A hoodie with the University of California, Lo...
Image via Wikipedia

When site licensing to academic institutions became the dominant form of digital journal sales, many publishers knew they were making a severe strategic compromise. But there was too much money to make in fairly short order. Plus, librarians were vociferous in demanding a future shaped by site licensing.

For publishers, site licensing became a tactic based on revenue and library relationships. For librarians, site licensing became a tactical bridge into the future, but a bridge built only with one plank.

Both librarians and publishers were not acting strategically when they settled on the site license, and there have been substantial implications to the resulting tactical choices.

For publishers, site licensing led to a loss of contact with actual customers — readers — which is starting to make marketing departments and publishers squirm all the more in this era of social media, device-based publishing, and other individual consumption choices. Not having a strong trendline of customer habits and usage patterns creates blindspots in customer knowledge and leaves gaping holes in customer databases. A decade is a long time to lose touch with students, trainees, and young researchers. It also led to extremely discounted per-user prices in publishers’ core markets, often cents on the dollar, despite the perceived steep prices on an aggregate basis.

For libraries, budget reallocation has been the story of the era, with materials budgets shifting to site licensing. There has also been a perceptual price to pay, with faculty seeing the library now as a glorified purchasing department rather than an intellectual partner. Now, publishers are beginning to look beyond the library for revenues owing to flat or declining budgets. Librarians aren’t an appropriate arbiter of a revenue model beyond the librarian-managed site license, and the site license is shrinking.

Recently, a familiar front was opened up in the ongoing “library as purchaser” territory battles, with Nature Publishing Group attempting to increase prices by 400% to University of California libraries. In an open letter, UCLA University Librarian Gary E. Strong notes that some at a Vice Dean at UCSF is ready to boycott Nature journals, urging faculty to:

  • Decline to peer review manuscripts for journals from the Nature Publishing Group.
  • Resign from Nature Publishing Group editorial and advisory boards.
  • Cease to submit papers to the Nature Publishing Group.
  • Refrain from advertising any open or new UC positions in Nature Publishing Group journals.
  • Talk widely about Nature Publishing Group pricing tactics and business strategies with colleagues outside UC,
    and encourage sympathy actions such as those listed above.

Nature’s response has been equally predictable, with phrases we’ve all read before elsewhere, like:

NPG adds huge amounts of value to the very best quality original research, and this situation was simply not sustainable. It is our belief NPG titles represent excellent value for money, whether measured by cost per download, or perhaps more accurately, cost per local citation.

The “rinse and repeat” approach is clear in this exchange and its echoes. From CDL, there are the usual arguments about how many papers are submitted, reviewed, and edited by University of California faculty; claims that Nature is “at war with its own customers”; and the whole “aren’t we supposed to be on the same team” theme. From Nature’s side, there’s the economic argument, the value-add argument, and the per-usage argument, including a nice little trap Nature has made for itself — i.e., since CDL is paying less than others, effectively being subsidized, should CDL’s prices go up, the logical follow-through would be that prices at other institutions would be reduced. I doubt this would happen.

Aside from the fact that libraries aren’t the customers of Nature’s information (students and faculty are — the library is just the purchasing agent), it’s the usual rollercoaster ride of emotional appeals, economic arguments, and shaming, all in an attempt to maintain the library-as-purchaser status quo.

Why that status quo hold sway is worth considering. For librarians, there is the obvious safety that managing these important budgetary items provides. In fact, given the changes in library priorities, it could be argued that these budgets and the digital presences emanating from them are the last bastions of librarianship on any significant scale for academic centers. If these budgets shifted from institutional purchasing departments and digital access management to a technology department, for instance, librarians would find themselves severely marginalized. It’s a situation created by dependence on site licensing as a model, and sustained by the same — but that doesn’t make it right or sustainable.

From a broader university administration point of view, it’s a great situation. Librarians will fight like dogs to keep these budgets as small and predictable as possible, while tuition, fees, and overall revenues can climb independently for academic institutions. In short, there is no incentive to change where or how site licenses are budgeted. What universities have in place now is a very effective way to control costs in this area.

And the costs are certainly controlled. Even with an alleged 400% increase in prices, Nature is projecting that CDL will pay $0.56 per download. It would be interesting to compare the site licensing economics of Nature’s titles to those of, say, PLoS ONE. From what I can see from a sampling of 3-month-old papers, authors are paying between $1.00 and $2.00 per download at the current rates PLoS ONE charges. For a top-tier publication to settle for what appears to be a fraction of what an author-pays publisher with admittedly lighter peer-review standards generates shows how well costs are controlled in a site licensing model.

Academics playing into the boycott give one the impression of possessing a forced naivete, as Keith Yamamoto said to the Chronicle of Higher Education:

Although researchers still have “a very strong tie to traditional journals” like Nature, he said, scientific publishing has evolved in the seven years since the Elsevier boycott. “In many ways it doesn’t matter where the work’s published, because scientists will be able to find it,” Mr. Yamamoto said.

As if making an article available to Google is what motivates researchers. An acceptance letter from a Nature journal would be 99% likely to break this boycott for any individual scientist.

The arguments about contribution of peer-review, authorship, editorial expertise, and so forth don’t impress me. Academics contribute these things for all sorts of selfish, career-driven reasons, and the prestige of any one journal often exceeds the prestige of their own institutions, at least in the land of the CV. Peer-review is a great way for an active academic to keep an eye on research in her or his field. There is little altruism propelling this behavior, so to portray it as unselfish contribution is just emotional balderdash.

Shaming for-profit publishers is also not an especially compelling argument to me. Most universities of any size have large stock market and mutual fund portfolios. They depend on dividends from for-profit businesses for a significant proportion of their revenues and important cash flows. Just reflect on how many problems universities encountered when the real estate bubble popped. Layoffs and budget cuts were commonplace. Their money doesn’t waft down from heaven on angel kisses. They, and the parents of their students, and their grant-writing bodies — all were hurt by the rapid decline of the equity markets. (By contrast, non-profit publishers have no requirement to return money to investors, and universities can’t benefit financially from the surpluses they generate.) Librarians should think about this each time a paycheck is deposited. At some level, sniping at a for-profit publisher may be nipping the hand that feeds you. And I’m sure publishers are among the most laudatory members of any institution’s investment pool, when you take a close look at it.

So, here we are, in 2010 — we have librarians holding on to their last major source of power in the digital age, the site licensing budget, which is shrinking or flat; university administrators who likely sense they’ve stumbled upon a great budget watchdog situation for information expenditures; faculty who can be swayed into action but who mostly are still incentivized to publish in high-profile journals while at the same time enjoying the benefits of ubiquitous online access to same; university budgets that depend on equity markets (for-profit companies); and publishers of all kinds who are stymied by this strange brew of tactics, compromises, and short-term thinking while facing increasing expenses and pressures to invest in new markets.

As Eric Hellman sees it, the NPG/CDL battle is the result of a “communication failure.”

I think it’s deeper than that, and not amenable to the band-aid of “improved communication.” I think this is the inevitable upshot of a fundamentally flawed situation.

There are some important points buried in the rhetoric of those complaining about these higher prices NPG is attempting to reach, but I think you need to look at them slightly differently:

  • “Aren’t we on the same team” — The fatal flaw of site licensing is that it pits traditional allies against each other, effectively putting libraries and publishers on opposing teams. This flaw points out how non-strategic site licensing has been for both libraries and publishers.
  • “Things have changed since the Elsevier boycott” — Things certainly have changed, meaning that the desktop=Internet correlation is eroding quickly, while mobile=Internet is emerging. At the same time, we understand more completely how entrenched researchers’ citation motivations are, even in the digital age — and, interestingly, they’re more likely to shape the behavior of young scientists.

Where will it all end? It seems to me that the shift away from the desktop is an opportunity for everyone involved to rethink this witch’s stew of tactical compromises. As a recent article in the Atlantic puts it:

The era of the Web browser’s dominance is coming to a close. And the Internet’s founding ideology — that information wants to be free, and that attempts to constrain it are not only hopeless but immoral — suddenly seems naive and stale in the new age of apps, smart phones, and pricing plans.

Wireless, 3G, 4G, and mobile will be the Interwebz of the near future, and already we sense the answers to new questions:

  • What is wrong about individuals paying for information they want?
  • Do we need to bundle all devices into site licenses and perpetuate the problem, or is there another way?
  • Can publishers move to re-engage with readers again?
  • Can librarians become guides to information and intellectual partners again in the device-driven age?
  • Should publishers just ride out this site licensing motif without damaging their reputation while working hard to move commerce to the emerging ecosystem?

Is the site licensing era something everyone involved would be better off leaving behind over the next decade? If so, how do librarians and publishers execute a strategy that no longer pits them against each other?

Enhanced by Zemanta
Kent Anderson

Kent Anderson

Kent Anderson is the CEO of RedLink and RedLink Network, a past-President of SSP, and the founder of the Scholarly Kitchen. He has worked as Publisher at AAAS/Science, CEO/Publisher of JBJS, Inc., a publishing executive at the Massachusetts Medical Society, Publishing Director of the New England Journal of Medicine, and Director of Medical Journals at the American Academy of Pediatrics. Opinions on social media or blogs are his own.


26 Thoughts on "The Latest “Library as Purchaser” Crisis: Are We Fighting the Wrong Battle?"

A key strategic blunder of the recording industry was thinking that the record store was their customer. Likewise, when you talk to trade book publishers today, they think that the bookstore is their customer. You have correctly identified the analogous blunder for STM and scholarly publishers: falling into the trap of thinking the library is their customer (meaning who they are producing information for, not who writes the check at the present time). This problem is less pronounced with society publishers who still focus on membership and with membership the bulk of their subscriptions (at least in terms of volume, revenue may be a different story). Fortunately, the overlap with author and reader communities means that most STM and scholarly publishers have close connections with their readers even if they are deriving most of their revenues from site licenses (clinical titles aside).

I don’t know that I agree that pursuing a site licensing strategy was a strategic error, however. Directing sales resources to a few thousand libraries is far more manageable than directing those same resources to tens (or, in some cases, hundreds) of thousands of individuals. In that mid-to-late 1990s, I think this strategy made both strategic and tactical sense. (This is only true, however, if one does not forget that one’s customer is the reader, not the library, and one directs marketing resources accordingly and maintains strong author relationships).

I agree that we have reached the limits of this strategy, however, but not for reasons that were, from the vantage of the 1990s when all the site licensing started, obvious or even inevitable. While one can point to incidences of pricing excess by publishers, I think a far greater problem is the failure of librarians to make a case for increases in information budgets — and the failure of publishers to help support librarians in making such a case. The output of information has increased dramatically in the last 15 years. Moreover, the investment in digital publishing has lead to that information being both more accessible and more useful than ever before. And yet librarians like to use metrics like inflation and the CPI to show that information prices have increased out of proportion to, say, a static unit of cotton (as opposed to using a more relevant indicator like the cost of tuition at universities or the cost per downloaded article). Phil Davis has long since debunked this entire line of thinking here in the Scholarly Kitchen (see: http://scholarlykitchen.sspnet.org/2009/06/01/consumer-price-index-oa/ ) and yet it persists.

In addition to the questions you pose above, I would add the following: Can publishers and librarians work together to make a better case to institutions about the value of STM and scholarly publications (when measured against cost per article, faculty’s time spent on research/writing, impact of scientific research, and other more meaningful metrics) and need for more funds to be directed to information budgets?

“Who is the customer?” is an important question. I believe that the people who send you checks are your equilibrium customers, and that when that’s not true, you’re in a fundamentally unstable situation.

I would argue that readers are in no sense the customers for scholarly journals. The true customers, in the present era, are the authors. All decisions made by journals, no matter the business model, are geared towards authors, not readers, not libraries.

That why CDL struck a nerve, and why NPG reacted so strongly- both are appealing to the true customers.

I disagree with your argument that “readers are in no sense the customers for scholarly journals.” Authors become readers after they’re done writing a work. “Author” and “reader” can be roles of the same person, or they can represent different people. For small-circulation journals, the overlap can be significant. For large-circulation journals, the overlap can be slight. Journals vary in many ways — by field, within fields. One of the site licensing model’s flaws is that it attempts to treat them all the same. Publishers have carved out some pricing concessions, but actually, there’s probably a lot more pricing flexibility needed to reflect real value.

Saying that “all decisions made by journals . . . are geared toward authors” is simplistic and clearly wrong. If you’ve never worked at a journal, you may not know the number of reader surveys, focus groups, and interviews publishers do. We talk with authors and reviewers, too, but readers are the primary focus. Site licenses have made readers harder to identify and understand.

OK, in the morning light, I agree my late-night comment was overbroad to the point of silliness. Nonetheless, the role of the author as customer is attested to by the amount of marketing being aimed at the author base. For example, consider the marketing of special issues. My unscientific survey suggests that special issues tend to be marketed more in advance, as calls for papers, than after publication.

There is certainly a market for journals that are geared primarily to service the needs of authors with little thought to readers. These are the vanity press journals that see little (if any) readership and little (or no) citations with the exception of self-citations.

Nature, however, does not fit into this market.

What about the second-tier journals that acquire papers by publishing conference proceedings. Are these papers acquired for the benefit of the reader?

OK, if site licensing was strategic, please state the strategy for me.

As one wise person once said, learning to carve your skis better is a tactic. Choosing to ski in Utah is a strategy. Everyone skis better in Utah. Are you telling me that site licensing was like publishers and librarians mutually agreeing to ski Alta? Or rather, was it finding harsh terrain and carving the turns harder?

Strategy is formulated in response to conditions. Depending on the time of year and El Nino pattern, it may be that one wants to ski A-Basin or Taos instead of Alta. Additionally, what skis do you bring? Do you bring an all mountain ski or bring one’s fat boards and hope for big snow?

Site licensing was a strategic response to the decades-long decline in individual subscriptions (a trend that Don King has documented goes back long before the advent of the Web). Moreover, while researchers may subscribe to a few core titles individually, they need access to a much broader set of content for research purposes. They are not going to subscribe to tens, never minds thousands, of titles. Nor are they likely to take out an individual “package” subscription to dozens of publisher sites (e.g. an individual subscription to Science Direct, Ovid, Wiley Interscience, etc. if such a thing existed). The site license is a mechanism to provide access to the greatest possible number of titles for the whole faculty of an institution. And it has worked. More people have more access (and useful access, at the tips of their fingers) to more titles than ever before. And publishers have by-and-large done well by the site license. I think it has been win-win thus far.

But that is not to say that a given strategy works indefinitely. Conditions change. The El Nino patterns change. Sometimes one has to shift from Utah to Colorado or New Mexico. It might even be time to trade in the old long boards for a pair of reverse camber parabolics. The site license strategy may have run its course but that does not mean it was a bad strategy given the conditions for the last 15 years.

Agree with Michael. Site license models work well (for ex, in software) where a large customer is extremely profitable compared to a small customer b/c selling and servicing a big co is more scalable, such that the vendor/publisher can give away margin. Years ago, that made sense to publishers. Particularly as big deals evolved such that you needed sales reps to negotiate immensely complex contracts.

Today where the “service” cost of delivering e-journals is negligible on a marginal basis (no longer cheaper to serve a big customer than a small) and where in theory any customer/librarian can buy from you if the publisher had an online order entry system for subscriptions (i.e. do you need so many sales reps?), there is less of a compelling economic argument for having site licenses. As for the argument of getting to know your customer again, my understanding is that publishers never knew their customer – that before there were site licenses, there were reseller ‘agents’ who served as intermediaries. If so, really this is the potential dawn of a new era for publishers to FINALLY know the customer.

Of course that assumes that publishers have the resources and corporate DNA to understand this new market segment, and more importantly, does the customer want to know the publisher as opposed to say Google?

Maybe I’m dense, but does the shift to greater internet use with mobile platforms change that much about the consumption of research? A download is a download, and a microbiologist’s research interests aren’t going to change just because he can access journals from his iPhone. In any case, it’s not yet obvious to me what librarians would gain from splitting mobile access from our site licenses. Either we pay for access or our users do. If we pay, then incorporating moble access into some form of site license seems inevitable. At least, that’s what it will amount to, even if we aren’t calling it that. If we get our users to pay for their own access, then it’s not clear what we’re for. Free tech support to students with access problems? Cataloging and local search maintenance?

I would have said that we’re witnessing the ascent of the web browser, not its end. Mobile devices display web pages with browser software just as surely as laptops and desktops do. If app use data were handy, I have the sneaking suspicion that browsers would be used far more often than any downloaded apps.

My eyebrow was permanently raised as I read that Atlantic article. Old media companies were among the first to offer–“devalue,” as the author put it–their wares online for free? I’d love to know what he has in mind. Apple has always fought the idea of open? Their stance has been much more complex than that, as Steve Jobs’ famous open letter against DRM shows. And a great many of us under-30s are quite content to pay for our entertainment. We can quibble about the amount we prefer to pay, but the notion of paying for entertainment isn’t shocking or foreign to us.

As for information wanting to be free, that was the guiding philosophy of some of the builders of the internet. It has never been ubiquitous. Does anyone think this was the guiding philosophy of DARPA? And people still remember AOL’s walled garden on the internet, don’t they? It was AOL, not Apple, that was the first to make big plans for a curated internet. Microsoft and Real Networks certainly had other plans for the internet. And some of those who agree with the sentiment have idiosyncratic definitions of free; just ask the Free Software Foundation what they mean by free. I wouldn’t call it malicious, but I suspect it’s a vastly different meaning than what most ordinarily mean when they use the term.

This is an important and timely post that raises important issues for me and my fellow librarians about the ongoing viability of the site license model. That relatively brief era seems to be running its course. I made a similar point, less eloquently, here: http://mbanks.typepad.com/my_weblog/2010/06/university-of-california-libraries-vs-nature-publishing-group.html

I work at the University of California San Francisco, which has been the most outspoken UC campus about the possibility of a Nature boycott. While the “publisher is evil” argument is usually an annoying strawman, here it seems to me that Nature has been amazingly obtuse. The state is broke and library budgets are reeling. This wasn’t the year to attempt to rebalance prices, especially by resorting to the notion of a “list price” that nobody pays in the first place.

With regard to author motivations in the scholarly space–sure, let’s grant that it is all selfish. The California Digital Library letter doesn’t say that authors work for altruistic reasons, just that they contribute a lot of uncompensated time to the work of sustaining the peer review system. Maybe the incentives will always run toward the established, and new models will never prevail. Even in this universe, there should be some way to factor in a university’s intellectual contribution in the prices paid for journals. Without the labor of UC scholars, and scholars elsewhere, Nature would not exist.

There is another way, that of “library as publisher.” Glorified purchasing agent sounds horrible, but publisher could have promise. We already do this at the University of California, with the eScholarship platform: http://escholarship.org/. Sure, eScholarship is David vs. Goliath vis-a-vis Nature. Then again, David won.

I’m not really buying the argument about benefits to a University from the financial success of a for-profit publisher in which the University owns stock. The Universities as a whole make up a much larger fraction of the customer base of the publishers than they do of the stockholders. Instead of there being lots of other customers buying the publishers’ products and therefore indirectly supporting the Universities, there’s a sort of inefficient loop in which the other stockholders benefit from the purchases by the Universities. And even if the University holdings of publisher stocks were significant, wouldn’t this in effect be a transfer of wealth from the poorer Universities–those that don’t have large investments in the publishers–to those richer Universities that do have such investments. And for the case of the University of California, the per-campus endowment is rather modest and would rank around 100. My reading of their financial position tells me that endowment income is only a few percent of revenue. But all these things are public, quantifiable information and the post would be a lot stronger if the relevant numbers were collected and crunched.

On another point, I don’t see how the “loss of contact” is relevant at all to academic journals. It’s not like academic journals–even the for-profit Elsevier ones–had advertising or other content that’s now being missed in the era of online downloads. Journals, as a rule, are rather Spartan containers for articles with little else. The primary interface for most researchers to journal content has been, for quite some time, citations from other articles, databases like INSPEC or Science Citation Index and a handful of tables of contents from the top journals. Nobody browses, say, Elsevier’s Thin Solid Films to see if there is anything new and exciting this month. Except for journals that come with society membership–like Science–individual subscriptions–like, again, to Elsevier’s Thin Solid Films–are rather rare. Publishers are now getting more information about the way the journals are being used–it’s easier to count how many people subscribe to an email table of contents than to guess how many people make a monthly trip to the library to browse a new issue. And it’s easier to count pdf downloads than it is to guess how often an individual article was photocopied.

Well, of the $2.3 billion in equity assets managed by the one dozen or so equity managers in the University of California system, I’m sure some of this includes profits from publishers. As for the effect of markets on university budgets, note that the University of California system’s endowment fell from $73.4 billion in 2007 to $55.1 billion in 2009. Most of that loss comes from loss in equities, real estate holdings, and other assets. My point was that “profit” can’t be a dirty term in the university setting because so much of a university’s budget depends on it, from investments to tuition and fees. Economics is about a cycle of value.

How did Elsevier know to publish “Thin Solid Films”? Contact with the audience. Losing this contact stunts changes in journals, makes it harder to know when to launch new titles, etc.

These arguments in defense of the publishers are pretty weak. A few examples:

Your stats re. article pricing are misleading. As you can see from this page: http://www.plosone.org/static/journalStatistics.action over the lifetime of an article the price/download less than or about equal to $.50/download, at least for PLoS Bio, which I calculated. And this is the “reduced” price which CDL is paying.

Whether or not authors/editors are altruistic is completely irrelevant to the question of whether they contribute unpaid work to publishers. Universities pay the authors/editors for the work they do for publishers, even if they do turn around & use that work to judge the author/editor.

Whether or not universities depend on for-profit firms for investment income is again, irrelevant. The question is whether universities should pay for the content that their researchers generate; that is, how much value are they getting for the money spent? Would you argue that a university should outsource, say, its dining services to a company that makes large profits when the work could be done by the university for less, because the university has investments?

Finally, whether users should pay for content generally is a different question than whether they should pay for scientific knowledge. Certainly there are costs involved but there are important questions about who bears these costs. Should knowledge generated by state subsidy be locked in a world of for-profit companies, with those unable to pay unable to use it? We are not talking about Harry Potter here. This is a separate issue.

(Regarding apps, wireless, etc., while I have no problem with content owners charging for content, there are some of us who a) think that the web is still the future; and b) will fight against the closed world of apps, etc., a world that locks out the kind of experimental work that created the web in the first place, a proprietary closed world where users are stuck on separate platforms, unable to develop new things, to collaborate, or to do anything besides what the platform owners tell them they can.)

Can you show me the math you used to get the $0.50/download for CDL for PLoS Bio? I can’t see how you could possibly arrive at a per-usage rate for CDL based on the aggregate data for all PLoS usage.

Altruism is actually vital to the question. Numerous studies have shown that work diminishes — especially high-level, autonomous work — when people are paid for it. The altruism of the system may be vital to its sustainability. Universities don’t “pay authors/editors for the work they do for publishers,” if I’m understanding you. They pay academics to teach, do research, and attract grant funding. Work for publishers is tangential.

Many universities outsource catering and cafeteria services, then take a cut of the excess that’s generated by the for-profit companies. But that’s not a good analog. Universities are not equipped to conduct independent, third-party review of their faculty’s output. University politics would overwhelm objective assessment attempts. They aren’t paying for the content — they’re paying for the placement of that content in a more objective and removed merit system, as well as the services publishers provide directly to improve manuscripts. If universities can do it more cheaply and better, they should prove it. The fact is that most university presses struggle to create journals because of parochial pressures. Some are able to elevate, but only by becoming someone else’s third-party arbiter of quality and relevance. And they usually aren’t cheaper by the time they do so.

I don’t understand your point about not paying for scientific knowledge generated by state subsidy. How did that money get to the state? Through taxes. Who paid those taxes? Individuals and companies. Yet universities largely benefit from the grants for scientific research. So, should universities be free? After all, in addition to funds they receive from for-profit entities through their investments, they also receive plenty of state subsidies. If state subsidy means that nobody should pay for what’s subsidized, then universities should be free.

Sorry, I wasn’t very clear. If we look at the oldest articles in PLoS Biology, we can see that they have been viewed (downloaded) 11600 times on average. Since PLoS Bio. charges a $2900 author fee, this works out to 25 cents paid per download (by the author). If we take the total downloads (each year’s n * avg) and divide by the total articles published * $2900, we get about $.43 download that authors are paying. You looked at the downloads over the first 3 months since publication & conclude that an article costs about 1-2 dollars. But we need to look at the lifetime of the article, because the 56 cents that NPG is projecting is for any article.

The point re. altruism is this. Take the example of unpaid interns. They are not doing the work out of altruism, they are doing it to advance their career. But that does not mean that we should not be concerned that companies create a situation in which people are expected to do free internships to advance their career.

Universities pay for researchers & professors to work. Part of this work is writing & editing publications for journals. Journals which generally do not pay. To me, this means universities pay for researchers & professors to do that work.

I’m not saying that universities should found journals, I am suggesting that they need to consider if the money is best spent on for-profit – or, sometimes, not for profit – journals, or on open-access journals. I think the data from PLoS show that it is more cost-effective. Furthermore, PLoS has the advantage of being free to everybody in the world.

My point about “who pays” is you cannot argue that people should pay for scientific knowledge based on the economics of Harry Potter. Whether users should pay for Harry Potter is different from whether they should pay for scientific (and academic, generally) knowledge for a number of reasons. 1) people have a right to that knowledge, as far as I am concerned, especially when the marginal cost of an additional download is basically zero; 3) the state, via grant funding agencies (NSF, NIH, etc.) and via public universities, and via tax exemptions for private universities, subsidized loans for students, etc. has already greatly funded (probably the majority) of the production of that knowledge, so that knowledge ought to be used for the benefit of all. Furthermore it seems that we can do this while spending less money that we currently are (if the PLoS example above holds true.)

One thing Nature doesn’t make clear is whether their $0.56 cost/use projection is over the lifetime of the article or an annual average. Because site licenses tend to be annual, I had assumed that the cost/use projection was based on a per-year average, not a lifetime average. If that’s the case, then comparing the lifetime average for PLoS Biology to annual averages for Nature isn’t a direct comparison. But it’s not clear from Nature’s statement if they’re using an annual figure or a lifetime figure. When I sampled articles, I allowed for expanding the 3-month figure into an annual figure (x4). It still came out to be $1-2 per use. A lot of the PLoS articles aren’t read very much. But because authors pay one-time and institutions pay on a rolling, annual basis, perhaps the best way to compare PLoS to Nature costs is to determine it at the institution level rather than backing out of the author fees. Then, we’d have to know usage at CDL and what they pay PLoS per year, and do that very specific math. I don’t have access to those figures. I would love to, and that was my implicit request in the post.

Interns aren’t a good example because there’s a clear exchange — grades and low pay for work experience. High-level professional aren’t interns. They are seeking what everyone seeks who is intrinsically motivated — autonomy and competence. By serving as peer-reviewers, they have a way of pursuing their love of their field in a way that creates pathways toward increased autonomy and competence. They become smarter, have more insights, feel more in control, and gain intrinsic rewards. It’s the same with blogging here. I don’t get paid to do it, but here I am responding to your comment on a post I didn’t get paid to write. But I like to blog and do it probably more intensely than I would if I were paid. There’s a huge literature on this. Read Dan Pink’s “Drive” or Clay Shirky’s “Cognitive Surplus” to understand the studies showing that unpaid volunteer work done for the love of something and in search of autonomy and competence is often more intense and productive than paid work. In fact, paying can actually curtail the work’s amount and quality.

Universities have established the “publish or perish” culture we currently live in, defining what professors have to do to survive. They could just as easily change that to a “mow the lawn or perish” culture, and professors would be motivated to keep their lawns mowed. Would they be paying for lawn mowing? No, but they’d be incentivizing it. They’d be paying them the same, just for different work. Some universities emphasize teaching over publication, but pay the same.

PLoS has yet to prove that it’s cheaper. In fact, there’s evidence that it’s not. Many OA publishers are for-profit, by the way. Why would they not maximize profit over time, as well?

Your classic litany of online economics needs to be revised. Fixed costs are huge in online publishing, dwarfing those in print publishing. The marginal cost per download has to absorb ongoing SEO, hosting, upgrade, maintenance, email system, file system, backend, staff, administration, and editorial costs. Each download is encumbered with these, and because each file’s fate is uncertain, it’s hard to know how to spread those costs. Print created a system in which the publisher’s job ended largely when the issues went to press. Now, the publisher’s job is constant, and it’s like each is running a 24/7 press. So, the marginal cost per donwload IS NOT ZERO AND DOESN’T EVEN APPROACH IT. In fact, as technology becomes more complex and publishers have to support Kindle, iPad, iPhone, Droid, Facebook, Twitter, and other outputs, their costs are going up. And if downloads are what we can mainly monetize, then the cost/download is ACTUALLY GOING UP! PLoS is also finding that it’s more expensive to publish, I’ll bet. Their article-level metrics, back-end payment systems, archival systems, email hosting, ad systems, and all sorts of other things are adding to their cost basis. They have to reclaim these costs somewhere. Watch as over the years they become more expensive rather than less. It’s already happened, in fact.

It’s great that the government funds some research and gives universities some subsidies and tax breaks. But that doesn’t make information curation, dissemination, or promotion free or cheap. It just makes it easier for universities to run tenure systems and pay faculty and researchers. But given universities reliance on for-profit equities, I’d be interested to compare government contributions to university budgets compared to equity contributions and support from for-profit entities.

We were discussing cost/download.. CDL (I should have said before, I work for CDL, but not in licensing or anything related, and of course do not speak for them) will apparently pay 56 cents/download. I showed that PLoS – counting author fees only – has a lower cost per download. I don’t see any confusion here. The cost per download factors out the annual cost, and it is comparable, in a back-of-the-envelope way, with my PLoS calculations.

Unpaid interns are a good example, because although unpaid interns and academics are in different places in their lives, both perform unpaid labor in order to advance their careers.

Altruism, by the way, does not mean “unpaid”, though it could include things that are unpaid. (As an aside, shouldn’t the logical conclusion of your argument be that if authors have to pay to have their work published, they will work even harder?) You seem to think that altruism means that we work for free so that companies can profit. Well, count me out.

I guess by evidence that Open Access is more expensive than traditional publishing, you mean a study that says countries could save money with Open Access.

The marginal cost does approach zero. I am talking about marginal costs, not fixed costs, in the technical sense.

Govt funds “some” research? In 2006, the federal govt funded 59% of basic research in the US; universities funded 10%, and state & local govts funded 3.5%. Industry funded 17%. [1]

I am not saying that “information curation, dissemination, or promotion” is “free or cheap”. I am not even, necessarily, saying that OA is cheaper – although I think it will prove to be. The point is, we are 90% of the way to a world in which – at least in the US – we can provide free (as in price, and as in freedom) access to basic research for all our citizens, and indeed the whole world. We (as taxpayers and as universities) currently pay for most research, and we pay for the publishing of that information through subscriptions and site licenses. We pay – with perhaps a few exceptions – all the fixed costs. And since the marginal cost approaches zero, we can open this system up to everybody – US citizens & beyond – without additional costs. In a world in which we can do this, we have no right not to do this.

1. http://www.nsf.gov/statistics/nsb0803/start.htm


I still don’t think we’re communicating. CDL pays PLoS a certain amount each year. This amount, divided by the number of downloads attributable to CDL’s “site license” is the cost/download for CDL. Aggregating across all of PLoS is not comparable. Being OA, this is especially true.

You’re missing my point about altruism. Read the literature tying altruism to motivation. Doing things without being paid (volunteering, doing something for the love of it, having a hobby, doing something because you think it matters) leads to more and better work than being paid. By this logic, authors paying to be published might work less hard, actually, come to think of it. Again, read the literature. The sources I mentioned are good places to start.

The evidence I’m referring to about OA being more expensive has been cited on this blog before, and can be found circumstantially in the increased fees PLoS and others have found they need to charge. And you didn’t respond to my point about for-profit OA publishers.

To your research statistics, the article you cite notes that basic research is 18% of all R&D, and that the federal government comes second to industry for the total R&D expenditure. So, I stand by my statement about the government funding “some” research.

Marginal costs are additional costs. Variable costs are typically contrasted with fixed costs. There have been long discussions in other comment threads on this blog explaining the importance and necessity of attributing fixed costs to ongoing operations in order to recoup them. Marginal activity (another download, another page view) has to absorb fixed costs and variable costs as part of it creating a marginal cost.

Rhetorically linking “free” and “freedom” is nice, but property rights and debt are two things that really made individuals free from tyranny (property rights by moving property to the individual, debt by aligning the interests of the powerful with those of the debtors). We, the taxpayers, pay for a lot of things, but that doesn’t mean they’re free to us. Corn? It’s subsidized by taxpayers. Tomatoes are subsidized in the EU — should they be free to EU citizens? Of course not. Getting them from farm to market is expensive, and that’s not subsidized. It’s the growing/farming that’s subsidized, not the shipping, merchandising, and so forth. Publishing is similar — the research may be subsidized, but the publishing isn’t. Even when it is, it isn’t subsidized sufficiently because maintaining, extending, and enhancing publications over decades is expensive.

If OA is destined to be cheaper, show me when in the last 10 years an OA publisher has reduced fees. I can’t think of an instance, and certainly these mysterious efficiencies would have started to show after a decade.

One more thing — most OA publishers, PLoS included, rely on volunteer labor. In fact, in defending PLoS from perceived attacks, this is often invoked. People love PLoS, which is great. But are they exploited? What if PLoS starts generating surpluses, which they fully intend to do? Where will those be shared? Will all this free labor be exploitation then?

Hi Kent,

Thanks for your clarification. You are correct, we were not communicating. I was looking at the total price paid for a paper in cost/download. So if a paper costs $10 to publish and is viewed 100 times, the cost/download is 10 cents. My calculations were back of the envelope ones assuming that 56 cents is a standard cost per download for a NPG article, and that the cost that PLoS charges authors covers their costs. (While UC does pay PLoS, from what I read on the PLoS web site, this payment results in less fees for authors, so it could be a wash.) Take my calculations as you will.

I’m not missing your point about altruism. You are mistaking cause & effect. You are assuming that x causes y based on the fact that x & y occur together; that is, that because people work hard on things without pay, therefore not paying people makes them work hard. This is a fallacy. I will look at some of the literature, but the relationship of motivation to pay & other things is a complex one, & in the past I have not been impressed with Shirky’s (I don’t know Dan Pink) approach to complicated questions.

Re. R&D, as far as I am concerned basic research is really what matters here. R&D for a new soap brand is not of any particular importance to society as a whole.

If you don’t think the marginal cost of a download approaches zero as the number of downloads increases, then you don’t understand the concept of marginal cost. http://www.google.com/search?q=marginal+cost

Sorry I was unclear about “free” and “freedom”. I was not rhethorically linking “free” and “freedom” but distinguishing “gratis” and “free” as the Free Software Foundation does.

I do think that if government & universities pay for the research, and most of the costs of publication via subscription fees, & the marginal cost of access approaches zero, then this research should be free to all.

I do not know if OA is cheaper or not, or if it is getting cheaper. The evidence seems to point that way, for those without blinders on, but frankly I don’t think it is that important.

I can only assume that your final comment is some sort of joke. Are you arguing that OA publishers might, possibly, become as bad as non-OA publishers?

OK, let’s go through this.

Why are you and I debating this, and not getting paid for it? Because we care, and because we’re human. When people care, they tend to do things intensely and without remuneration. In fact, studies have shown that remuneration actually inhibits activities when people care about something — pay spoils the vibe, creates cynicism, and quashes idealism. As I noted, a lot of PLoS and OA people pride themselves on their free labor contributions. Nothing wrong with that, but they’re making the same gesture as unpaid peer-reviewers for subscription-based society journals, whether they’re run by a for-profit publisher or by a society. As networking effort becomes easier, there’s every reason to think that more collaboration (without pay) will occur. The proliferation of OA editorial boards offers some evidence here. I just don’t think you can call one set of people dupes and another idealists when they’re both contributing to money-making operations. And, like it or not, PLoS and other OA publishers want to make enough money to sustain themselves and pay a little forward for investments. It’s clear in their communications about finances, and in the fact that they got seed money to start, raised prices when this ran out, and continue to pursue that elusive “break even” in their annual reports.

Marginal costs are extremely elusive in the networked world, as the model is really around manufacturing output. So, you’re right, we shouldn’t have even been using that term. Let’s shift to average cost, since this is clearly more relevant in the digital world — there is no “next production” of an article, only one production event, but that has to carry the fixed costs associated with it. So, thanks for moving me ahead here. I will now and forever recognize this trap of referring to the next digital copy via “marginal cost” as a red herring. We should always refer to these as “average cost.”

So, then, if the average cost of publication (or access) doesn’t approach zero — and, again, if it truly did, then PLoS would be rolling in money — then government offsets for research and so forth need to be recalculated to address average cost, not marginal cost. Again, that was a great point you made, and showed me just how misleading this whole “marginal cost” schtick has been.

Evidence is showing (and prices at OA journals reflect this) that OA is getting more expensive. Again, fixed costs are rising, so the average cost of articles is going to go up. There’s no way around this.

My final comment was not a joke. How do you reconcile a for-profit OA publisher with your world view?

Hi Kent,

Thanks for your reply. I agree with most of what you have just said. Publishing (on the internet) is a world of large up-front fixed costs (the first copy) and marginal costs that approach zero.

I think you misunderstand my worldview. I don’t think it is “dupes” vs. “idealists”. There is probably a place for for-profit OA publishers. There is certainly a place for people making a living.

I like the open access model for the same reason that I like free & open source software. I think it is the right model in a world in which copies of software or articles can be made freely. I think it is wrong to be unable to share knowledge with others. I want to enable that sharing. And proprietary software, and proprietary publishing, prevents people from sharing that knowledge. (Harry Potter is probably a different thing; I don’t know what the right model is for creative entertainment).

Regarding the question of PLoS generating surpluses. First, PLoS is a non-profit, so anything above operating costs would have to be put back into activities that support their core mission. But say a for-profit does. What is the problem? Authors retain the copyright. If the authors don’t like the model, they can publish with a different publisher. If the readers don’t like the model, they can get all the content from the publisher (or author) and set up a new web site. I just don’t see the issue. On the other hand, with a publisher that obtains copyright and charges for access, the authors cannot move their content out of the publication, and readers cannot access it.

The issue of fixed costs vs marginal costs is an important one -though those are probably not the best terms to be using. In the print days we talked of “first copy costs” in journal publishing. First copy costs were the fixed costs and included everything necessary to produce the first copy of a journal issue: editorial costs, production costs, office overhead, etc. The marginal costs were then the cost of printing and shipping each issue.

The problem is that this model does not port to online publishing because there are a host of costs associated with an article after is is published. There are online hosting costs, bandwidth costs, archiving costs, and costs of conversion to new formats. Journals that began publishing in the mid-1990s have likely gone through several iterations of SGML and XML DTDs – requiring conversion. Articles published prior to online publishing have had to be scanned (for PDF) and/or keyed (for XML). Reading on ereaders requires further conversion to EPUB. The list goes on.

The point is not that the cost of a download is/not marginal, the point is that the costs associated with an article do not end at the time an article is published as they once did. And that all the future costs associated with an article are not even known at this point. Who would have known 50 years ago that there would be a cost to converting an article published then to PDF and XML? Who thought of ongoing costs for digital archiving incurred by the publisher? in the print world, libraries payed for the costs associated with archiving.

I am not aware of any publishers setting up a fund from current OA publication fees to support the future costs associated with those articles. Therefore, an issue to be aware of in thinking through the economics of author-pays OA is that new articles have to also cover the ongoing costs of upkeep and conversion of older articles. This is perhaps one reason why article publication fees are on the rise.

Comments are closed.