Leaked Dutch Contract with Elsevier Raises Significant Alarm Bells

In recent weeks it has emerged that Elsevier is negotiating a new deal with VSNU, a consortia of Dutch Universities. According to press reports on leaked details of the deal, Elsevier is discussing a contract to provide Dutch universities with access to its journals at no extra cost (a major concession after decades of significant annual increases for most of their customers). However, the deal comes with significant new strings: Elsevier will essentially accept a “zero revenue growth” position for its journal in exchange for the universities purchasing a large set of their data analytics products. While the exact details of the deal are unconfirmed (and Elsevier has indicated that there are several inaccuracies in the leak), we have no reason to believe that the main storyline is incorrect.

There are many reasons why signing a deal like this would represent a very insidious precedent for the academic community. 

1. Linking publishing and data contracts is problematic.
   It is easy to understand why Elsevier may choose to bundle publishing and data contracts. Elsevier is under pressure from the academic community to lower its future revenue growth from journals, and, while the transition to Open Access may prove less disruptive than feared, there is still considerable uncertainty over both the end game of this transition and the complexity of managing a protracted transition phase. Contracts like this provide some degree of “insurance” from the risks posed by these issues.

   Academic institutions, however, have no reason to go along just because the contract is in the interest of Elsevier. By linking two separate products together, the academic community limits its future flexibility to cancel or discontinue either product. These products serve different purposes, are acquired by different entities within academic institutions, and compete with different alternatives. It is questionable enough to acquire a large number of journals with scarce or no interest to the research community in order to have access to the leading Elsevier journals, but it is even more questionable to have access to Cell or the Lancet contingent on purchasing  a research management system like Pure.  

2. This deal structure is bad for competition in data analytics.
   Data analytics businesses are naturally likely to become concentrated oligopolies or monopolies. In scholarly publishing there are cost areas such as sales, administration and IT that benefit from company consolidation, but many other costs (editing and peer review administration, for example) are dependent on the number of articles published. As a result, publishing has concentrated, but there are still many publishers, particularly in niche positions, and the industry is an oligopoly with three major competitors and a large group of smaller companies providing diversity and alternatives. Data analytics, on the other hand, is naturally a highly concentrated oligopoly or a monopoly because users want access to the best data and analytics or to the broadest reach, regardless of cost. It is enough to think about search engines or social networks to see how powerful these oligopolies have become even in the consumer market.

   This deal structure inflicts further damage to competition in the data analytics business. Elsevier, in fact, is the only company offering both journals and analytics. Its competitors in data analytics simply cannot match this offer. Clarivate, Academic Analytics and Digital Science sell products that compete directly with Elsevier’s data analytics, but it is difficult to justify acquiring the products of multiple vendors. It is difficult to see how Clarivate, Digital Science or Academic Analytics can match the package offered by Elsevier and remain competitive.

   Arguably, this may lead to consolidation between publishers and data analytics companies, with each leading publisher merging with an analytics company (for example, it would appear natural to bring together Springer Nature Group with Digital Science). Even so, the Elsevier publishing contract is typically the larger one a library signs, and therefore there will be an incentive for libraries and consortia to sign a “Bigger Deal” with Elsevier rather than with any other company.

3. Reduced competition will negatively impact customer leverage.
   With less competition, universities may have fewer options on all terms and conditions of their contracts. Spending is one possible area of concern, but so are issues around the transparency of algorithms, the retention of ownership of the data, the retention of the right to use outputs should contracts be discontinued, non-disclosure agreements and other clauses that have historically penalized academic institutions in their dealings with scholarly publishers.  In a market with robust competition, academic institutions may obtain favorable conditions on these issues; in a quasi-monopoly, that appears unlikely.
4. This deal structure is bad for the health of scholarly publishing.
   Historically, critics have been incensed by the high profitability of Elsevier and other large publishers of scholarly journals. However, should all revenue growth for Elsevier start to be driven by data analytics,  and expectations that  ”zero revenue growth” will become the norm across the industry, there would be margin pressure on smaller publishers. In turn, this would reduce the availability of capital to fund new technologies, new journals, etc.

   We realize that, in the near term, many critics of commercial publishers would rejoice in their difficulties. Paradoxically, however, Elsevier is best equipped to deal with “zero revenue growth” and would suffer less than anyone else (and perhaps even profit more) by shifting revenue and profits growth away from publishing and into data analytics. Smaller publishers and scholarly societies would be in an even more precarious financial position.

5. Spending is the least relevant issue.
   While the total cost of data and data analytics is likely to go up as the business becomes less competitive over time, this is the least significant problem The real problem posed by a monopoly (or quasi-monopoly) on data analytics is the loss of diversity.

   If the academic data analytics business becomes a quasi-monopoly, it is easy to imagine a dystopian future. One company (and one algorithm) may heavily influence decisions on which departments should grow in size and budget, which research projects should be funded, who should be promoted, etc. As we know, algorithms contain errors and biases, and those errors and biased could affect a vast amount of academic research. Maintaining and encouraging diversity appears necessary for reasons that go well beyond the spending.

   In addition, lack of diversity may influence even “what” academic institutions measure, not just “how.” If data analytics becomes a quasi-monopoly, what the quasi-monopolist decides to offer for sale may well become the metrics that academic institutions use to evaluate their research. Elsevier is attempting to do this through its efforts to play a leading role in defining standard for research assessment. If unchecked, Elsevier may well define on behalf of the academic community what constitutes good research and sell the tools to perform the actual assessment.

For all these reasons, these contracts are highly problematic. Institutions and consortia should pause to consider and robustly debate all the ramifications of these decisions, before pursuing what may prove apparent and short-lived benefits.