisher is affordable?
As you note, we’re currently in a transition phase, shifting funds from paying to read articles to paying to publish articles, so in some ways it makes sense for the library, which is already handling these payments, to manage that transition. But once finished, does it make sense to centralize those payments through one entity or will each department want to handle their own funds and set their own rules? Or if it makes sense to centralize, is the library better suited to manage purchasing decisions than the university’s purchasing department?
I would remind David and others that when university presses at state universities like California were first started, they were intended to serve as outlets for the publications of faculty in all departments. So they operated then as general service agencies for the whole university, a role that seems to have fallen to libraries in our digital age. The California Digital Library, I believe, publishes some series for some departments, but does not attempt to service every department.
Trying to reply to Sandy and David but WordPress’s “Reply to Comment” seems to be wonky.
Sandy, I don’t know much about the history of university presses, but it seems like you’re saying that libraries failed to seize opportunities to hold ourselves responsible for the health of university presses. Is that accurate? If so, then I think we’re in agreement that perhaps libraries should reconsider this as part of our responsibilities to support scholarly publishing. I hadn’t thought of it this way before. Thanks for bringing this up!
David you said, “One of the questions that I keep coming back to is whether research departments will be okay with the library administering funds for their publishing… Does the Chemistry Department want librarians making decisions about where and how its faculty can publish? Does the History Department want librarians making the call on whether their chosen book publisher is affordable?” I’m not talking about supporting APCs or BPCs at the article- or book-level. Rather, I’m imagining offering financial support at the publisher or publication level. Wearing my library liaison hat and working at a mid-sized, liberal arts institution, I work with faculty to
decide where to spend subscription monies to support faculty teaching and student learning needs. In an open access world, maybe instead I work with faculty to decide where to spend publishing investment monies to support publications, publishers (including presses). As liberal arts faculty, we recognize the value that niche publishers and publications offer. So, in short, the library doesn’t “decide” on publishing activities any more than it decides on subscribing activities, and the library continues to collaborate with faculty on publishing activities as it has on subscribing activities.
I also don’t want to forget to say thank you to Lisa for your post!
Thanks for this very clear and insightful piece, especially on the limitations of the model. I would like to add two points on the assurance.
1/ The assurance model is also present at least in some Read & Publish agreements. De facto, big publishers don’t get anymore paid for reading but get the “subscription price” back through Hybrid OA payment. For a complicated “double-assurance” deal see the Springer-DEAL agreement with thresholds over and under the “substrction price” which resulted in a €2750 “fee par article”, see this analysis
https://polecopub.hypotheses.org/1587
2/ More generally, S20, just like the OLH model or any “participative payment model” has the huge advantage of being also an assurance for paying institutions. Rather than getting into the wild with models depending on the level of production in which they could pay much more (mostly APC, but new deals like ACM share the same “producer pays” flavor), they have the assurance of a flatter distribution of payment amongst libraries/institutions.
I appreciate these comments. The use of the word “assurance” here is in a more specified/technical sense. Consortial (DEAL) and membership (OLH) agreements are not assurance in this sense. But, yes, taking the more general usage of the word, any payment/contract is assurance in some way. But, the assurance comes not from the structure of the inducements to participant but from the contracts themselves.
Thanks for the article Lisa. A couple of responses:
“….in the long term, Subscribe to Open limits the potential for revenue for publishers, since after this initial period there is little further expectation of subscriber growth, and the only way to raise revenues will be to increase prices.”
No, these are open question. We will be encouraging organizations that have substantial usage of the Subscribe to Open titles to become subscribers.
“The status of the remaining four titles in the Annual Reviews pilot program is not yet known and Annual Reviews has not yet determined if it will offer additional titles in 2021.”
Annual Reviews publish throughout the year and we collect orders right up to the publication date. To maximize subscriptions, we are announcing the results at the time of publication. The situation for the other titles is encouraging at the moment.
We shall determine whether to offer additional titles based on the subscription and usage data collected through this year.
“If libraries see a drop in their institutional usage, even if overall usage of the content goes up due to open access, they may reconsider their subscriptions.”
We are confident that usage at subscribing institutions will increase based on how search engine algorithms rank content.
Thanks Richard. I’ll take your comments in order.
1) While it may be an open question for AR specifically re if this limits potential for revenue growth, S2O inherently does present this limitation. A publisher cannot rely on S2O alone for increasing revenue – as you mention by pointing to the AR plan to present encourage organizations that have substantial usage of the Subscribe to Open titles to become subscribers. The “ouch” phrasing I’ve heard people use in responding to this is – “so the follow up is some sort of shakedown – like is AR going to try to publicly shame libraries into paying?” I have said that I don’t think there will be public shaming! But, given that reaction, I think the possibilities here are limited for increasing revenue. I think it is also important to recognize that the strategy you are pursuing requires infrastructure for tracking and analyzing use without authentication and that not all publishers have that capacity. FWIW, though, I agree though that this is en empirical question to be tested over time (see my concluding paragraph).
2) Ok. I think you are just confirming what I wrote?
3) This I would really like to hear more about. How are you going to attribute usage that comes to the site via search engines to institutions? Much of that usage (extrapolating from studies documenting that people are often off campus and not authenticated) will come from IP addresses that are not recognizable as being in the namespace of the user’s institution.
It’s important that libraries prepare for a likely “drop in the usage of Subscribe to Open titles on their usage reports… [since users] will no longer have to pass through an authentication system to get to the content and so any off-campus, mobile device, etc. usage may end up untraced to the institution.” Perhaps voices from smaller, liberal arts institutions provide a strength or complement to those offered by larger institutions.
Liberal arts institutions cannot rely solely on quantitative data to assess the value of our diverse programs. Philosophy for instance, may have fewer faculty, fewer courses, and fewer majors but that does not devalue what it offers to our institution and students. Similarly, if cost-per-use stats are ill-suited to capturing the value that open investments offer to our local communities, perhaps libraries need to consider alternative (and perhaps more qualitative) means to assess such value. Furthermore, if institutions support the creation of scholarly content, and the value that niche publications offer our faculty scholarship and student learning, perhaps S2O and similar initiatives provide ways to engage our faculty in the stewardship of library “collections as creation” funds.
Subscribe to Open is a good fit for Review Journals since they publish a fixed number of articles for a reasonably-predictable cost and readership. In the case of Review Journals authors are invited and often paid, so APCs are out of the question (unless you run a vanity press). Regular research journals process a variable number of articles and face greater risk in terms of cost and readership.
Richard Wynne – Rescognito
Gee, maybe university presses should announce that if they get a threshold number of orders for a new monograph, say, 300, from libraries, then they will make the monograph OA so that everyone can benefit. How does this differ from the approach described here to journal publishing? It sounds liker a subsidy system to me. Isn;t that how Knowledge Unlatched works?
