Anyone who has spent time in a room where open access business models are being discussed knows the confusion and chaos the topic can induce. Someone mentions a transformative agreement, only to be asked by a colleague whether it is read-and-publish or publish-and-read; a third person wonders aloud whether Subscribe to Open is really a transformative model or something else entirely; and a fourth notes that the journal they edit is funded in a completely different way. And, of course, a publisher may give a formal name to their offering of a general type, creating additional complexity that challenges our ability to make sense of the landscape. 

Tasha Mellins-Cohen offered a seven-category framework in 2024, which has proven useful for the field. Today I offer a three-category framework as an alternative, not so much in disagreement, but as a way of seeing how contours of the business model landscape have smoothed out after a period of experimentation and to aid in matching models to open access mandate requirements.

In my view, models for funding open access journal article publishing can be sorted into three categories, organized by who pays and what they pay for: transactional, collective action, and subvention. 

These categories do not describe the “color” of the resulting article — all of them produce a “gold” article, i.e., the version of record is open on the publisher platform — but the financial mechanism that funds it. Sorting the models this way does not make them simpler than they are. The complexity is worth tracking. But it helps to clarify what is actually different from one arrangement to the next, and therefore what is at stake when a library or a publisher chooses among them.

Before turning to the categories, let me add two observations worth keeping in mind. First, a given publisher may implement any or all of these models across its portfolio. For example, the presence of a transformative agreement with one segment of a publisher’s list does not preclude a Subscribe to Open offering for another segment. Second, a given publication may be funded by one model, or a subset of models, but not by all of them simultaneously — some combinations are incompatible at the publication level. Keeping the publisher level and the publication level distinct helps avoid some of the confusion that can arise.

Transactional 

Transactional models are those that fund the publication of a single article or the article outputs of an institution’s authors published with a single publisher. 

The most familiar model, and in some respects the base unit of the others, is the micropayment for a single article. The article processing charge (APC) is a fee — paid by the author or on behalf of the author by their funder or institution — to publish an article open access.

What is striking is how much of the current model landscape can be read as an effort to move away from the individually invoiced APC while at some level retaining its logic. The reasons are fairly obvious. Individual payments carry significant administrative overhead — invoicing, tracking, bad debt — and aligning payment with publishing creates friction. A publisher must decide whether to delay publication until payment clears (which slows scholarship) or to publish before payment (which may remove the incentive to pay). Institutions, for their part, struggle to see their own “APCs in the wild,” scattered across departmental credit cards and grant lines, and to link those expenditures to funders. The pressure to aggregate, or bundle, is a combination of the drive to reduce costs and bad debt, while also increasing auditability and improving reporting.

Aggregation takes two primary forms — the transformative agreement and the pure publish agreement. 

The transformative agreement seeks to shift a library’s or consortium’s contracted payment to a publisher away from subscription-based reading and toward open access publishing. As I described in my primer on transformative agreements, these come in two flavors — read-and-publish, in which payment for reading and payment for publishing are bundled into a single contract, and publish-and-read, in which the publisher is paid only for publishing and reading is included at no additional cost. Some argue the distinction is immaterial, but I think it matters, particularly in how consortia allocate costs. These agreements are undertaken especially by the largest publishers and research libraries or library consortia, because they can justify the investment in negotiation and implementation demanded from both parties. 

The pure publish agreement serves the publishers for whom a transformative agreement is not available — for those publications without a subscription base to transform. Here the contract funds an institution’s authors to publish in a publisher’s fully open access journals, whether as a capped or uncapped commitment or as discounted per-article payments. The pure publish agreement is, in effect, the fully open access cousin of the transformative agreement, and it is often paired with a transformative agreement for publishers that have both hybrid and fully OA journals.

Collective Action

The second category groups models that pursue open access publishing through collective action. What distinguishes these models from the transactional models is that no specific author is supported for open access publication; instead, a threshold of collective participation determines whether all content opens in a publication, regardless of who authors it.

Subscribe to Open is the most prominent model in this category and requires an existing subscriber base for a closed or hybrid publication or set of publications. The publisher’s offer to libraries is straightforward: continue subscribing, and if enough subscribers do so, the content opens to all; if they do not, it remains closed. Because the subscribing institution retains access either way, the model is fairly described as a no-risk opt-in. Its virtue is that it demands almost no change to existing subscriber workflows. Its limitation is that neither authors nor their institutions have agency over whether any particular article ends up open, and — as I have noted before — it is best matched to highly valued content, where the risk of losing access is what sustains the subscription in the first place.

Crowdfunding — a timed pledge cycle in which libraries and others commit funds toward opening a publication — is the other approach to collective action. The Open Access Community Investment Program (LYRASIS) is an example of this approach. 

Subvention

The third category covers business models in which the money to publish comes from sources other than payments tied, however loosely, to researchers or libraries and their budgets. In the membership model, members receive other benefits, such as governance participation, rather than access to content, which is, by definition, available to everyone. The Open Library of the Humanities is an example of this model. Sponsorship approaches fund publications from an organization’s other revenues, as with US federal government serials, or College & Research Libraries, the flagship journal of the Association of College and Research Libraries. A related model is volunteerism, in which scholars contribute their labor and use the resources that institutions typically make available (e.g., web servers and word-processing software) to publish a journal. The distinction between sponsorship and volunteerism is whether the organization itself is the publisher or if it makes resources available but does not direct their use for publishing. And finally, there is philanthropy, where monetary gifts support open access publishing, such as the anonymous donor who supports the Communications of the American Mathematical Society.

How Categories Help

Understanding the category to which a model belongs helps illuminate whether it can satisfy a given funder open access mandate and whether it fits a given institution’s circumstances. A funder requiring immediate open access with a CC BY license on the version of record is well served by most transactional models, but a Subscribe to Open offering only helps authors comply if the threshold is met, and thus the author (and their institution) may need for there to be a transactional model backstop for cases where the threshold is not met but their funder requires open access. For an institution, the question of which models support open access publishing for its own authors may be most salient, particularly if the funding used is part of the overhead on federal grants or is charged directly in grant budgets. Given these considerations, it is not surprising that the uptake of transformative agreements and the resulting publication of open access articles is growing rapidly. 

That growth is itself part of why three categories now suffice where perhaps more were needed a few years ago. The period of intensive experimentation produced many creatively named offerings but comparatively few distinct answers to the question of who pays and how. Rebate structures, membership-linked agreements, and other innovations have largely receded, and transformative and pure publish agreements have converged on a recognizable form. Publishers will continue to name their offerings distinctively, and there may be subtle differences. But the underlying mechanisms are fewer than the names would suggest, and knowing which of the three categories is in development or on offer is usually enough to determine which questions to ask next.

Acknowledgment: I am grateful to Tasha Mellins-Cohen for her feedback on this essay.

AI Disclosures: Claude was used as a conversational interlocutor in developing the essay. Grammarly was used to polish the writing.