A few years ago, I was invited to give a keynote on ‘any area’ of my research. Surprisingly, for me, it was for a psychology conference – neither my work on scholarly communication infrastructures nor on maritime history seemed an ideal fit. It turned out that the organizers were interested in another Torsten Reimer, an actual psychologist. They had traced his publications back to me, despite our profiles not overlapping in the slightest.
This anecdote highlights a problem in the scholarly communication system: names do not reliably identify individuals. Even if there was only one Torsten Reimer, spelling mistakes, varying use of initials or future name changes lead to confusion, as do changes in institutional affiliation. This is a problem for individual authors as much as for research organizations, funders, publishers and others who need to link authors and publications across systems and data sources.
The ORCID1 identifier (iD) offers a solution. It gives authors a unique iD that stays with them throughout their career, no matter how often they change name or institution (or whether they have one at all). Publishers can add the iD to the metadata of outputs, linking it to the author. ORCID is free for individuals: over 1.6m are registered globally. Institutional members pay an annual fee and get access to features such as a members’ application programme interface (API). The member contributions help to make ORCID sustainable as a community owned, not-for-profit organization.
Imperial College London2 was amongst the first UK universities to become a member of ORCID. In early 2014, the Provost’s Board approved a proposal to join ORCID and to create iDs for academic and research staff. A few months later, the College joined with seven UK universities implementing ORCID in a pilot supported by Jisc and the Association of Research Managers and Administrators (ARMA).3 This case study describes Imperial’s ORCID implementation project4 and sets it in the wider ORCID context.
Why engage with ORCID?
ORCID addresses a real problem: reliably linking authors to scholarly outputs and helping them to receive credit for their work. The unique iD also allows institutions to automate information exchange with other organizations – especially when combined with institutional and funder identifiers – thereby increasing data quality. This saves academics time and institutions money.
Ideally, authors should only have to provide information on an output once. On submission, authors share their ORCID iD and other relevant information with the publisher. On acceptance, publishers embed this information in the output’s metadata. The metadata then travels automatically through relevant systems: publisher, ORCID, research institution and funder. The following examples show how ORCID can support different workflows such as funder reporting, publication of research data and open access (OA).
To apply for a grant, academics usually submit a CV with a list of publications. If successful, they will have to report on project outputs. Researchers have to spend time curating publication lists, and must then copy, paste and reformat them to fit funder templates. Instead, ORCID’s registry allows funder systems to ingest information automatically, with the author’s permission. Alternatively, institutions can use the iD to match data and supply information across systems. Assuming ORCID iD and funder information are embedded in an output’s metadata, funders could even get publication information directly from sources such as Scopus or Web of Science.
ORCID is not limited to journal articles. The principles can be applied to any type of output – data sets, code, theses, patents – and even records such as employment history. Universities using a current research information system (CRIS) already receive publication data, but the feeds are not complete and academics are still required to claim or reject articles manually as the systems cannot always accurately match author and output. No such general workflow exists for research data, although solutions for specific systems are emerging. DataCite and ORCID are working on tools to increase the discoverability of data sets and to link DataCite digital object identifiers (DOIs) to ORCID iDs.5
ORCID can also help with open access. In the UK, the Research Excellence Framework (REF), a regular review of scholarly outputs, determines funding for universities. To be eligible for the next REF, accepted manuscripts of journal articles and conference proceedings must be deposited in an OA repository within three months of acceptance for publication.6 Even where papers are published very soon after acceptance, there is no guarantee that institutional systems will receive notifications of publication before the deadline. This means that all authors (including co-authors who themselves get no publication notification from the journal) have to record metadata and upload manuscripts manually. Later, they must remember to match the manual record to the published information. ORCID can automate this workflow. On acceptance, the publisher mints a DOI and links it to the iDs of the authors. Institutional systems ingest the metadata from sources like CrossRef, remind the authors to deposit, and update the metadata automatically when the paper is published, via the DOI.7 Using a system like Jisc’s Publications Router8, publishers could even provide the manuscript to universities. Not only would such a solution increase data quality and ensure open access to outputs, it would also generate substantial savings. In the case of Imperial College, with over 10,000 journal and conference publications annually, just one minute saved per output equals roughly 0.1 FTE of staff effort!
As these examples show, ORCID provides more than ‘just’ a link from an output to an author. It has the potential to improve the global scholarly communications system significantly for the benefit of authors, publishers, funders, research organizations and readers. This is why Imperial College chose actively to engage with ORCID.
The Imperial College ORCID project
At Imperial, ORCID was implemented through the framework of the College’s open access project. This ensured early engagement from senior colleagues, including the Associate Provost chairing the OA steering group, the College Secretary, the Director of Library Services, the Director of the Research Office and representatives of the academic faculties. Following discussions within the group, a proposal for the College to join ORCID was approved by the Provost’s Board.
To oversee the roll-out, an implementation group was set up with representatives from ICT, the Library, the Research Office and the academic community, overseen by the Vice Dean of the Faculty of Natural Sciences and managed by the Scholarly Communications Officer. The group consulted Legal Services and HR on legal and ethical aspects, and the Communications and Public Affairs team and the Provost’s office on communications planning.
Two approaches to ORCID
There are two distinct approaches to institutional engagement with ORCID:
- Encourage staff/students to register
- Create iDs via membership API.
Both approaches have advantages and disadvantages. It is important to consider which fits best with the institutional culture.
Encouraging staff to register with ORCID is the more common approach. It does not require an institution to implement any technical infrastructure or to become a member of ORCID, and it makes it clear that the iD belongs to the academic. There is no risk of creating an iD for someone who does not want one or who may already have registered. However, it does require academics actively to engage, and to enter information manually, necessitating greater effort and risking errors. A recent check revealed that the ORCID registry contains at least 26 manually entered variants of the name Imperial College London. ORCID treats these as separate organizations.
Creating accounts on behalf of staff allows an institution to ensure that everyone has an iD with relevant information pre-populated, including the correct institutional affiliation and, where available, a curated list of publications. It also ensures that there is an institutional record of the match between internal and external identifier. However, there is a risk of creating duplicate records or creating iDs that will never be used. There are also issues of ownership, consent and privacy. A new ORCID workflow, ‘create on demand’, offers a solution to some of these problems (see below).
Imperial College opted to create iDs on behalf of staff, reflecting an institutional culture that values automation of data-related tasks. ORCID removes a manual step from the process of populating staff records in the College’s publications management system, Symplectic Elements: outputs with an ORCID iD in their metadata can be assigned to researchers automatically. Another reason for creating iDs was that the academics involved in the project saw the creation of pre-populated ORCID records as a time-saving service. There was a strong feeling that ORCID is a ‘no-brainer’ and should be part of a professional academic identity.
Bulk creation of iDs was discussed with HR and Legal Services to address potential privacy concerns. Creating an iD requires the organization to share personal data – at minimum name and e-mail address – with ORCID. The College recognized that not everyone wishes to have a public profile, and academics who had opted not to be on the public College directory were automatically excluded from the bulk creation. For other staff an opt-out was set up. The project decided only to share with ORCID what was already public and that academics would normally want to be widely known: name, institutional affiliation and publications. In the ORCID bulk creation process, profiles are automatically made public after ten days, even when researchers do not interact with them. To address this, profile information was set to private (apart from ORCID number and name as these fields are always public), but academics were encouraged to make information publicly visible.
Early on in the project, Symplectic Elements was identified as the best system for storing the iDs internally. Academics can easily link their iD to their Elements profile, enabling the system to associate publications with an ORCID iD with a researcher’s institutional profile. The roll-out took place over two months. Communication in the first month focused on awareness raising, publicising the opt-out and encouraging staff with an existing iD to add it to Elements, to avoid creating a duplicate. After this initial communication, ORCID iDs were created and staff encouraged to claim their iD and link it to Elements. (See Table 1.)
|06/11/14||ORCID web pages go live
ORCID support in Symplectic Elements goes live
E-mail from the Provost to all staff
|14/11/14||Follow-on e-mail from ORCID project
Supporting communications: staff briefing, info screens, etc.
|20/11/14||Reminder about ORCID distributed via Heads of Departments|
|27/11/14||Final day to opt out or add an existing iD to Elements
Technical preparation starts
|03/12/14||E-mail to staff informing them that ORCID creation is imminent
ORCID iD creation process
ORCID claim e-mail sent from ORCID
|11/12/14||ORCID identified 325 staff who already had an iD but did not link it to Elements before 03/12; as a result no iDs were created for them. The project e-mailed these colleagues, encouraging them to add their iD to Elements.|
|08/01/15||E-mail to staff who had not linked their iD to their Elements profile|
The communications plan, developed in consultation with ORCID, featured:
- web pages with information on ORCID, the project and the opt-out facility
- mention of the initiative in the online staff briefing
- ORCID featured in presentations to departments
- ORCID featured on College info screens
- an e-mail campaign.
Technical preparation involved setting up the script to create the iDs, compiling the list of staff, extracting publications metadata from Symplectic Elements and matching data to fields in the ORCID registry. College publications data is reasonably comprehensive. The XML-encoded data sent to ORCID included over 240,000 academic works relating to 3,226 iDs generated for College authors. While we do not know how many academics are using their iD, we can track the number of iDs linked to Symplectic Elements. Actively linking the iD to a College system shows that the academic understands how ORCID works, and is happy for the College to engage with them through their ORCID iD. (See Table 2.)
|Project achievements in numbers||No|
|Overall number of staff included initially||4,347|
|Staff excluded (those not listed in the College’s public staff directory)||332|
|Staff opting out through the online form||25|
|Staff who added iDs to Symplectic Elements before the roll-out||439|
|Staff with existing iDs, as identified through ORCID de-duplication||325|
|New staff iDs created||3,226|
|Metadata on publications (‘works’) added to the ORCID registry||>240K|
|Staff asking for their newly created iD to be deleted (usually because they already had one that was missed by the de-duplication)||7|
|Staff iDs linked to Elements as of 19/01/15 (project end date)||1,155|
|Staff iDs linked to Elements as of 28/09/2015||1,406|
|College-created iDs claimed as of 20/10/2015 (~65% of the 3,226 created)||2,088|
Academic response, success factors and lessons learned
Overall, the academic feedback has been very positive. During the first stage of the project only 25 colleagues decided to opt out; those who gave reasons stated that they were close to retirement or were about to leave the College. Several colleagues who opted out had an iD already and missed the message that adding it to their Symplectic profile would avoid the creation of a duplicate. Feedback from presentations to academic audiences was very positive. Referring to the opt-out, one Head of Department commented, ‘I can’t conceive why anyone would want to opt out of something that sounds so useful.’
Two academics raised privacy concerns, and were satisfied with the response from the project. Seven colleagues asked to have their newly created iD deleted; five of them already had an iD (but had not shared it previously, and were not identified by ORCID de-duplication), one colleague was about to retire and another academic did not want to engage with ORCID.
The feedback and uptake should be seen as a success and a sign that academics understand the benefits. Several colleagues asked when more features could be expected (such as two-way synchronization between ORCID and Symplectic) and when funders’ systems would use ORCID to automate research outcomes reporting. Others reported that they had begun to share their iD with publishers. Researchers are looking forward to full ORCID support from publishers, system vendors and funders.
Imperial’s experience, and that of other institutions, suggests six success factors in ORCID implementation:
- early engagement with all stakeholders
- consideration of institutional culture
- co-operation across central services
- academic involvement in the project
- support from senior colleagues
- clear, focused communication.
At Imperial, the framework of the OA project was invaluable for delivering ORCID. It ensured senior buy-in and input from academics across the College, and provided a well-established collaboration between ICT, Library and Research Office. Consulting HR and Legal Services early on ensured there were no delays over concerns around privacy and consent.
Whichever approach to ORCID you choose, clear communication is essential to ensure academic engagement. Target your communication, focus on benefits to the academics and make it clear what you want them to do, and how to do it. Face-to-face meetings and presentations are effective to raise awareness of ORCID, but depending on the size of the university such an approach may not always be feasible. Communication via e-mail scales better, but e-mails are easily missed. It seems obvious, but short e-mails, with bullet action points, really do make for better uptake. The more channels of communication you use, the more likely that you will reach the audience – as long as you avoid being seen as spamming them.
If Imperial were to undertake this project again, we would probably keep the approach, with one key difference: instead of ‘bulk’ creating iDs, we would choose to ‘create on demand’.9 With create on demand, a university still generates iDs for staff, but only after an academic clicks a personalized link giving their consent. This makes an opt-out unnecessary, and ensures that iDs are only generated for those ready to engage with ORCID. We hope to see such an approach implemented in CRIS systems, so that universities will not have to set up their own infrastructure. Symplectic Elements, for example, allows academics to create an iD, but it does not currently populate the ORCID registry with information such as publications or employment details.
Ongoing ORCID engagement at Imperial College
Imperial’s formal ORCID project has concluded, but ORCID engagement continues. Requests to add ORCID support to the College repository, Spiral, and to the staff web pages have been added to ICT’s development queue. The College continues to work towards the goal that all staff involved in research use ORCID. New starters are encouraged to create an iD via Symplectic Elements. ORCID is still featured in presentations and electronic communication, including in response to external requirements, such as the Wellcome Trust’s ORCID mandate10 for new grant applicants.
Of the 3,226 iDs created 2,088 have so far been claimed by College staff – or about 65%. Currently, 1,406 iDs are linked to Symplectic Elements. This seems only a modest increase over the 1,155 in January 2015, but some 700 academics have joined and left the College since the roll-out. The vast majority of new staff were not aware of ORCID, whereas those who left included ORCID users. Universities will have to continue to engage with staff on ORCID.
ORCID uptake also depends on further system integrations that deliver real benefits for academics. Imperial College is working with partners to increase ORCID support and to make new features available. On 28 September 2015, we hosted a meeting with over 50 UK universities, ORCID, Jisc, platform vendors and funders to share best practice, discuss community requirements and launch Jisc’s UK ORCID membership consortium. Requirements from the community include full ORCID integration into the systems and workflows of funders and publishers as well as enhanced ORCID support in CRIS and repository platforms – for example, prompting users to create iDs and pre-populating ORCID accounts with institutional affiliation and other information.11 Ease of use and automation of data entry tasks were seen as key factors for academics uptake, and we are looking forward to working with the ORCID community on these and other issues.
The author has declared that he is currently a member of ORCID’s communications steering group.