Elsevier and the Dutch Open Science goals
The VSNU, NFU, NWO and Elsevier have announced a national deal that bundles Open Access and data services. Is the deal consistent with Dutch Open Science goals, and will undesirable platform effects be avoided?
Yesterday it was announced
that the Association of Universities in the Netherlands (VSNU), the Netherlands Federation of University Medical Centres (NFU), the Dutch Research Council (NWO) and Elsevier have reached a national deal that includes Open Access publishing and reading services. The deal had been long in the making, and the road was bumpy. An important feature of the deal is that it moves beyond Open Access: it is also an agreement about the joint development of new research intelligence services. The bundling of Open Access and data services was formally announced in December, but the news leaked a month earlier. The contract with Elsevier has been made publicly available, except for the financial agreements made.
Because of this link between access and research intelligence services, the VSNU, NFU and NWO installed an expert Taskforce at the beginning of 2020 to “address issues around the responsible use of research information and the role of commercial third-party providers in particular.” I am a member of the Taskforce. The relevance of the Taskforce is explained in our assignment:
“There is an emerging market for third party providers offering services to satisfy the growing demand for research information and evaluation. Using large scale data collection, aggregation and analysis, these services provide new prospects for assisted decision making on, for example: - funding opportunities, - publishing venues, - identifying upcoming research fields, - alternative metrics. As critical functions of the scholarly enterprise become increasingly dependent on such services, it is critical that the academy itself carefully considers risks involved in becoming too dependent on specific market players and their tightly integrated solutions. The increasing interwovenness of information about research (research intelligence) and research itself raises a number of challenging issues both for users and for producers of this information.”
Two important aims of the Taskforce are to:
- Establish a set of terms and conditions under which metadata of public research output can be (re)used and enriched by public and private organizations, in accordance with research ethics and public values and avoiding undesired network or platform effects
- Describe the concept of an Open Knowledge Base (OKB), in which an open infrastructure is developed in cooperation with private parties for the responsible management of research information and data, consistent with Dutch Open Science goals
An open consultation is currently underway on the Guiding Principles the Taskforce formulated for collaboration between Dutch research institutions and third-party organizations in developing new services based on (meta)data use. We have formulated principles on 1) Ownership of (meta)data; 2) Enduring access; 3) Trusted and transparent provenance; 4) Interoperability as part of community owned governance; 5) Open collaboration with the market; and 6) Community owned governance. These principles have also been offered to the steering committee of VSNU, NFU and NWO, who were responsible for the negotiations with Elsevier.
In cash, or in kind?
So what exactly does Elsevier get with this deal? And how does the deal hold up when compared to our Guiding Principles?
To begin with, it is definitely laudable that Elsevier committed to a set of collaboration principles, including data ownership (researchers and/or institutions own their own research data), interoperability, and institutional discretion on the use of the services. But there is also cause for concern.
The contract mentions examples of potential pilots to develop research intelligence services. The first potential pilot, for example, aims to “[i]mprove findability and visibility of NL research outputs by aggregating and deduplicating separate CRIS systems into a Pure Community module available to all institutions which can serve as a building block to a NL open knowledge base.” (p. 103)
As part of a community-owned Open Knowledge Base, such a pilot does not seem acceptable to me. I can well imagine the interest research institutions may have in such a pilot; an enrichment of metadata is interesting for them. But it would basically let a public infrastructure be controlled by Elsevier modules by building them in from the start. The research institutions should first carefully consider whether or not a pilot with a Pure Community Module is truly in line with Dutch Open Science goals. They should also more carefully consider whether or not this pilot would undermine plans for a community-owned OKB.
There is a simple solution though. This is to disconnect the Pure Community Module from the contract for now. Subsequently, a tender could be issued for a Current Research Information Systems (CRIS) module that does something similar, and that can also process (meta) data from other sources. Elsevier can then tender for this contract. This would also give research institutions the opportunity to make an inventory of what alternatives are currently available for the Pure Community Module. If Elsevier wins the tender and their proposal is implemented, it will have been properly weighed against alternatives.
A crucial difference is that Elsevier is then paid in cash for their services, and not in “kind” by giving them an insurmountable competitive advantage in terms of access to research intelligence. This is a very important point of principle.
All the pilots that Elsevier wants to commit to as well as the OKB itself aim to establish relationships between items, mainly by linking metadata. The Guiding Principles of our Taskforce stipulate that universities and other relevant institutions must have access to Dutch research information (metadata), including "derived" information. Therefore, a very important point is whether sustainable access is negotiated to relationships between entities established by third parties, in this case Elsevier. If I read the negotiation agreement carefully, I wonder if this is the case. Will Elsevier deposit proper metadata to Crossref? Will this metadata be made fully open, also in the case of citation data?
Limited financial means
This agreement will obviously give Elsevier a competitive edge, because they have already secured a number of pilots. And this agreement also implies that universities will have less money left for issuing contracts to other parties. The financial ramifications are especially relevant now that we are confronted with major economic consequences as a result of the COVID-19 pandemic. Should the competitive advantage be seen as a compensation for Elsevier’s willingness to engage in a read and publish deal? I am not persuaded. It would have been far more preferable if the research institutions would have first formulated general principles for collaboration with private parties, and only then had started to engage in projects and look for third-party interest. It seems sensible to me to stick to such a model as much as possible, for example by decoupling certain projects from the agreement and putting them into a general tender. In my view, funders would do well to also consider this model for the Funder Information pilot mentioned in the agreement, which has the stated aim to “link NL research outputs to grants and funders (EC, ERC, NWO, RVO, ZonMw), to allow for improved tracking / assessment of impact of funded research.” Here, too, it is crucial to avoid vendor lock-in. And of course, Elsevier should be given the opportunity to apply for these tenders. The option of issuing a tender per pilot has been left open in the contract with Elsevier.
Value-extraction from public-sector information
It is evident that Elsevier has quite a lot to gain from this deal. What is in store for them is a unique research intelligence infrastructure, because it is not only a national-level arrangement, but on top of that, the information that goes in is also validated by the institutes and research funders themselves. Obviously, Elsevier wants to manage a wide range of research intelligence to offer analyses from the perspectives of multiple stakeholders. Therefore, the company aims to establish a comprehensive linkage of metadata related to institutions, researchers, output, funding, and other resources. The Open Knowledge Base that the research institutions are themselves considering has similar ambitions. What remains to be seen is whether our research institutions are willing and able to ‘step up to invest in home-grown research infrastructures’. It is also an open question whether Elsevier systems will be made open and inclusive enough to comply with the Guiding Principles we formulated on behalf of the research institutions. I am not persuaded by the contract, and still find it disconcerting that this deal may effectively transfer crucial means to influence Dutch science policy to a monopolistic private enterprise.