We've failed: Pirate black open access is trumping green and gold and we must change our approach



Key points

  • Sci-Hub has made nearly all articles freely available using a black open access model, leaving green and gold models in its dust.
  • Why, after 20 years of effort, have green and gold open access not achieved more? Do we need ‘tae think again’?
  • If human nature is to postpone change for as long as possible, are green and gold open access fundamentally flawed?
  • Open and closed publishing models depend on bundle pricing paid by one stakeholder, the others getting a free ride. Is unbundling a fairer model?
  • If publishers changed course and unbundled their product, would this open a legal, fairer route to 100% open access and see off the pirates?


At the 2017 UKSG Conference, a show of hands during Stuart Lawson's plenary on ‘Access, ethics and piracy’ showed that barely anyone present blocked access to Sci-Hub or considered it should be blocked (Wright, 2017). What I find startling is not that a room full of UKSG delegates seem to be condoning piracy and supporting black open access (Björk, 2017), but 15 years on from the Budapest Open Access Declaration, a pirate site is needed at all. After all, the pirates have long since been chased out of the music business (Gapper, 2017).

It is not as if the past two decades have been spent idly. Open access advocates have busily encouraged stakeholders into funding both high-profile ‘alt-publishing’ efforts and lower-profile more-or-less open access journals (e.g. https://doaj.org/) and brought forth a roar of policies and mandates that aim to oblige authors to change their publishing habits (https://roarmap.eprints.org/). All this accompanied by a controversy of sessions at conferences like UKSG, which have debated the many facets of open access to the point where conference organizers must be getting desperate to find an original angle. There is even a tool to help you see where a journal sits on an open access spectrum (http://www.oaspectrum.org). Let us also recognise that every single stakeholder in the scholarly communications industry claims to be supportive of open access, yes, including publishers, commercial or not (STM, 2016), and have set up associations to help things along (e.g. https://sparcopen.org/ and https://oaspa.org/). Now, some European countries are trying a new approach: to demand of the major publishers nationwide open access contracts, such as Projekt DEAL in Germany (Vogel & Kupferschmidt, 2017).

Yet, while we have been bickering about the true path to open access nirvana, the pirates have crept up on us, especially in the form of Sci-Hub, which is self-reporting more than 60 million articles freely available (Sci-Hub, 2017) and could have harvested nearly all scholarly literature (Himmelstein, Romeo, McLaughlin, Greshake, & Greene, 2017) – if true, Sci-Hub has single-handedly won the race to make all journal articles open access.

Set against this are the combined efforts of stakeholders in scholarly communications who, after two decades, have managed only to get around half the world's research articles open, with the rest still behind a paywall 3–4 years post-publication (Boselli & Galindo-Rueda, 2016; SIMBA, 2016). If past performance is any guidance, around four-fifths of all new scholarly articles in 2017 will be unavailable for most people on publication via legal channels (Research Consulting, 2017; Taylor, 2017). It does not look impressive: black open access has trumped green and gold.

For books, despite initiatives like Open Access Publishing in European Networks (OAPEN), Knowledge Unlatched, and Open Book Publishers, progress has been glacial. At the time of writing, there are just over 8,000 titles listed in the Directory of Open Access Books (http://www.doabooks.org/), which – considering that Springer alone offers nearly 280,000 titles from its online bookshop – suggests that the proportion of books published open access has yet to reach 2%.

Not for the first time, pirates are delivering where the established players and legal channels are not (e.g. https://en.wikipedia.org/wiki/Radio_Luxembourg). To see off pirates, the music industry recognized that they had to shift from relying only on legal means to shut down the pirates and also revolutionize their business models so that it was more attractive to users than a free, pirate service (The Economist, 2009). Not only did they find one with pay-to-download services, like iTunes, but they are now well on their way to complementing it with another service built around streaming (Arthur, 2016).

The evidence above says that green and gold open access models are not the revolutionary business models we need because, if they were, then they would have >80% market share already, the pirates would be looking elsewhere for opportunities, and I would not be writing this piece. True, there are some sustainable open access successes like BioMed Central, PLoS, and arXiv, but their share of all articles, let alone books, remains marginal. To see off the pirates (and to nick a line from O Flower of Scotland), we need tae think again.

So, why have green and gold failed? I do not blame political leaders, the power of monopolies, or unwilling academics (Rice, 2017). Rather, I think that these models are flawed for two key reasons: change and bundle pricing.


As the famous management guru Peter Drucker observed, even when everybody has accepted that change is unavoidable, change is like death and taxes – it should be postponed as long as possible, and no change would be vastly preferable (Drucker, 1999). In a recent report, six roadblocks to open access were identified involving at least five stakeholders, and all six blockages would need to be removed to deliver 100% open access in quick order (Research Consulting, 2017). If Drucker was right, then attempting to coordinate this number of stakeholders on six different fronts, across the globe, to deliver change is unrealistic.

Let us look at one stakeholder, the author, as an example. For green open access, at a minimum, the author needs to change from his or her traditional approach as follows: he or she would need to select a journal that allows a green version to be posted in a repository and then find a suitable repository on which to post it. This does not sound like much, but without mandates, only around a fifth of authors actually make the effort to deposit green versions, a figure that struggles to rise above 70% with mandates (Gargouri, Larivière, & Harnad, 2013; Poynder, 2011). A study of Spanish researchers in 2016 showed that, when allowed, just 13% of authors posted green versions on their institutional repository, and allowed or not, just over half posted full-text versions on ResearchGate (Borrego, 2017). It seems that even a little change involving no out-of-pocket cost is a tough ask for authors, even when backed up by a mandate.

For gold, at a minimum, the author needs to find a suitable gold open access journal and, sometimes, find funds to pay the publishing bill. As we have seen above, with less than 20% of all new articles published in gold journals, for most authors, not changing to gold is vastly preferable.

Space precludes going into the changes for the other actors, but I am sure readers will know in broad terms what changes are needed by each, even at a minimal level, and the results above suggest that most are in ‘postpone change’ mode.

By my reckoning, all six actors would need to change something in their old work patterns to deliver green open access, and all but librarians (in reader-support mode) and readers would have to change to deliver gold open access (see Table 1). The bottom line is that for both green and gold open access, a lot of actors need to change what they do.

Table 1. Stakeholder changes required to move to green or gold OA.
 Green OAGold OA
  1. OA, open access.
AuthorChange requiredChange required
Author's institutionChange requiredChange required
FunderChange requiredChange required
Librarian (reader-side)Change requiredNo change required
PublisherChange requiredChange required
ReaderChange requiredNo change required

To make it more challenging still, for the green open access model to be a success, all six actors need to change some aspect of their behaviour in concert – the model fails if any one actor does not change or fails to cooperate with others. For example, the main reason Spanish authors did not post green versions in their institution's repositories was because they were not aware of its existence (Borrego, 2017), suggesting a breakdown in communication between the author and whoever ran the repository. For gold, four of the six must change from past practices, again, in concert.

The changes need to be made in concert because green and gold are complex processes dependent on several actors for completion: some stakeholders need reports from others, new relationships and business processes have to be negotiated, and new internal workflows and roles are needed too (Kingsley, 2017). In this light, and if Drucker was right and our preference is not to change, is it any wonder that green and gold open access have failed to scale?

That we have managed to get roughly halfway there in terms of articles being open access in some form is, in this light, not a bad result, but it has taken nearly two decades and the pace of change for journal articles may be slowing (Boselli & Galindo-Rueda, 2016; SIMBA, 2016). But if green and gold are not going to get us to 100% open access, maybe it is time to look for another route that requires less change, but which? A clue might be found in the airline industry.


In the traditional airline industry model, to get to B from A, one used to purchase a ticket, which covered the cost of a travel bundle: you were carried, fed, watered, entertained, and could take as much or as little baggage as you wanted. Today, led by low-cost airlines, the product has been unbundled: food, drinks, seat allocation, baggage, changing tickets, and even the way you pay are now being priced as extras to the core service of getting you to B from A. Today, low-cost airlines earn a significant amount (20–40%) of their revenue from unbundled, non-core services (IdeaWorks, 2017).

The airline business involves a large number of stakeholders across multiple jurisdictions subject to a cat's cradle of international policies and agreements. The passengers may be subject to rules imposed by the passenger's employer and then there are parties at either end of a flight who have an interest in the journey too. This is not entirely dissimilar to the multi-stakeholder, global, and increasingly policy-bound environment of scholarly communications, so making a parallel between the two industries is not as ridiculous as it might seem.

In unbundling, airlines exposed optional extras to market forces and if they were not wanted by passengers they could be cut, saving cost from the overall system. Equally, to survive in a competitive environment, airlines had to be ruthless in terms of streamlining processes, focusing on the customer's core needs, and being creative in looking for new services and partnerships to generate additional revenue (CNN, 2016).

Scholarly communications also comprise a large number of services (Anderson, 2016). In the traditional closed access model, these services are bundled into a single product that is purchased by the reader or reader's representative. In the gold open access model, these services are bundled into a single product that is paid for by the author, author's representative, or benefactor. The green open access model is different in that it comprises two parallel products: one offering closed access to the version of record, paid for by the reader or reader's representative, and the other offering open access to a ‘green’ version, paid for either by the author's employer and/or by one or more institutions or benefactors with an interest in the subject.

In all three models, closed, green, and gold, an author's manuscript is transformed into a published work that is made accessible to readers. Yet, the process has value to other stakeholders too: employers, funders, librarians, publishers, and policy makers benefit in the form of enhanced reputation, mission advancement, reader services, return on investment, and societal benefit, respectively. Yet, because all of the value is captured in a single bundle price paid for by just one of these stakeholders, it could be argued that the others are getting a free ride.

What if, like the airline industry, publishers unbundled their product and started to test the value of some of the elements that form the bundle? To achieve a basic level of open access (after all, this is our collective goal, right?) publishers would have to offer a basic read-only service for free and work up from there, seeking revenue from premium services that surround the content.

Offering a basic service to readers for free would qualify as gratis open access (Suber, 2008). This would be like reaching base camp in terms of climbing the open access mountain, but as we are only halfway to base camp with journal articles today and have barely started lacing our boots for books, that would be a significant improvement. Looping back to Drucker's point about change, launching a free, read-only service would require only one stakeholder, the publisher, to make a change, so it would be much easier to deliver.

The question then is this: if everyone could read all scholarly content for free, is there sufficient value in additional services to generate the revenues needed to fund both a read-only service and for those other elements of the scholarly communication process that, once unbundled, survive exposure to market forces?

To try and answer this question, let us go back to the airlines. Not everyone wants to stow multiple bags, eat, drink, change their tickets, and choose a bigger seat, but a number of passengers do. Employers sometimes pay for business class so their employees are more productive at their destination (or, at least, feel important and valued by their employer); the wealthy pay for fast-track services, the privacy of private lounges and comfy seats. Low-cost airlines have found that two-thirds of customers will choose to purchase optional extras, and the airlines also earn a striking amount of money from partnerships, notably from credit cards issued with frequent flyer programmes (IdeaWorks, 2017).

So, in scholarly communications, where might publishers find revenue from services that go beyond the provision of a basic free service for readers? What partnerships could be fruitful?

On the author side, peer review management, copyediting, and language services could be offered along with services to promote the publications and make them more accessible to non-technical audiences. On the reader side, services could include higher-utility versions of articles, books, and data sets; access to productivity tools like downloadable citations, semantically driven navigation, and alerting services; and tools to report impact to funders and employers. Librarians might value rich metadata and feeds to build catalogue databases, usage reports, and user support services. Funders and employers could be offered reports by subject area. If the airline industry is anything to go by, publishers will diverge in terms of the services they offer and become less like each other as they find what works and what does not (IdeaWorks, 2017).

In unbundling the product, low-cost airlines democratized air travel, reducing prices such that the market for flying grew (Aviation Economics, 2014). Likewise, in democratizing scholarly communications by offering a basic service to all for free, readership will expand and could stimulate a market for services beyond the existing customer base, tapping into new budgets. Similar to other digital businesses, which build large audiences with free services, there will be value in ‘selling the audience’ to advertisers and others with an interest in accessing niche audiences. As Springer-Nature is finding out with its Nature Partner Journals programme, there are also opportunities in building partnerships.

The evidence that a larger readership is there if only the works were available at low or no cost is growing (Adie, 2014), and our own experience at Organisation for Economic Cooperation and Development (OECD) is that total readership expanded more than five-fold following the introduction of our read-only, free-of-charge service, and it did not cannibalize our ability to earn revenues from premium editions and associated services.


In this paper, I argue that while green and gold open access have been put forward as two worthy emperors, it is time to conclude that they are unclothed. This is not to shame them or those who argue in their favour; but surely, it is time to recognize that, in this age of digital disruption, there must be something structurally wrong with an approach that after two decades of collective effort has yet to reach base camp. The pirates have recognized this, and it is time we did too.

Green and gold open access can only scale if all stakeholders change in concert. In view of Drucker's observation on our collective preference not to change, is it any wonder these models have not worked and that black open access is leading the way? Consider the relative ease of agreeing with your co-authors on which institutional or subject repository to post which version of your work in line with multiple stakeholder policies (employer, funder, and publisher) as compared with self-posting in your own time to an academic sharing network of which you are already a member or, indeed, doing nothing and leaving it to Sci-Hub.

Yes, initiatives like CHORUS ease the pain and reduce the level of change needed (https://www.chorusaccess.org/about/about-chorus/), and publishers like Springer-Nature are experimenting with free article sharing (http://www.springernature.com/gp/researchers/sharedit), and initiatives like Projekt DEAL might work in some countries, but in looking for a new emperor, I suggest that we might be encouraged by the airline industry and unbundle the product. This would make all content free to read, answering the plea that the results of publicly funded research be available to the public, reveal the true values for the existing bundle's component parts, and lead to a situation where each stakeholder has the choice to pay for the particular benefit they get from the scholarly communication process. This might prove to be a fairer, cheaper, more sustainable, and less controversial model in the long run.

Critically, only one actor is needed to start this process of unbundling: the publisher. In making a basic, legal version free for anyone to read, gratis open access is achieved at a stroke, and it would start to make the pirates redundant.