Opinion: Why Researcher Profiles Need One Entry Point, Not Many

A recent guest post on The Scholarly Kitchen, "Improving OpenAlex Data: Why Publishers Should Help Authors Improve the Scholarly Record" by Paul Tuinenburg, calls on publishers to encourage authors to claim and curate their OpenAlex profiles at acceptance and publication. The aim is to improve author disambiguation in a dataset that many downstream tools now rely on.

I read the piece with great interest since it describes a problem we deal with every day to support our BIP! Services. I fully agree with the diagnosis. I disagree, however, with two points in the proposed remedy. First, I believe that asking researchers to curate yet another profile creates a parallel structure, when the effort should instead go into strengthening the infrastructure that already exists for exactly this purpose: ORCID. To be fair, the post also encourages authors to link their ORCID iD, but as one step within an OpenAlex-centred workflow, rather than treating ORCID as the place where curation should happen. Second, OpenAlex is not the only open scholarly graph, so a remedy focused on one graph is incomplete.

Following the spirit of transparency of the original post, I should note that the services my team develops (BIP! Services) are built primarily on top of the OpenAIRE Graph, though we also use data from OpenAlex and other open sources. I should also disclose that I served as OpenAIRE’s Development and Operations Director from June 2022 to November 2024. With that on the table, here is my perspective, which reflects my personal views. 

Where I agree: errors propagate

The central observation of the post is right: when a widely used scholarly data source contains author disambiguation errors, those errors spread silently to every service built on it.

We see this first-hand. Both BIP! Scholar, our researcher profile service, and our recently released researcher-level BIP! Badges aggregate information across a researcher's full body of work. If two researchers are merged, or one researcher is split into three profiles, every aggregated number shown for them is wrong, regardless of how carefully the indicators themselves are calculated.

I also agree that researchers themselves are best placed to correct these errors. Algorithms can go a long way, but the person who actually wrote the papers is the best authority on which works are theirs. The real question is where that human effort should go.

A design choice we made with BIP! Scholar

When we started designing BIP! Scholar, we faced this same question. A researcher profiling platform needs to identify which works belong to each researcher, and building a dedicated claiming and curation workflow can be tempting. We deliberately decided against adding a second data entry point for researchers. Instead, BIP! Scholar profiles draw on the researcher's ORCID record, enriched with indicators, topics, access information, and other metadata from open scholarly data infrastructures.

We made that choice for the following reasons:

• ORCID was created for exactly this purpose. It is a dedicated infrastructure for researchers to assert which works are theirs.

• ORCID is a non-profit, member-governed organisation. Its Board is elected by its member organisations, with a majority of non-profit representatives, so effort invested there is not tied to any single aggregator or service.

• ORCID is well established and well connected. Crossref and DataCite can automatically push works to ORCID records, many publishers collect ORCID iDs at submission, and many institutional systems can synchronise with it. 

ORCID coverage is not perfect: some researchers do not have an ORCID iD, and many records are incomplete. But this is a reason to improve ORCID records, not to build a parallel system. When researchers want a complete profile in BIP! Scholar, we encourage them to complete and update their ORCID records. That benefits every service built on ORCID, not just ours.

Why a parallel curation layer is the wrong answer

The original post anticipates an obvious objection: researchers already maintain profiles on Google Scholar, ResearchGate, ORCID, institutional systems, and elsewhere. Is it reasonable to ask them to maintain yet another one? The post argues that OpenAlex justifies this additional effort by combining a commitment to building and maintaining a comprehensive scholarly record with full openness of both its data and its code. These are undoubtedly valuable qualities, but they are not unique to OpenAlex. The OpenAIRE Graph, for example, is also a large-scale initiative to build a comprehensive scholarly record, with openly licensed data and open-source code; ORCID, for its part, is also open and community-governed. In my view, therefore, these qualities do not resolve the underlying objection. If anything, they point towards a different approach: investing researchers' effort in ways that strengthen and benefit all these infrastructures collectively, rather than asking them to contribute separately to each one.

The scarce resource here is researchers’ attention, and parallel profiles divide it. Every new claiming workflow competes for the same few minutes researchers are willing to spend on profile maintenance. The likely result is not one more complete profile, but several partially complete ones, each accurate in different places.

OpenAlex’s claiming workflow relies on the same researchers, with the same limited time, who today leave their ORCID records incomplete. If they do not keep their ORCID records up to date, there is little reason to expect them to maintain their OpenAlex profiles. If ORCID adoption has fallen short, a parallel claiming system may fall short for the same reasons. The remedy is to make maintaining one record easier and more rewarding, not to add more records.

Finally, corrections made in one aggregator stay there. The original post rightly notes that a correction made once in OpenAlex improves every platform built on OpenAlex. But a correction to ORCID, a shared source, improves every platform built on any aggregator that ingests ORCID records. 

I am aware of the counterargument that ORCID records are not always accurate, and that treating them as an authoritative source of truth has led to problems in the past. This is a valid concern. I agree that aggregators will always need their own disambiguation mechanisms, particularly for researchers who do not have an ORCID iD. However, when a researcher corrects an error in an aggregator, that correction should ideally be propagated back to ORCID. In fact, ORCID already provides a framework for this: its best practices for discovery systems recommend that services building their own researcher profiles allow researchers to push new information back to their ORCID records.

OpenAlex is not the only open map

The original post describes OpenAlex as the “map of who published what” used by a growing number of systems. OpenAlex is indeed an impressive and valuable resource, and I warmly welcome the move of institutions away from proprietary databases, in line with the principles advocated by the Barcelona Declaration. But it is not the only open scholarly graph.

As mentioned, the OpenAIRE Graph is another large-scale open scholarly knowledge graph, aggregating metadata from thousands of sources, including repositories, publishers, Crossref, DataCite, and ORCID. Many services, including BIP! Services, are built on it. Other open resources, such as Crossref, DataCite, OpenCitations, and Wikidata, also play important roles in the ecosystem.

Asking publishers to direct their authors to one specific aggregator gives that aggregator a clear advantage over the others. The open scholarly ecosystem benefits from having several independent, open infrastructures that complement and cross-check one another. Having moved away from dependence on a single proprietary database, we should avoid recreating that dependence on a single open one.

This is another reason why the source layer matters. Improving ORCID records, as well as the publisher metadata that incorporates ORCID iDs and persistent affiliation identifiers, benefits every open graph that ingests this information. It could also benefit OpenAlex if it plans to implement a deeper integration with ORCID in the future. 

What I suggest instead

I fully support the idea that publishers are well-placed to nudge authors at the right moment. I would simply point that nudge at the shared source. Specifically:

• Publishers can collect authenticated ORCID iDs for all authors at submission, include them (along with persistent affiliation identifiers) in the deposited metadata, and use acceptance emails to remind authors to keep their ORCID record complete and to allow trusted organisations to update it automatically.

• Aggregators and researcher profile platforms can treat ORCID as the home for researcher-asserted authorship, keep improving their algorithms for everything ORCID does not cover, and send any corrections researchers make on their platforms back to ORCID.

• Researchers can spend their limited profile-maintenance time in one place, their ORCID record, and let every service built on top of it benefit.

This is the approach we follow in BIP! Scholar, and we believe it scales better for everyone involved.

Fix it once, at the source

The original post closes with a line I fully endorse: open knowledge is maintained by all of us. The point is that "all of us" works best when we maintain it in the same place. The answer to incomplete researcher records is not another data entry point for research contributions, but better ORCID records that feed a diverse ecosystem of open scholarly graphs rather than a single one.

Author Note: AI-assisted editing tools were used to help revise wording and structure for clarity.