In a recent post, I came up with a per-article costing figure, based on Ubiquity Press's economy of scale, for a learned society to go open access. I received two responses to the figure given. One said that it was extortionately high. The other queried whether it was perhaps unrealistically low. What are the real costs of scholarly communication?
Often it is very hard for us to know. Some publishers (as is well documented pretty much everywhere in OA circles) do very well out of selling research material back to academic libraries at extortionate rates of profit. Other mission-driven publishers, though, do not. Despite similar levels of pricing, if a certain scale isn't reached, it is impossible for a fair comparison to be made between entities. Individual pricing does not equal revenue, even if the underlying labour costs are the same (but the profit different).
One of the most common arguments used to drive down the price of web publishing is Clary Shirky's assertion that publishing isn't an industry now; “it's a button”. In some spheres, this may be true. But how are you defining publishing when you say this? Is it about “making public”? In which case, then, yes, you can make something public for very very little money. This is not unique to the digital sphere, though. To borrow an example from Michael Bhaskar's excellent book, if you print a single copy of an article and put it on a bench in a park, is this “published”? The difference in the analogy is that the systems of discoverability on the internet mean that putting up a single copy can give you an extremely wide reach. It's not quite the same, but there is more to publishing than just putting something online. As Mike Taylor has brilliantly pointed out, though, there's no reason to couple the value-adding services that we want to a mode of copyright transfer and sale.
That said, I'm still not persuaded that we should race to the bottom on the costs of running a successful new scholarly communication enterprise. What about the labour of publishing? arXiv, the most frequently cited source for low-cost publishing (as in “making public”), still had a pretty strong revenue requirement even back in 2010: “The calendar year 2010 budget for arXiv is $400,000, which includes costs for personnel and operating expenses (G&A overhead, hardware, hosting, and network charges [...]). Staff salaries account for about 80% of total annual expenses. We expect the annual budget to increase to $500,000 by 2012 to facilitate necessary upgrades and enhancements”. It's cheaper than profit-making entities. It's hardly nothing, though,
For an enterprise undertaking more than arXiv, there are costs of membership to valuable organisations that fulfil different functions: CLOCKSS, CrossRef and COPE, for instance, are all pretty crucial parts of the ecosystem (preservation, addressability and ethics). There are the costs of the platform hosting and the server itself. Sure, go with Wordpress, but there's still a total cost of ownership and maintenance if you go at scale. If you're hosting something for free somewhere, is this ad-supported and might that compromise your effort differently? Are you going to do XML typesetting for forward migration of semantic content to avoid bit rot? There's a cost.
What happens if a publisher gets sued? Do we want such entities to be one lawsuit away from bankruptcy? Do we want publishers to be able to cross-subsidise the publication of research where there isn't a market? (This could be either APC/BPC-waivers in an OA environment or risk-taking in a book-list in a hybrid sales environment.) Do we want new publishers to have some cash so that they can transform the space and compete on some level? If not, surely existing entities will throw money at the problem to keep the system as it stands in their favour, leading to stasis. There's also a recommended legal amount of surplus to legally wrap up an entity if things go pear shaped.
This is not to approve of the current practices of many of the larger academic publishers. As I wrote elsewhere, recently, our forebears spent a long time building the principles of freedom of inquiry into the university. One of the most crucial of these measures was that academics should not have to produce research that would sell. In fact, the choice was deliberately made to separate the selection of research topic from market populism because it is clear that the investigation of esoteric areas, selected by experts, can yield better results than financial incentives posing as democracy. The freedom from having to sell.
These principles are, of course, perpetually under assault in the contemporary marketisation of higher education. However, a more long-standing problem comes in the form of subscription and sales publishing for academic research. While we gave researchers freedom to investigate wheresoever they felt there might be merit, we concurrently outsourced publication to entities that were reliant on the market for their income. Furthermore, they were often profit-driven entities whose motivation was shareholder, rather than mission, orientated.
Which is all a rather long-winded way of saying that I think the true cost of scholarly communication lies between the camps. Again, I think that Mike Taylor has done a good job, for instance, of setting out some of the stuff that we shouldn't be paying for. That said, there are, to be blunt on the ethics, good and bad types of surplus in my books. I am not a fan of the shareholder, for-profit model for scholcomms. I do think, though, that a surplus used to transform the field for the better, via mission-driven routes, is a good way to proceed. Any effort that wants to seriously compete to transform the scholarly communications ecosystem, at scale, will need to build some degree of surplus into its model for both growth and safety reasons. To over estimate these costs in the interest of commercial gain isn't right and puts private ahead of public interest; which should not be the goal of sciences or the humanities. To under estimate them in the rhetorical service of cutting costs is not strictly accurate and makes it harder to create effective new enterprises that can really compete with existing players for the greater good.