5 minute read.Why PID strategies need more than PIDs: our first position paper
PID strategies are being written around the world right now, and the decisions being made will shape the scholarly record for decades. After 25 years running open scholarly infrastructure—now on behalf of 25,000 members in 167 countries—Crossref has an informed perspective on what those decisions should ideally rest on. Today we’re setting it out in our’s first position paper: Persistent identifiers in research infrastructure policy: the need for a holistic approach. You can read it online or download the PDF; it’s a 16-minute read.
Why now? The world looks very different from when Crossref was set up, and what we steward now reaches well beyond our membership: open metadata flows into national assessment systems, discovery services, integrity tools, and the AI systems learning from the scholarly record. Increasingly, decisions about research infrastructure are made by people who have never registered a record with us. Setting out our positions in public on key issues—to be referenced for years to come—seems a useful way to take that wider responsibility seriously.
The central argument of this paper is that decisions about PIDs are often framed as a choice of identifier, whereas what is really being chosen is infrastructure. Identifiers alone—however well-designed—cannot deliver the connected, open record that researchers, institutions, funders, publishers, and policymakers depend on. Effective research infrastructure rests on three interdependent elements: open, persistent identifiers; rich, open, and linked metadata; and the sustainable governance and operation of the organisations involved.
We urge those developing PID strategies and open science policies to evaluate all three together, looking beyond the identifier itself to the metadata, services, interoperability, and long-term sustainability that provide the real-world value they seek.
One thing the paper is clear about: not all DOIs are equivalent, and there is no centralised DOI metadata store—meaning a policy that specifies ‘use DOIs’ without accompanying requirements for metadata, services, and governance, can leave important questions open. We therefore recommend that policies spell out the real-world outcomes needed, not just the identifier type.
The paper is also explicit about what persistence actually takes. Real persistence is active: a social contract in which members take on the long-term stewardship of their own records. And it has to be paid for; infrastructure serving the record in perpetuity should demonstrate financial resilience. The paper shows how Crossref delivers both: our members actively maintain their records, and our membership model has kept us financially resilient for the long term while also paying it forward and supporting others.
The paper also addresses digital sovereignty, a live concern in many national strategies. Our answer is that sovereignty is safeguarded by design, not just by trust. National institutions are not customers of this infrastructure; they are members and governors of it, they remain the authoritative source of their own metadata, and their investment strengthens the shared record rather than fragmenting it.
Who is this paper for?
Policy strategists get the macro view: why identifiers, metadata, and governance need to be evaluated together, and how infrastructure choices shape the long-term success of a research strategy.
Decision-makers get criteria to require before a decision is made: how to evaluate PID infrastructure as a whole, including interoperability, sustainability, and governance.
Practitioners get a view of what good implementation looks like before committing: practical requirements for the metadata and services that make an identifier system useful in daily operation.
Key positions
PIDs are a means to an end. The value of an identifier system lies in the richness of its associated metadata, the scalability of the services it enables, and the needs it meets for the communities it serves.
PIDs can be complementary rather than competitive. Research uses many identifier types; they coexist productively, each serving the community and use case for which they were designed.
Not all DOIs are equivalent. Different DOI Registration Agencies offer fundamentally different services, metadata, and governance. Concrete requirements and expected outcomes for each entity being identified—not just identifier type—are therefore essential in any policy or implementation.
Interoperability is critical. Implementing identifier systems without evaluating their interoperability undermines the goal of a connected research record. National strategies must ensure that proposed PID infrastructure connects to the broader global research ecosystem.
Sustainability and governance matter as much as technical standards. Community governance and demonstrated financial resilience should be taken into account in any assessment of infrastructure organisations.
Open infrastructure—not PID adoption alone—is the foundation of a connected, open research record. Strategies should measure success based on the openness, connectivity, and reach of the surrounding infrastructure and its services.
Join the conversation
The paper includes worked examples from the UK, Portugal, Canada, and South Africa—policies that specify needs and outcomes rather than prescribing identifier systems, which is what we think good PID strategy looks like in practice. If you’re developing or reviewing research infrastructure policy and want to talk it through, we’d welcome that. Start a discussion on our community forum, or email us via feedback@crossref.org anytime. This paper is also the newest addition to our publications area, which gathers 25 years of Crossref reports, guides, and perspectives in one place.
Cite the paper as: Crossref (2026). Persistent identifiers in research infrastructure policy: the need for a holistic approach. Crossref Position Paper. https://doi-org.turing.library.northwestern.edu/10.13003/q4vu-l2mw
