As many Scholarly Kitchen readers will know, in late January it was announced that the Association for Computing Machinery (ACM) had signed new open access (OA) publishing agreements with four major US universities. There has been lots of public conversation about these agreements, but I decided to go to the source for some additional information. Scott Delman of ACM graciously agreed to respond to some questions.
Please tell us about your role at ACM, and maybe a little bit about what you’re currently most concerned about or focused on in that role.
First, let me thank you Rick for the opportunity to talk about some of the issues most important to ACM and the computer science community we serve. I really appreciate it. Over the years I’ve had a number of roles at ACM, including editorial management roles and responsibility for ACM’s Digital Library sales department, but for the past 3 years I’ve served as ACM’s Director of Publications with day-to-day operational responsibility for ACM’s Publications program and more generally responsibility for ACM’s strategic direction with respect to Publications.
The things that keep me up at night are both strategic and operational in nature. Obviously, open access is the most visible thing and potentially the thing that will ultimately have the greatest impact on ACM, our authors, and the scientists who read and use our publications in their work. But that’s not the only thing or necessarily even the biggest challenge we have in front of us. If I had to name the top three in no particular order, they would be (1) competing effectively against larger and better resourced societies and publishers in terms of authorship and readership (2) transitioning to a sustainable open access model “at scale” and (3) preparing ACM’s staff and volunteers for a publications future that looks quite different from the organization we have today. Of course, these are all connected and very much dependent on each other.
ACM recently announced a new transformative agreement with four major research universities. How did this deal come about?
This all started at an ACM Publications Board meeting in San Francisco in June of 2018, when we invited Jeffrey Mackie-Mason, the Dean of Libraries at the University of California at Berkeley, to participate in the part of our meeting that was focused on open access. During that meeting, we talked about some of the challenges ACM had with transitioning its publications to open access, with financial sustainability being the major one. It became clear that moving to an open access model at any sort of large scale would require the big research institutions like UC Berkeley to pay substantially more than they historically paid ACM for read-only access to our publications. Much to our surprise, Jeffrey responded favorably and gave us a very positive indication that UC Berkeley would be prepared to do its share to help ACM through this transition. At that time, there was no detailed model, just a vision and a hope to build on top of. Within about six months from that time, we started working with Ivy Anderson at California Digital Library towards developing a model. Over time, we also started working with Ellen Finnie at the Massachusetts Institute of Technology, Keith Webster at Carnegie Mellon University, Curtis Brundy at Iowa State University, and Sunshine Carter at the University of Minnesota toward the same goal. At a certain point — and I don’t recall exactly when that was, but probably around twelve to eighteen months ago — we started working together as a single group, and then last spring we decided the best way to take a bigger leap forward was to spend a full day in a small conference room at MIT in Cambridge, Massachusetts. During that meeting, we focused on some of the most important goals of each organization and the concerns our university partners shared about the kinds of Read and Publish models coming out of the large commercial publishers, which were all heavily based on individual Article Processing Charges (APCs). So, we started developing a model that relied less heavily on APCs and more on revenue replacement for ACM. The deal that was announced a few weeks ago is the result of dozens of video conferences, hundreds of emails, and a few in-person meetings at various workshops different members of the group participated in together during the summer and fall of 2019.
Sometimes there’s ambiguity or confusion about the term “transformative agreement,” in part because it’s not always clear what such an agreement intends to transform. In this case, what is ACM’s transformative goal?
Great question. ACM has two primary goals for this new model, which is now being referred to as ACM Open. The first is to transition ACM to an open access publisher for the entire portfolio of peer reviewed research articles we publish. Last year, this amounted to approximately 21,000 original research articles from over 40,000 authors, and this amount is growing between 5-6% annually across all of our journals, conferences, and technology focused magazines. Since 2013, we’ve been tracking the average downloads and citations of articles published in ACM’s journals, conference proceedings, and magazines, and comparing articles sitting behind our Digital Library “paywall” and those published in front of the paywall. As other publishers have found, our data is relatively clear that when an article is published on an open access basis it receives significantly more usage and citations, in other words “impact.” As a non-profit society publisher, our highest priority is to advance science and accelerate the rate of scientific discovery. To the extent that breaking down the subscription paywall can do that, we have an obligation to move in this direction. In this regard, we are entirely aligned with our members, authors, and the scientific community. However, some in the community do not appreciate that simply tearing down the paywall immediately would be the wrong thing to do, because it would eventually result in a lack of financial resources to continue publishing the community’s work. Communicating that to those who feel the most strongly about moving to a complete open access model immediately is not easy, but it’s absolutely essential because ACM is one of the few highly respected scholarly publishers in Computer Science and if we disappear as a result of a premature transition to OA, the only publishers left to serve computer scientists will be the large commercial publishing conglomerates, and we don’t feel that is in the best interest of the community. This leads me to our second goal, which is to manage the “transition” or “transformation” of our publications to an OA model in a sustainable way, so ACM can continue to maintain and grow its publications and invest in new publishing technologies and services, like the new ACM Digital Library platform that just launched on January 1, 2020.
There is also often confusion about the difference between “read and publish” and “publish and read” models. (Lisa Janicke Hinchliffe published a very helpful primer on this topic in the Kitchen last year.) Does ACM’s deal fall neatly into one of those categories, or is it something different?
I’m not sure if there is a formal dictionary definition that distinguishes between these two models, but for ACM, here’s what it means: First, it’s important to know that ACM publications are supported by selling access to its Digital Library to approximately 3,000 institutional customers from over 100 countries around the world. Since we only have 4 staff who are responsible for all of this licensing activity worldwide, we license access via institutional library consortia. These are effectively “buying clubs” that negotiate terms and pricing and purchase access in bulk. We work with approximately 100 of these consortia around the world, including North America, Europe, the Middle East, Africa, Australasia, and Latin and South America. When we looked at these deals and which institutions the publishing income and articles are coming from, we found that we essentially had three main categories of potential future open access “transformative” deals. The first category was institutions and consortia where the total deal value would be primarily based on open access publishing instead of “read access.” These were deals with the largest research universities that publish the most with ACM. Taking into consideration that many of these deals would need to see significant increases in overall income based on publishing output, we categorize these deals as “Publish + Read” deals. On the opposite end of the spectrum, a very large number of our subscribing institutions (~50%+) publish very little with ACM. These institutions are primarily concerned with having “read access” to our publications. As more of the Digital Library becomes OA, we expect these deals to decrease in value over time. Because “read access” is the predominant value proposition for this group, we categorize these as “Read + Publish” agreements. And then of course, there is a large group in the middle where the future projected income is split very much down the middle between OA publication income and read-access income. We don’t really have a name for these, but they are of course the “sweet spot” For ACM and account for probably 25-30+% of our institutional customers. They get good value reading and downloading articles and publish quite a few articles with ACM as well.
There’s been lots of chatter about this deal already on social media. Some are suggesting that a deal like this one could be implemented more or less universally between publishers and libraries. Do you agree?
That is great to hear and of course the “chatter” has been almost entirely positive. We’ve had some form of communication directly or indirectly with over 500 universities in North America, Europe, Asia, and South America over the last six months and we hear this sentiment a lot. They like the model and the pricing is extremely fair, but at the same time budgets are always the challenge, so the ultimate success of the model for ACM will depend on whether the largest research universities are willing to accept the reality that to join the model with ACM means a substantial increase in their overall spend. As you well know at your university, there’s enormous pressure to keep costs “neutral” in this transition to open access publication. For ACM, “cost neutral” is the goal at the global level, as we are not looking to use this model as a way to grow our publishing income, but rather as a way to sustain what we have today with nominal increases sufficient to support our publishing activities. But considering the unique financial challenges ACM faces with its current pricing model, this means that some universities will need to pay substantially more, some will pay the same, and others will pay significantly less as more of our Digital Library opens up. Will this model work for everyone? Unfortunately, this is a really hard question and the answer entirely depends on the publisher, what their own unique characteristics of publication and publishing income look like, and of course the relationship they have with their customers.
Given that the costs of doing business generally increase for everyone (certainly including libraries) over time, what kind of provisions does this deal make for cost increases from year to year?
During our work with the four universities mentioned above, we analyzed and shared a ton of ACM publications and financial data with the group. What some of this showed is that ACM’s Digital Library income comes primarily from approximately 3,000 universities paying for “read-only” access to our publications, and there is a fairly even distribution of income from large, mid-sized, and small institutions. This even distribution means that the largest research universities around the world pay approximately the same amount to access the ACM Digital Library and download published articles as the smallest universities and colleges. As we started analyzing the impact of different OA models, it became clear that we had a problem in that the vast majority of articles we publish each year comes from a very small percentage of our institutional customers.
Here’s some data to highlight the problem: the top 1,000 of our institutional customers represent 80+% of our published articles, but only 32% of our Digital Library revenues ($20-25M when factoring in individual member subscriptions, industry subscriptions, and government subscriptions). The problem gets more pronounced as we analyzed the data further. The top 250 institutions represent approximately 50% of our published articles, but only 11% of our Digital Library revenues and is even more extreme when you look at the top 100 institutions that represent approximately 33% of our published articles, but only 5% of our revenue. When we looked at usage-based metrics across our entire institutional customer base, such as “cost per download”, we found very similar percentages to the articles published percentages. The largest research universities around the world were downloading a tremendous number of articles compared against the long tail of smaller institutions, as one would expect to see.
For any publisher that derives most of its publications revenue from read-access licenses, this would be problematic, but when you are operating in a market that is moving quite decisively towards a model where the author/funder, or university, is paying for the OA publication on the front end instead of paywall access on the back end, this becomes even more problematic, because as we move from a “reader/library pays” model to an “author/funder pays” model and more and more of our publications become OA in the ACM Digital Library, this long tail of small institutions are likely to cancel or put increasing pressure on ACM to reduce pricing, which of course would be a reasonable thing to do from their perspective. Taken to the extreme, when all of our original research articles are OA in the ACM Digital Library, which is of course the ultimate goal, that means that approximately 70% of ACM’s Publications revenue is at risk. This was a challenge for ACM before the OA movement began roughly 15-20 years ago, and we are committed to addressing this challenge as part of our overall OA transition strategy to minimize the risks I just mentioned.
As you look into the future, how do you anticipate that you and your partners will assess the success of this deal?
For ACM and for our community, there is a lot of pressure to transition our research publications over to a sustainable OA model in the next 5-7 years. We think this is achievable, but it depends how quickly the large research universities adopt our model and sign on with ACM. Success from my perspective means reaching a tipping point. Quantitatively, that means something between 250-350 of our largest university partners coming on board, which would make approximately 50-60+% of the articles we publish each year OA and put us in a strong position to decrease pricing for the long tail, which by the way is a commitment we make contractually to all ACM Open partners, not just a statement in an interview like this. Qualitatively, this means that those institutional partners have figured out the “funding challenge.” The partners we’re already working with now are committed for a three-year period and most of the potential partners we’ve talked to are also likely to make a three-year commitment. But during and after that first three-year period, the challenge will be to fund these deals with ACM and others over a much longer period of time. That means tapping into large and growing sources of funding. The most logical source is the research funding dollars that university researchers obtain to conduct their research — places like the National Science Foundation and the Department of Energy in the United States, the European Commission, and research funding from government and private funders elsewhere around the world. Will libraries or the universities themselves be willing and able to establish these kinds of OA fund at scale is unclear today, but that is what I believe is going to need to happen for these kinds of transformative deals to not only happen at scale, but to happen over the long term. Thanks for giving me an opportunity to share some information about our new model and where we’re heading with open access.
8 Thoughts on "ACM’s New Open Access Agreements: A Q&A with Scott Delman"
A very interesting article highlighting the holistic challenges faced by scholarly publishers boards. I guess Scott will have to make some drastic strategic trade-off. Thank you Scott for sharing your strategic decision process. Good luck.
Sebastien, thanks for the feedback. Can I ask what you see as the “drastic strategic trade-off” we need to make here?
Hi Scott. I guess you will have to concentrate all your resources on 1 customer segment, the 250 you refer to since you know very well their usage and needs. The other customer segment might be better off using other solution provider. We faced the same situation in a company I worked for some years ago. Different industry. Still, it helped us to focus. BTW, What about coming up with a service to help your key accounts to get the extra funding they need to finance the premium you ask for? Be careful, these might be too generic strategies only relying on your above info. You know your business more than anyone else.
I almost forgot to address Sam’s second point. In reality, ACMs publications revenue is not significantly higher than our publishing costs. The numbers presented in ACMs annual reports do not present a totally “functional” view of all direct and indirect expenses and some in the community have called for more transparency on this, so we’ve done quite a lot of work over the past few months on this to take all publishing related costs into account and will be providing more information about these figures publicly over the coming months.
Thanks for this. While we are, of course, talking to the large research universities in the US, Europe, and Asia first, we are taking a multi-pronged approach to the roll out and using our extensive network of consortia and subscription agents to promote the model to our mid-tier, and long tail university partners as well.
Unfortunately, this interview doesn’t discuss two important questions that need to be considered when talking about open access at ACM. First, why is the cost so high? We know that the per-article cost at open access publishers such as arxiv and Dagstuhl are an order of magnitude smaller than ACM. Reducing ACM’s excessive costs would make these agreements more affordable for everyone. Second, ACM publishing revenue is much higher than publishing costs, with the surplus going to subsidize the ACM in general. My sense is that few people who publish with the ACM think this is the right approach, and would like publishing costs to reflect the costs the ACM actually incurs.
Sam, thank you for your question. First, the cost per article published is actually significantly lower with the ACM Open model than any other publisher out there. Compare the recent deals announced by Springer, Wiley, or Elsevier or talk to a number of the large societies that publish original research for the computing community. Most are charging in the $2,500-$3,000+ range per article published. For institutions that participate in ACM, the effective per article price would vary depending on which Tier the institution falls into based on the number of articles they publish with ACM each year. Take Carnegie Mellon, for example. As a Tier 1 institution, they will pay $100k annually, but last year they had over 230 articles published with ACM where the corresponding or lead author was from CMU. If you do the math, they are paying less than $500/article and that also includes unrestricted access for all CMU factually and student to everything ACM publishes.
I also noticed you listed arXiv and Dagstuhl as publishers to compare price against, but this is not an apples to apples comparison, because these are not publishers. ACM has a long history and deep relationship with arXiv and I also participate as a member of the arXiv advisory board. While there’s no question that these are important services and valuable to the community, arXiv is a pre-print service that serves primarily as a “host” of content that is either not yet published or hosts versions of published works that are published by actual publishers. arXiv provides a valuable service, but that service is only one aspect of what scholarly publishers do. Some of those other services publishers like ACM provide are peer review (systems, software, organization), typesetting, rich platform features like those found in the ACM DL, rich metadata tagging including DOIs, participation in various discovery services beyond Google, author marketing, long term digital preservation, ethics & plagiarism policies and infrastructure to investigate claims on behalf of authors, rights management, etc. the list is actually quite long and while I appreciate that some of this is quite invisible to the community, authors who work with ACM and other publishers do understand the distinction. There is also significant value for authors in the branded publications ACM publishes themselves. Over time, different communities get to know which are the high quality journals and conferences that are looked most highly upon by tenure committees and colleagues in a Niche community. Submitting to those publications with respected editorial board and EIC that will give valuable feedback to authors is an essential part of the publication process and value for authors. It’s absolutely true that posting in arXiv can also result in feedback from the community, but submitting to specific journals or conferences ensures targeted feedback from peers. In any event, ACM is a strong supporter of arXiv, but we don’t think it is a competitor or an either-or proposition for the community and the comparison from a cost perspective is simply not accurate. I’m happy to discuss this in more detail offline. Please feel free to reach out to me.
Fair point regarding ArXiV. What about Dagshul? Would you say this a not a “publisher”? The APC are 100$ there (last time I checked) with board of highly regarded researchers. How would you explain/justify the cost difference?