writing a letterMy colleagues and I spend a great deal of time advising professional society publishers on their programs, and not infrequently open access (OA) comes up in the discussions. Sometimes this is because OA is the focus of a project, sometimes it is because OA is on the minds of the society’s leadership. What we find overall is that OA is little understood by society members, and often feared. (I have written about this before on the Kitchen.) We have found it useful to send along a briefing memo or to make a presentation to the governing board about OA, to take questions, and to quiet frazzled nerves. This is what we tell them:

Dear Society Board:

We want to provide some background on OA because it came up, unprompted by us, in several of the interviews we conducted. To many members of your society, OA is “out there,” lurking dangerously, waiting to pounce. Since a significant portion of the society’s income derives from its toll-access publications, a world in which OA has completely obliterated traditional publishing models is far from welcome to many society publishers.

We spend a great amount of time in the OA world — though not, we hasten to add, in the echo chamber, where OA is believed to be destined to save the world, perhaps by tomorrow. Our view, in contrast, is that OA is not inevitably going to wipe away toll-access journals. Rather what we are seeing is that OA and toll-access publications have developed a somewhat anxious coexistence. (See “Additive, Substitutive, Subtractive.”)

This does not mean that OA is something your society should ignore. Our recommendation is the opposite: you should get engaged with OA to forge new, profitable services. Furthermore, we believe that profit is the appropriate measure of an OA service because it implies that the community values the service highly enough to allocate resources to it.

Some numbers on OA:  The total market for scholarly journals is thought to be approximately $10 billion a year. At this time revenues for OA are running around 3-4% of the total, that is, around $300 to $400 million. That’s not a negligible amount, but it is of course a tiny piece of the overall journals market. Another number to note is that the average amount of revenue journal publishers receive per article published is about $5,000. This number is derived by a back-of-the envelope calculation: divide the total market size of $10 billion by the number of articles published each year (about 2 million) and you get $5,000. We recommend that you do this very calculation on your own journals:  total revenue divided by number of articles published each year. Some OA services charge authors under $1,000 to publish an article; PLOS ONE, one of the market leaders, charges $1,495. Some long-established publishers with big brands have OA programs that carry author charges of $3,000 and higher. But it’s not hard to see the problem here: it would be very difficult for a publisher used to receiving $5,000 or more per article under the toll-access model to get by on $1,500. So if OA became the norm, which we don’t anticipate anytime soon, publishers would indeed be under considerable pressure.

Ironically, the largest commercial publishers, which are regularly derided for their greed in library circles, have revenue per article near the average. For example, the last time we sharpened our pencils on this matter, we discovered that Elsevier received a bit over $5,000 per article. The highest figures are to be found with not-for-profit professional societies. We worked with one not-for-profit society client that receives $16,000 per article. Some receive much more.

OA is not actually a business model but a property of a publication, namely, it designates content for which there is no charge for access. In some contexts, though mostly not in STM publishing, “open” means not only free to the end-user but also configurable by that user. This configurability is anathema to scholars who value the definitive, fixed text for their research. Open configuration is a feature of the OER — Open Educational Resources — movement that is beginning to penetrate the college classroom market.

For STM publishers there are perhaps five business models, not of equal viability, that have the property of OA:

  • Green
  • Gold (author-pays)
  • Platinum
  • Hybrid
  • Advertising-supported

In the Green model authors seek publication with publishers operating under the traditional model and then deposit a copy of the article in institutional, subject or funder repositories. Green OA is thus parasitic: it depends on the toll-access model of the traditional publisher for its existence. Publishers have feared that Green OA would result in libraries cancelling their subscriptions, but in fact this has not happened to any significant degree. One reason for this is that articles are increasingly sold in bundles, i.e., as part of a “Big Deal: even if a particular article is available as OA somewhere, libraries still want to license the rest of the package.

The Gold model is the most robust and accounts for the lion’s share of OA revenue. Sometimes called the “author-pays” model, a Gold publication is one where the author (or a funder sponsoring the author’s research) pays to publish. The article is then made open to one and all. Gold OA was pioneered by a for-profit firm, BioMed Central, which is now owned by Springer Nature. The Public Library of Science dramatically raised the profile of Gold OA with its hugely successful PLOS ONE service. That service has been copied just about everywhere, leading to sharp growth in profitable OA publication.

In the Platinum model neither the user nor the author pays. It would be easy to dismiss this model as utopian, but in fact a small number of institutions, philanthropies, and government agencies take on the financial responsibility for a limited number of journals. This model does not scale, however, as there are not enough sponsors to support a large amount of research output.

Hybrid models typically involve layering a Gold OA option onto a toll-access journal. Thus a journal that publishes ten articles in an issue may find that two or three authors may pay a fee to make their individual articles OA. A variant of the hybrid model is the cascade, in which articles that don’t quite make the cut at a highly selective toll-access journal are published in Gold OA form by an affiliated journal. Cascades can be highly lucrative.

Finally we have advertising-supported OA. We include it here mostly for the sake of completeness, but in fact no scholarly publisher has made a success of this model; it is simply too hard to sell online advertising.

Our recommendation to your society regarding OA is not to fear it but to embrace it and exploit it. A good OA strategy would have the following elements:

  • Coopt OA in your subject area. Once a society accepts that OA will enter its market, that society should develop the strongest OA service and squeeze out rivals and upstarts.
  • Consider a cascading strategy. We have seen this work with wonderful effect. A prestigious toll-access journal typically rejects a number of articles that are scientifically sound, but may be narrow in scope or simply not paradigm-breaking. These articles are candidates to be directed to a Gold OA service bearing a variant of the society brand. Note that this is high-margin business because the articles have already been reviewed by the toll-access parent publication.
  • Work with partners on new OA options. We are impressed with the success many established publishers have had. In some instances you may want to create your own OA service, but there may be cases where you should consider piggybacking on an established OA program.
  • While this may sound paradoxical, your society should simultaneously explore the creation of new toll-access journals and other publications. The business aim is not merely to publish the best literature in a field but to dominate the category altogether.
  • Build your brand. OA publication can help to support this.

We base these recommendation on our view that in the coming years, OA is likely to be additive to, not subtractive from, traditional journals.


Octarine Consulting, LLC

Joseph Esposito

Joseph Esposito

Joe Esposito is a management consultant for the publishing and digital services industries. Joe focuses on organizational strategy and new business development. He is active in both the for-profit and not-for-profit areas.

View All Posts by Joseph Esposito


57 Thoughts on "What We Tell Society Publishers About Open Access"

Hi Jan,

The group Joe is addressing here is looking to publish their own journals, not to find existing OA journals in which to publish.

Terrific! A super and concise description of the state of play. Thanks.

Just curious why you didn’t include embargoed open access as an option? Pay for newest articles, the older scholarship is all available at no charge

But if this bad practice is mandated for a significant fraction of a Society’s journal articles, say by the US Public Access Program, what should the Society do? This may be a common case, not to be ignored. Should they seek out non-US funded articles, or flip to gold, or tough it out, or what?

As laid out in the White House OSTP memo, the US public access program doesn’t mandate embargoes. It says that “each agency plan shall use a twelve-month embargo period as a guideline… however, an agency may tailor its plan as necessary to address the objectives articulated in this memorandum, as well as the challenges and public interests that are unique to each field and mission combination.”

Isn’t that just Green OA? Subscription articles become freely available after an embargo period.

I believe it is generally called delayed OA and is fairly common. Laakso and Björk identified approximately 500 journals that published around 110,000 articles in 2011. Many were high impact and about half were society journals.

Laakso, M. and Björk, B.-C. (2013), Delayed open access: An overlooked high-impact category of openly available scientific literature. J Am Soc Inf Sci Tec, 64: 1323–1329. doi:10.1002/asi.22856


Again, embargoed OA is not a federal policy. The embargo guideline is only a guideline, as pointed out above.

I think it is a bit of a stretch to say that the OSTP memo is not Federal policy, especially since the “P” stands for policy. It allows for flexibility (that no agency has opted for) and every agency has complied. It is viewed as an order.

I didn’t say that the public-access requirement isn’t federal policy. I pointed out that, as written, this policy does not require an embargo. The language from the policy that makes this clear is cited above.

You are correct about the memo language, but every Federal agency has adopted the 12 month embargo as their policy, so that is the present policy. That these agency policies can change does not change this fact. Moreover, HHS, Edu and Labor are required by law to use a 12 month embargo, so there is no option and they provide over half of the basic research funding. So the 12 month embargo is present Federal policy. As with any Federal policy, that can change and it may well under the new Administration and Congress.

“a limited number of journals”–Joe, you know (or should know) better than that. 6,749 journals active in 2015 (more than twice as many as there are gold journals with APCs), publishing 250,954 articles. The bulk of publishing in Latin America. I suppose anything short of infinity is “limited,” but still…

How does the article count compare between non-APC Gold/Platinum journals and APC-funded Gold journals? (Comparing journal title counts is good and useful, but the rise of the megajournals has rendered that comparison less meaningful than it used to be in isolation.)

I haven’t seen a comprehensive survey of all article output–and let’s add to that: we need such a survey broken down by field. My hypothesis is that more complete information would not change the numerator very much (the total market size), but the denominator (number of articles) could well be much larger. That in turn would bring down the average revenue received per article. See Walt Crawford’s comment on this post, which notes a very large Platinum segment (but smaller than I would have thought with Walt’s international perspective). Note that the post was not about OA in general. It was about guidance for professional societies, whose finances are intertwined with their publishing programs. For many societies (probably most, but I have not done a survey) OA is viewed negatively and even with paranoiac alarm. My point to them is that they should coopt OA. This is an unabashedly partisan post.

To be meaningful, OA article counts will also need to include those published in hybrid toll-access journals. Account for those, and the market share of APC-funded open scholarship rises further.

I’ll reply to both responses: Joe, you haven’t seen such a survey because you haven’t (apparently) read the 100% free CC-BY “Gold Open Access Journals 2011-2015” or the subject offshoot, “Gold Open Access Journals 2011-2015: A Subject Approach”–easiest way to get to both is from the project page, http://waltcrawford.name/goaj.html — More than 8,000 downloads to date of the overall book (there’s a paperback version, but at $6 it does carry a production charge.) Rick: for 2015, the APC-charging article count was 315,968. [There’s also a newer survey of “gray OA”–gold OA journals that were *not* in DOAJ on 12/31/15–and that’s published as Cites & Insights 17:1, http://cical.info/civ17i1.pdf–and, as I suspected the Shen/Bjork “420,000 2014 articles in predatory journals” was wildly off the mark; AFAICT, the appropriate number is 113,996 articles, with 29,947 of those in journals where Beall actually made some sort of case rather than “trust me.”]

Thanks, Walt. So there’s about twice as many non-APC-charging OA journals as there are APC-charging ones, but the APC-charging ones produce 26% more scholarship.

One more quick question: does that figure (315,968) include APC-supported articles that were published in hybrid journals?

No. It’s a study of DOAJ-listed journals. I don’t know of any way to find the number of “hybrid” articles. And yes, I’ve been noting for a while that most gold OA journals don’t charge, but most gold OA articles appear in APC-charging journals.

This speaks to Joe’s mention above of this model failing to scale. It’s reasonable to run a non-charging journal with volunteer labor when you’re only dealing with a handful of submissions at a time. Using the numbers above, 250,954 articles in 6,749 journals means around 37 articles per journal per year, three a month. But if you’re too successful and start having to deal with hundreds of submissions per month, you end up needing full time employees to cover things, which is harder to do with volunteers.

I don’t disagree, although non-charging doesn’t imply no-revenue. There are lots of places where very small journals make sense, esp. in the humanities, but yeah, trying to run PLOS ONE without full-time employees would be a non-starter.

To be fair, Walt, your rounded number for 2014 is 255,000 articles based on a newer version of Beall’s two lists than S&B used. That is still a lot of science. I agree that “gray OA” is a much better term than “predatory” because the content is real. This low cost business model needs further discussion.

David: I think I ws being fair: I used versions of Beall’s lists that were as close as possible to what Shen/Bjork used, arriving at 113,996 for 2014. Even 255,000 is a LOT fewer articles than their 420,000 projection–and nowhere in their article do they suggest that they’re including publishers and journals *not* yet in Beall’s lists. Yes, even 114,000 articles is a lot of stuff (not all science by any means), and yes, it’s an issue, but the repeated drumbeat of “almost as many predatory as in DOAJ ” is simply not true. Not even close.

My point is that a great deal of content is being ignored because it is supposedly predatory, when it is in fact an emerging business model. In that context the difference between 420,000 articles and 255,000 articles is negligible. I come from a gray lit background, specifically Federal research reports, which are important but not peer reviewed. See my https://scholarlykitchen.sspnet.org/2011/09/21/taxpayer-oa-is-already-here-in-principle-in-reports/. These gray OA journals look like the real deal to me.

Whether the articles represent legitimate science or not, if the journal doesn’t perform any sort of peer review, then we have no way of knowing, and at best, these could be seen on the same level as a preprint. I agree with you that a substantial number of what is published in “predatory” journals is done so by authors that may have done legitimate research, but who are taking the path of least resistance in publication, because their employers don’t perform proper oversight on career advancement. If all it takes to get a raise/promotion is a paper published in a journal, then why not pay a small amount for an immediate publication with no review rather than going through a slow and rigorous process?

The value of the content of those papers is going to vary widely, but we have no way of knowing (and as far as I know, there have been no studies on the “validity” of the research published in these sorts of journals). But given that the authors appear to be pulling a fast one, taking a shady route to publication rather than going through proper oversight by legitimate parties, that would seem a strike against their credibility.

And no, I do not consider fraud to be an “emerging business model”.

I’m not comfortable with “emerging business model”–since the model is “OA that’s not in DOAJ,” rather a nebulous model–but I’m not willing to assert that none of these journals have peer review either. That case has not been made; for that matter, Beall usually doesn’t make any case *at all* for adding things to his lists.

Shen & Bjork estimate that the average APC is about $200 and if you back out the handful of big publishers that drops to $100. This is the emerging business model — very low cost journals for developing economies that are now investing in research. Publishing hundreds of thousands of articles a year, they need to be taken seriously, not dismissed as somehow predatory. There is no evidence that these articles are somehow all invalid. The claim strikes me as elitist.

Shady? A $100 APC journal cannot look or operate like a $2000 or $3000 APC journal. You seem to be invoking a rich country standard, as though science required it, but it does not. Try this. Cut your budget by 95% and see what you have left. Journals do not create research, researchers do that. Extensive peer review is nice, but not necessary.

Sorry, no. You are wrong here. There is a basic minimum of oversight necessary to instill trust. If the journal promises that oversight and fails to provide it, it is committing fraud. Yes, there are many legitimate local and low cost journals that exist in markets throughout the world. But not every low cost journal is equally legitimate. One cannot make a sweeping generalization that every article in every journal has been adequately reviewed so as to be trustworthy when there is clear evidence that this is not the case.

Of course I am making no such sweeping generalization, but you seem to be, based on no evidence. Every article that I have looked at in these low cost journals on Beall’s list has been legitimate and I think the great majority probably are. I cannot imagine researchers cranking out hundreds of thousands of bogus articles a year. You are demeaning them.

As for trust, I have no idea what you mean. Are you claiming that the only trustworthy scientific content is that found in expensive peer reviewed journals? I think not. Scientific communication is a vast realm.

Note too that what a journal promises on its website is irrelevant as far as the legitimacy of its articles are concerned. Nor do we have any data on how prevalent this advertising practice is among the thousands of journals involved. My impression is that many of these so-called “predatory” journals actually do some peer review, but it cannot be much for the small money involved. As has been pointed out frequently here, extensive peer review is expensive.

There are several overlapping discussions going on here. I don’t consider most or all of these articles shady or invalid; I’d bet most of the journals do some level of peer review and most of the articles are good, but narrow, science. But if the journals want to be taken seriously, they need to step up and qualify for DOAJ–which is free, certainly not requiring high APCs. I also am bemused that, given that Shen/Bjork’s article estimates are wildly off, you assume their APC estimates are correct and wholly ignore my 100%-actual-sample numbers: an average of $299 per APC-based article in 2014. You know, all those numbers are freely available…

Of course I am making no such sweeping generalization, but you seem to be, based on no evidence. Every article that I have looked at in these low cost journals on Beall’s list has been legitimate and I think the great majority probably are. I cannot imagine researchers cranking out hundreds of thousands of bogus articles a year.

These seem contradictory statements. I make no sweeping generalizations, but my anecdotal evidence allows me to make the following sweeping generalizations. What percentage of articles/journals have you sampled? What is your randomization methodology? Which experts in which fields have you conferred with regarding the legitimacy of the articles? What plagiarism detection software have you run on the samples to determine their originality?

There is clear empirical evidence that at least some of these publishers do no peer review, given the Bohannon sting, and the many, many examples of people submitting gibberish papers and having them accepted. This tells me that the phenomenon exists, although not to what extent, but it then puts the burden of proof on the journal itself, and as Walt mentions in another comment, there are ways to to attain certification that everything is above board.

“Trust” must be earned. I read an article in a peer-reviewed journal that has earned my trust for editorial oversight in a different manner than I read a non-reviewed preprint.

You also seem to assume that localized, low-budget journals are a new phenomenon. They are not, and have existed for many decades. This is not a new discovery, but a long-running part of the publishing landscape.

David, you have stated that:

> I cannot imagine researchers cranking out hundreds of thousands of bogus articles a year. You are demeaning them.

> As for trust, I have no idea what you mean. Are you claiming that the only trustworthy scientific content is that found in expensive peer reviewed journals? I think not. Scientific communication is a vast realm.

If you take the replicability crisis seriously, then it is clear that even in the most expensive, reputed peer reviewed journal you will find that only a portion can be trusted, and that researchers really are cranking out hundreds of thousands of bogus articles a year (usually without any malicious intent). The implication is that we need far, far more rigour if we want to be able to rely on scientific research at all, not less. The business model of Gray OA may be viable, it may even be the best option for excellent and morally upright researchers in poorer countries to publish their results, but ultimately it doesn’t seem to address the urgent need for quality research, rather than just more research of any kind.

I do not take the so-called replicability crisis seriously, but I agree that publishing in a high priced journal does not guarantee trustworthiness, whatever that means. I like preprint servers for that matter.

You may wish to mention SCOAP3 as a rather significant player in the arena of Gold OA. It’s just a thought.

Keep in mind that it IS pretty much half of particle physics scholarly output at this point.

Yup — and it represents absolutely nothing in any other discipline.

Again, I’m not trying to diminish what SCOAP3 is doing — I just don’t think it makes sense to exaggerate its impact, either.

How is stating facts an exaggeration? I made a suggestion (and I was polite at least) and then I stated an accurate figure for its coverage in particle physics. I suppose if the model continues to catch on, then it will turn more heads. It’s a different funding model that leaves the authors alone to conduct their research. That makes it very significant in my view.

SCOAP3 is an interesting phenomenon, but it bumps into a hard truth, that it doesn’t provide a sufficient surplus (or growth potential) to participating society publishers. When I said that some societies are fearful of OA publishing, SCOAP3 would be an example of an outcome that they would not want to see.

Five of the eight titles in SCOAP3 are society publications. So, at least some are not fearful. However, I do recognize that fear of OA is a big problem.

Fear of OA is not a problem. The problem is not having a sufficiently commercial imagination to monetize OA.

Okay, fear is just the manifestation of “not having a sufficiently commercial imagination.” 🙂

I for one can’t wait for the Star Trek economy …

I fear we are heading for something more akin to the Game of Thrones economy.

How is stating facts an exaggeration? I made a suggestion (and I was polite at least) and then I stated an accurate figure for its coverage in particle physics.

I’m not disputing the figure you offered. I’m (politely, I hope) suggesting, for reasons I explained, that the characterization of SCOAP3 as a “significant player in the arena of Gold OA” is an exaggeration. I’m not suggesting it’s insignificant to people working in particle physics; I’m suggesting it’s not very significant to the larger ecosystem of Gold OA.

FWIW, I agree with you that if the model catches on, it may turn out to be more significant to the Gold OA ecosystem. But the fact that SCOAP3’s second phase shows the falling away of two journals and the addition of no new journals doesn’t give a lot of reason to believe that the model is catching on.

It would be important to note that both of the journals for which an agreement to continue was not reached represented less than 5% of the articles in SCOAP3 for 2014 through 2016. Also note that neither was purely a particle physics journal. New Journal of Physics was already open access and is reverting it’s small contribution to SCOAP3 (25 articles in 3 years) to its original Article Processing Charge model (sorry authors). Journal of Cosmology and Astroparticle Physics (JCAP) was more significant – roughly 30% of its articles were published via SCOAP3 throughout 2014-2016. JCAP still offers the author-unfriendly route to OA. Both are IOP journals.

Stay tuned for other news.

Nowadays all scientific papers are covered by e.g. Web of Science or Scopus, and all these contain: title, abstract and at least an e-address of a corresponding author.
Every corresponding author gets a free pdf of her/his paper.
So, any person – scientist, journalist, layman – can find any article based on key words for instance, and see the abstract and the corresponding author, from whom a pdf can be obtained.
This would suggest that we do not need at all any OA journal or artcle. We can just ask the corresponding author
Or do I overlook something here?

Access (getting a free copy) is just one part of OA. The other key concept is a lack of copyright restrictions on reuse of the material. Some feel this is important, and it is not covered by one’s ability to get a free copy from the author.

There’s also a pretty significant gulf between having immediate access to an article because it’s been made OA, and having the option of trying to contact the article’s author to request a copy. Such requests will in many cases be ignored, or the author may not be contactable, or the author may not respond for weeks or months. What the OA movement is trying to achieve is something much more useful and efficient than that process.

I do think you are overlooking a number of things. First, and my apologies that this will be blunt, people die. But, they also retire, leave research as a profession, etc. They also do not actually answer their email. Also, not all scientific papers are covered by Web of Science or Scopus. But, even if they were, those databases are not available to any person to search – they are available to individuals who are affiliated with institutions that pay a fair bit (some might say staggering) amount of money to provide access.

I am a little annoyed by the seemingly self-serving comment … “we discovered that Elsevier received a bit over $5,000 per article. The highest figures are to be found with not-for-profit professional societies. We worked with one not-for-profit society client that receives $16,000 per article.”

I would have been more helpful to compare Elsevier with a reasonable ‘not-for-profit’ publisher like the APS or AMS.

Given that the author does not work for Elsevier, how is this “self-serving”?

Regardless, it’s worth noting that this figure is Gross, not Net, so I’m not sure if this figure tells us much of anything. Given Elsevier’s economies of scale, their Net may end up higher than that of an independent publisher society that earns a much higher Gross.

I really can’t agree with you. The numbers don’t support your comment. If $5,000 is the average, surely that is the measure against which to compare Elsevier or anyone else. Note as well that some society publishers have higher operating margins than Elsevier. As for the comment being self-serving, I would take that as a compliment, but I am darned if I can figure out how I am serving myself. Should I expect a payment in cash from Elsevier in a brown paper bag? Happy to provide my address.

Another note, since the post relates to society publishers: Overall–and very much internationally–publishers that appeared to be societies (and a few government publishers) accounted for 1,086 journals, only 17% charging APCs, publishing 59,372 articles in 2015–39% of those in APC-charging journals. [Those counts don’t include all the society-*sponsored* journals published by other publishers, such as many of De Gruyter Open’s mostly-non-APC journals.]

Comments are closed.