How a Russell Group library built a CRediT helpdesk in 4 weeks

Profile

Composite institution profile

The challenge

What problem were they trying to solve?

In late 2025 the library received a sudden uptick in researcher queries about CRediT statements: UKRI had just confirmed that R4RI-style narrative CVs would be the standard for fellowship applications across all seven research councils, and several Wellcome and MRC schemes had started accepting structured contribution metadata as part of the application package. Researchers wanted answers to four kinds of question: which CRediT roles apply to me on this paper, how do I handle a paper where I disagree with my co-authors' role attribution, how does this map onto my R4RI module 1 narrative, and what do I do if my publisher captured CRediT but the structured data did not flow into Pure. The library had two librarians with relevant expertise but no formal service offering, no FAQ, and no triage protocol. A senior research-services lead set a four-week deadline to stand up a service before the autumn fellowship application deadlines arrived in volume.

The approach

How they implemented it

The library used a deliberately compressed approach: rather than building a comprehensive guide first, the team built the helpdesk first and let the FAQ grow from real queries. Week one focused on training the four library liaison librarians who would cover the service — using the CASRAI /for-institutions/training kit and the CRediT taxonomy reference at /credit/roles as the canonical sources, plus a half-day workshop on UKRI R4RI structure run by the library's research-impact lead. Week two stood up the helpdesk infrastructure: a dedicated email alias, a triage spreadsheet, a published service-level commitment of two working days for first response, and a one-page landing leaflet that pointed researchers at /credit/for-authors, /for-authors/narrative-cv, and the library's own LibGuide. Week three ran a soft launch with three pilot departments (biomedical, history, physics), capturing every query verbatim and using the patterns to draft the FAQ. Week four opened the service institution-wide and published the FAQ. The team made one architectural decision early that paid off repeatedly: the helpdesk would never tell a researcher which CRediT role they had played on a paper. The helpdesk would explain the 14 roles, point at the canonical NISO definitions, and help the researcher think through the assignment themselves — but the assignment was always the researcher's call, never the library's. This kept the library out of authorship disputes and made the service much faster to staff.

Timeline

Rollout phases

1. Week 1

   Liaison librarian training

   Four librarians completed the CASRAI training kit, a half-day R4RI workshop, and a "role disambiguation" exercise using twelve real (anonymised) papers from the previous REF cycle.

2. Week 2

   Helpdesk infrastructure

   Email alias, triage spreadsheet, two-working-day SLA, landing leaflet printed and distributed to faculty offices. Pure-side guidance added to internal LibGuide.

3. Week 3

   Soft launch (3 departments)

   Pilot ran with biomedical sciences, history, and physics. 47 queries received in 10 working days. Patterns surfaced: 60% role-mapping queries, 25% Pure integration, 10% R4RI, 5% disputes.

4. Week 4

   Institution-wide launch + FAQ publication

   Service opened to all 3,500 researchers. FAQ published on intranet with 18 entries derived from the soft-launch queries. First all-faculty newsletter mention.

5. Months 2–6

   Iteration + ARMA conference talk

   Helpdesk averaged 30–40 queries/week through the autumn application cycle. The library research-impact lead presented the model at an ARMA practitioner event in spring, prompting three peer Russell Group libraries to adopt a similar approach.

Outcomes

Illustrative outcomes

Every metric below is illustrative — synthesised from observed patterns across multiple adoption journeys, not attributed to a single real institution.

Lessons learned

What they would tell the next institution

• 01Building the helpdesk before the FAQ was counter-intuitive but right. The FAQ was much better when it was written from real queries rather than from the team's assumptions about what researchers would ask.
• 02The "we don't adjudicate role assignments" boundary was load-bearing. Without it, the service would have been pulled into authorship disputes that have no library-side resolution.
• 03Pure-side data fixes are the most expensive queries. Roughly one in five queries required a research-systems team intervention to correct ingested metadata; the library cannot resolve these but can triage them quickly.
• 04Library-led training scales further than expected. Cross-training eight additional liaison librarians in months 2–6 absorbed the autumn application surge without adding headcount.
• 05External validation matters. The ARMA conference talk gave the service legitimacy with university senior leadership in a way that internal metrics could not.

What's next

Planned next steps

The library is now piloting a CRediT-aware export from Pure into the R4RI module 1 narrative template — researchers can pre-populate their fellowship-application narrative with contribution patterns from their last five years of outputs. The library is also working with three peer Russell Group libraries to publish a joint LibGuide under CC-BY 4.0, derived from the CASRAI training kit, that any UK library can adopt as a starting point.

Q&A with the composite project lead

Composite project-lead Q&A

The questions and answers below are composite — synthesised from interview patterns across multiple real project leads. They are not attributed to a specific real person.

How did you handle authorship disputes?

We did not. The helpdesk's role was always to explain the taxonomy, never to adjudicate. When a researcher told us their co-authors disagreed with their CRediT assignment, we pointed them at the published NISO definitions, the relevant /credit/roles pages, and the institutional authorship policy — but the conversation between co-authors had to happen between co-authors. That boundary kept us out of trouble and the helpdesk staff sane.

What was the most common query?

"Does the Software role apply to me if I only wrote three Jupyter notebooks for the analysis?" — and variations on it. Roughly one in seven queries was a researcher trying to decide whether their level of contribution to a non-obvious role was significant enough to claim. The CRediT role definitions deliberately leave this judgement to authors, so our answer was always "here are the canonical NISO criteria; you are the best person to judge."

Did senior researchers use the service or just early-career ones?

Both, but for different reasons. Early-career researchers asked about role assignment on their own outputs; senior researchers asked about how to ensure their group's outputs were being captured correctly in Pure. The split was roughly 60/40 by query volume, and we ended up training a dedicated "senior researcher liaison" in month four to handle the second group, which has different needs.

How did you fund the service?

It was a re-prioritisation within the existing research-services library budget, not a new line item. The four-week build was deliberately cheap, and the ongoing cost was eight librarians spending roughly 15% of their time on helpdesk queries — well within existing duties. If the helpdesk had required new headcount we would not have got the four-week launch.

Cited CASRAI resources

Internal CASRAI resources referenced

• ›Dictionary: Narrative CV
• ›Dictionary: CRIS
• ›Dictionary: DORA
• ›CRediT role: Software
• ›CRediT role: Data Curation
• ›CRediT for authors
• ›Narrative CV — guidance for authors
• ›UKRI CRediT statement guidance
• ›Library guidance hub
• ›Workshop materials (CC-BY 4.0)
• ›Implement: Elsevier Pure
• ›Implementation checklist: responsible assessment