Recently, Dr. Louie Dean Valencia-Garcia addressed the ethical challenge researchers face when purchasing extremist materials. In buying books from publishers, subscribing to podcasts, or paying for membership in web forums, researchers are indeed helping to fund these extremist projects. While that support may be small, Valencia-Garcia points out, its ethical implications must be reckoned with.
However, in a world where commerce is mediated by platform giants such as Amazon, that question of funding becomes much more complex than a simple exchange of money and goods. Valencia-Garcia alludes to these questions in his excellent piece. Here, I expand on his thoughts, to sketch the practical dimensions of this ethical dilemma from the perspective of the economy of digital platforms.
The value of a sale on a digital platform such as Amazon goes far beyond mere dollars or euros. When we buy anything from Amazon, our purchase is noted in a vast reservoir of data, which then informs key decisions affecting Amazon’s future treatment of that product. Everything from the title’s ranking in search results, to its inclusion in the “Customers Also Bought” recommendation menu, to Amazon’s decision to carry the book in its warehouses, is positively impacted by purchases.
Crucially, because of a dynamic known as “network effects”, the cumulative value of this data is worth more than the sum of its parts. The value of two users’ data is worth more than double the value of one user’s data. The value of three users’ data is worth disproportionally even more. And that nonlinear value cuts both ways. For authors and publishers, multiple purchases on the Amazon platform offer a near-exponential value in the form of automatic recommendations and good search result placement.
Amazon typically keeps 60% of a book’s cover price from a purchase. A typical paperback from Arktos (a leading right-wing extremist publishing house) costs between $15-$20. Small commercial publishers such as Arktos rarely exceed 10% profit margin. And so, if scholars choose to purchase from Amazon, extremist publishers will obtain relatively little of the profit necessary to cover their expenses and to invest in further project development. So there is no doubt that purchasing a text directly from the publisher does increases per-unit profits.
When we consider this as researchers, it would seem that the profit kept from extremist publishers through Amazon purchases probably outweighs any contribution to network effects on the platform. However, as scholars, we do not operate individually. And network effects are very valuable. This is even more true for small, insular genres such as white nationalist literature. By purchasing an extremist text on Amazon, we collectively contribute to a linear decrease in publishers’ profits, while exponentially increasing the likelihood that extremist titles will appear in recommendations and search results. It does not take very many researchers purchasing Arktos titles before our contribution to algorithmic promotion outweighs the publisher’s lost profits.
We are left, then, to conjecture whether the “least worst” option for purchasing new extremist titles is not counter-intuitively to purchase directly from the publisher! Extremist publishers like Arktos face the same financial instability as other independent publishers. They often operate on shoestring budgets, and sometimes depend on cash donations to fund unprofitable business cycles. Let us not forget, too, that the very purpose of profitability for these companies is to increase the exposure of their noxious ideologies. By contributing to their algorithmic exposure on platforms such as Amazon, researchers might ironically be doing that publicity work on their behalf. Any extra profits obtained by scholars purchasing directly from extremist publishers may well be less valuable than the contribution to network effects when purchasing on a digital sales platform. If this is so, then attempts to deny these publishers extra monetary profits has the unintended outcome of contributing to their visibility on the platform—in activist language, “platforming” right wing extremists.
This is, of course, only a hypothesis. But it is a hypothesis with strong backing in the realities of digital business. And its answer should be easily quantifiable. Indeed, Amazon has built its sales empire on that very quantifiability. It is possible—trivial even—to measure the rate of sales and added exposure that come from Amazon’s data-driven recommendations and search placement. It is likewise possible to measure the value of the network effect on recommendations and search returns. Unfortunately, the core data that would enable researchers to definitively answer these questions are proprietary. These algorithms and metrics are the intellectual property of Amazon, a trade secret which Amazon does not share with the public, including scholars. This is unlikely to change.
In the meantime, researchers will need to develop ethical heuristics to guide the purchase of extremist texts. I conclude by noting some points where such a conversation might take place.
- In his piece, Valencia-Garcia settles on second-hand purchases as a tolerable ethical compromise. In doing so, he removes a title from circulation while denying further profits to publishers. Of course, this is not always possible. And as Valencia-Garcia notes, this approach does not solve the issue of algorithmic promotion. Extremism scholars should develop private regional archives, available to vetted researchers, and stocked with second-hand texts. This will minimize contribution to algorithmic promotion, while expanding scholarly access and denying extremists per-unit profits.
- As Valencia-Garcia discovered in his role as a Harvard lecturer, a single inter-library loan can lead an institution to purchase multiple titles from extremist publishers, not only contributing to their profitability but also legitimizing them with shelf space in prestigious universities. Researchers and librarians can work together to develop policies ensuring that research does not lead to legitimation.
- Researchers should always refrain from purchasing e-books, whose profit margins can run up to 75%. Pirated copies of extremist texts are widely available online. Extremist forums will often make “packs” of PDF files available featuring the latest in far right literature. Scholars should also network with one another to develop virtual archives where these texts may be found and shared freely. Scanning projects of extremist literature are excellent opportunities, particularly for graduate students, to familiarize themselves with the breadth of primary literature.
The economy of digital sales platforms such as Amazon have changed the conditions of the publishing market. While our moral intuition prompts us to deny extremist publishers every penny possible, the realities of the digital economy make such calculations anything but simple. Dr. Valencia-Garcia has opened up a critical avenue of conversation surrounding this topic. This is an excellent opportunity to develop practical, fact-based answers for the digital age and to collectively recognize field-wide best practices to the problems he has posed.
Mr Brian Hughes is an Early Career Research Fellow at CARR and Doctoral candidate at the School of Communication, American University. His profile can be found here:
© Brian Hughes. Views expressed on this website are individual contributors and do not necessarily reflect that of the Centre for Analysis of the Radical Right (CARR). We are pleased to share previously unpublished materials with the community under creative commons license 4.0 (Attribution-NoDerivatives).