This is the speaker notes of the author’s presentation given at IFLA 2016. View the video recording of the author’s presentation. The presentation begins at 13:18.


While technology has made producing copies of digital content almost entirely free, there is no escaping that publishing, according to most definitions of the term, still requires time and money. Any publishing service offered by a library must find a way to achieve financial sustainability—that is, operate without losing money.

However, even "losing money" is a tricky concept, especially when taking into account varying definitions of operating expenses (overhead costs) under different models for auxiliary services. Libraries are by their very nature cost centers, providing services without the expectation of recovering revenue, and are usually part of larger organizations that similarly provide services under partial or full subsidies. While libraries are often comfortable with charging for convenience services and for services to those outside their designed community of users, careful thought should be given to which costs a publishing service—or any new service—should be expected to recover.

I’m going to speak today about financial sustainability of scholarly publishing and libraries, but I think it doesn’t actually make sense to talk about libraries in isolation. The way to think about this topic is similar to how we talk about it for traditional publishers, and there’s actually been quite a bit of discussion in the past year or two about measuring and relating costs in university press publishing. I’m going to try to summarize that and relate it to library publishing.

Two quick assumptions that I think will be uncontroversial and won’t really surprise anyone: I’m using the term publishing to mean more than making creative works public by simply posting them online. Rather, a publisher performs some combination of selection, editing, design, marketing, and managing distribution. These stages of publishing require a combination of time and money to carry out. Publishing isn’t actually free in the age of the Internet.

If a library creates a publishing service, it might decide that the service should be self-sustaining (not require any subsidy). I see two possible reasons that might be given for this.

First, publishing might be seen as non-essential to the mission of the library—not an essential service. The thinking is that the library’s limited resources should only fund core, essential services. A classic example is a library publishing service that just provides hosting for open-access journals. In this case, the service might be seen as a convenience for journal editors to save them from finding hosting elsewhere. Another is a library-based service for authors of books, in which the library brokers relationships between the authors and freelancers or vendors that offer editing or design services. While authors could arrange this on their own, the library has expertise in finding appropriate service providers and can save the authors some time.

Second, since most publishers are run as businesses, a library’s administration might feel that a publishing operation should also be treated as one. I think a library’s administration is more likely to hold this view if the publishing service sells something to customers directly (like copies of books in print or digital form) rather than simply providing a service to authors or editors. In this case, you might be expected to recover at least some of your costs.

Let’s assume that your library publishing service is expected to cover at least some of its costs. The question, really, is: what sorts of things do we count as contributing toward the costs of publishing? In other words, what’s in scope for cost recovery?

In economics and budgeting, costs are generally divided into two types: fixed and variable. Variable costs, also called unit-level costs, are easy to measure in publishing: what it costs to produce each individual copy of a print book or ebook. Fixed costs, on the other hand, are incurred before any copies are produced—things like staff salaries and benefits, equipment to run the publishing operation, and workspace. They form a greater portion of the costs in publishing than variable costs, but theses costs are also harder to fully represent. Fixed costs are sometimes referred to as “first-copy costs” or “zeroeth-copy costs” (since you haven’t produced even the first copy).

Another way of dividing costs is between indirect costs and direct costs. Indirect costs are shared across all of a publisher’s projects: staff time spent on administration, equipment, workspace, and other things that you can’t attribute to specific projects. Direct costs, on the other hand, include staff time spent on a particular project, fees paid to editors and designer, and printing costs. Note that staff time is included in both indirect and indirect costs, depending on whether it can be attributed to a particular project or to the organization as a whole.

You can get into debates on how to divide staff time between direct and indirect costs, but if you want to know the total costs of a publishing operation, you just need to add up the direct and indirect costs. If you’re interested in knowing how your costs compare to those of other publishers, you can simply take your total costs for a fixed period of time and divide them by the number of books, journal articles, pages, or other metric, comparing with other publishers regardless of how those publishers count direct and indirect costs.

Ithaka S+R, with the support of the Mellon Foundation, published a study earlier this year based on data about publishing scholarly monographs from about twenty university presses. In the study, the accounting of direct and indirect costs was harmonized as part of an effort to calculate the average total fixed cost of publishing scholarly monographs. But the figures varied quite a bit, in large part because of in-kind contributions from the host institution. In-kind contributions are expenses provided for by the host institution without any expectation of accounting for them. For example, a university press might not be charged for office space or phone lines, or maybe a central unit on campus provides business services, like paying invoices, to the university press. Or maybe the press has to pay staff salaries but not benefits like health insurance, which might be covered by a state government.

As you can see, it’s not fair to compare fixed costs between university presses if they receive different in-kind contributions. Indeed, the Ithaka S+R report couldn’t fairly account for these contributions because there’s no hypothetical cost for most of these, so you can’t fully compare all costs of publishing between university presses. Furthermore, university presses receive varying levels of outright subsidy for their operations, making it even harder to compare their profitability.

I’ve gone down this long road of discussing university presses because the same considerations apply to library publishers, especially when discussing which expenses a library publisher is expected to recover. A library publishing operation looking to develop or justify a cost model might consider the following:

First of all, is this library publishing service considered essential for the users—a core service of the library? If so, I think it’s pretty clear that all of the costs should be fully subsidized by the library, just like other essential services of the library are. There’s no justification for the library not to fully support it and make it free for authors and readers.

Similarly, if it’s not necessarily considered core but it’s important that you not create disincentives for authors to use the service, all costs should also be covered. Let me give an example. Libraries often incur real costs in offering interlibrary loan services: charges from other libraries. These are quite measurable, so it would be easy to pass on those costs to users. However, in many cases the library chooses not to not charge users because it’s seen as important not to penalize the reader simply because the library doesn’t already have access to that particular book or journal. So, similarly, if it’s felt to be important that all eligible authors, journal editors, or others can use the library publishing service, you don’t want to create disincentives to use.

But otherwise, if you’re going to recover some or all costs, then I think the first question to ask is whether other parts of the library—perhaps interlibrary loan, document-delivery services, or a café selling food and t-shirts—are expected to recover their costs, in part or whole. If so, you’ll want to look at how they account for costs and do it similarly. It wouldn’t make sense for the library publishing operation to be held to a different standard.

The next question to ask is how much staff time you have available for detailed cost accounting. As you can imagine, it can be difficult to accounting for certain direct costs—after all, how many of us really work on just one thing at a time and track our time that way? Still, if you can manage to track staff time spent on particular projects, you could account for it in what you charge for your services or products. But if that’s not realistic, you should probably charge only for the obvious direct costs, like fees paid to freelancers or vendors or the cost of printing copies, and not to try to do difficult things like track staff time.

If you’re interested in this topic, I recommend the following for further reading:

Kevin S. Hawkins is assistant dean for scholarly communication for the University of North Texas Libraries, where he founded the library publishing operation and more recently began overseeing consultation on research data management and copyright, plus coordination of the Libraries' efforts to foster the transformation of scholarly communication. Prior to joining UNT in 2014, he was director of publishing operations for Michigan Publishing. Kevin has also worked as visiting metadata manager for the Digital Humanities Observatory, a project of the Royal Irish Academy. He serves on advisory boards for Project MUSE, the Open Access Publishing Cooperative Study, and Editoria, served as the first president of the board of the Library Publishing Coalition, and has contributed to major standards for digital publishing. He has BAs in Russian and linguistics from the University of Maryland and an MS in library and information science from the University of Illinois.