Digital library projects are easy to start and hard to keep. A grant pays for design, digitization and launch. Then the grant ends, the staff who built the system move on, and the collection enters a long phase in which nobody is quite sure who pays for servers, software updates or metadata fixes. Many projects from the 2000s did not survive that phase. Their URLs now return errors.
This post looks at the funding models that do keep digital libraries running, what each one asks of an institution, and how to choose among them before the money runs out rather than after.
The question asked in 2006
The question is not new. On October 6, 2006, the MetaScholar Initiative hosted the Sustaining Digital Libraries Symposium at Emory's Woodruff Library. The organizers noted that digital libraries had evolved quickly over the previous decade and that sustaining them had become a major concern across many kinds of institutions. The program brought funders, including the then-new NSF Office of Cyberinfrastructure, together with representatives of NDIIPP, the National Science Digital Library and the Digital Library Federation. It also invited groups that ran similar services but were rarely in library conversations, including Amazon.com and the Inter-university Consortium for Political and Social Research.
That mix was the insight. Sustainability is not only a library problem. Commercial services and data archives had already worked out how to fund ongoing operations, and they had lessons to share. The symposium's papers were later gathered in Strategies for Sustaining Digital Libraries, which we summarize in a separate briefing.
A few years later, the Blue Ribbon Task Force on Sustainable Digital Preservation and Access reached a similar conclusion in its 2010 final report. It argued that sustainable access depends on economics as much as technology: someone must see the value, have an incentive to act, and know that the job is theirs. See also our notes on the MetaScholar Initiative and its projects.
Five funding models
Most sustained digital libraries rely on one or more of five models. None is free of tradeoffs.
| Model | How it works | Strength | Weakness |
|---|---|---|---|
| Institutional base budget | A host library or university funds operations as a line item | Stable when the host values the service | Vulnerable to leadership changes and budget cuts |
| Membership or consortium | Participating institutions pay fees or contribute staff | Shares cost and builds a community of owners | Requires governance and steady recruitment |
| Grants | Foundations or agencies fund defined projects | Good for building and innovating | Rarely covers routine operations |
| Earned revenue | Fees for services, licensing, print sales or hosting | Ties income to use | Can conflict with open access goals |
| Endowment | Investment income from a dedicated fund | Long-term and independent | Very hard to raise at sufficient scale |
The MetaScholar projects illustrate several of these in sequence. Early work ran on grants from the Andrew W. Mellon Foundation, IMLS, NEH and others. The MetaArchive preservation network moved from a Library of Congress supported project into a member cooperative. The open access journal Southern Spaces grew up inside the initiative at Emory's library. Each transition was a deliberate choice about who would carry the cost next.
What actually costs money
Sustainability plans often underestimate ongoing costs because the visible work is over. A realistic budget lists the work that never stops.
- Infrastructure. Hosting, storage, backups, domain names and security certificates.
- Software maintenance. Security patches, framework upgrades and the occasional migration when a platform is retired.
- Preservation. Fixity checks, replicated copies and format monitoring for the master files.
- Metadata upkeep. Corrections, new vocabulary terms, broken links and rights updates.
- User support. Answering questions, handling permissions requests and fixing reported errors.
- Governance. Meetings, reporting, renewing agreements and planning for the next phase.
None of these lines is large on its own. Together they form a steady cost that outlasts any single grant. The planning mistake is not that projects ignore these costs. It is that they assume someone else will absorb them.
Why cooperatives deserve a closer look
Of the five models, the membership cooperative is the one most often misunderstood. It is sometimes described as a way to split a bill. It is closer to a way of sharing ownership. Members pay, but they also decide. They set priorities, approve budgets and accept new members. That control is what keeps institutions paying when their own budgets tighten, because leaving means giving up a voice in something they helped build.
The model works best under a few conditions. Members need a shared interest that is concrete, such as a subject area or a type of content. The cooperative needs a small, stable staff or a rotating set of volunteers with clear duties. Fees should be predictable and tied to a published formula, so finance offices can plan. And there must be a written agreement that covers what happens when a member leaves, including how its content is returned.
The MetaArchive partners built a cooperative agreement into their original project plan, alongside the technical network. That sequencing is the lesson. Governance designed during the grant is far easier to adopt than governance invented in a crisis after the grant ends.

Choosing a model early
The best time to decide how a digital library will be sustained is before it launches. By then, a funder may already ask for a sustainability plan. Treat that plan as a working document, not a box to tick.
- Name the long-term owner. Identify the unit that will be responsible once the project team leaves. Get that commitment in writing.
- Estimate the steady-state cost. Price the recurring work above for five years, including one platform migration.
- Match the model to the audience. A service used by many institutions suits a membership model. A local collection suits a base budget.
- Design for a smaller future. Decide what the service would look like if funding fell by half. Static sites, open formats and standard metadata make that version possible.
- Plan the end. If the service closes, where do the files and metadata go? A deposit agreement with a preservation network or repository is cheap insurance.
Smaller and simpler lasts longer
One pattern stands out across the digital libraries that have survived since the 2000s. The ones still online tend to be the simplest. Static pages and plain files outlive custom databases. Standard metadata, such as Dublin Core or MODS exposed through OAI-PMH, outlives local schemas. Content kept in a preservation network, such as one built on LOCKSS, outlives content kept on a single server.
That is not an argument against ambition. It is an argument for separating the durable core of a digital library, its files and its descriptions, from the interface layer that will be rebuilt every few years. Fund the core as permanent infrastructure. Fund the interface as a project. When budgets tighten, the interface can shrink while the collection stays safe and findable.
Sustainability, in the end, is a set of commitments by named people and institutions. Technology can make those commitments cheaper to keep. It cannot make them for you. The organizers of the 2006 symposium understood that, which is why they put funders, librarians, archives and a retailer in the same room. The question they asked still deserves the same broad conversation.
