101 Innovations in Scholarly Communication is a 2015–2016 survey and open dataset, built by Bianca Kramer and Jeroen Bosman of Utrecht University Library, that mapped how more than 20,000 researchers, librarians, publishers and funders were adopting digital tools across six phases of the research workflow. Revisited a decade on, the underlying dataset and taxonomy still frame how institutions read today’s tool landscape — and they show preprints moving from a peripheral publication tactic to a structural stage in the scholarly communication cycle.
101 Innovations in Scholarly Communication is best understood as both a crowdsourced database of research tools and a workflow taxonomy: it organises tool adoption into six stages — discovery, analysis, writing, publication, outreach and assessment — and remains the most cited empirical baseline for how digital tools reshaped academic practice before the current wave of open-science mandates.
- What is the 101 Innovations in Scholarly Communication project?
- What did the six-phase workflow reveal about tool adoption?
- Where do preprints sit in the research cycle now?
- How has the standards and identifier layer changed since 2015?
- Answer-first questions on the survey and its legacy
- What this means for institutions, publishers and developers
What is the 101 Innovations in Scholarly Communication project?
The project began in spring 2015 when Kramer and Bosman launched a survey asking researchers and research-support staff which tools and services they actually used, rather than which they were told to use. The survey ran from May 2015 to February 2016 and drew responses from over 20,000 participants worldwide, making it one of the largest empirical pictures of research-tool adoption ever assembled.
Crucially, the project was never a simple popularity poll. Kramer and Bosman built a crowdsourced resource database of research tools alongside the survey, and published the anonymised dataset openly — a practice that itself demonstrated the open-science values the project was measuring. The output most people recognise is the poster mapping roughly 101 tools by year of launch and workflow stage, first presented at Force2015 and reproduced across library guides and blogs ever since.
What did the six-phase workflow reveal about tool adoption?
The survey’s most durable contribution is its six-phase model of the research workflow: discovery, analysis, writing, publication, outreach and assessment. Each phase captured a different cluster of emerging tools, from literature-alert services in discovery to altmetrics providers in assessment, and the analysis found a median of around 22 different tools in active use per respondent.
That figure mattered because it quantified a problem institutions were only beginning to name: workflow fragmentation. Researchers were not choosing a single platform; they were stitching together a personal stack, often spanning free, subscription and institutionally licensed tools. The published analysis also flagged a specific risk — that publishers such as Elsevier and Springer Nature were acquiring tools across multiple phases, raising the prospect of vendor lock-in across an entire research workflow rather than a single product.
| Workflow phase | 2015 survey emphasis | 2026 landscape signal |
|---|---|---|
| Discovery | Alerts, Google Scholar, library databases | AI-assisted literature search layered onto existing indexes |
| Analysis | Excel, SPSS, early notebooks | Open notebooks and reproducible-analysis pipelines |
| Writing | Word processors, early collaborative editors | Cloud-native collaborative authoring as default |
| Publication | Preprints as an emerging, minority practice | Preprints as a funder-recognised compliance route |
| Outreach | Dropbox used more than GitHub for sharing protocols | Persistent-identifier-linked outputs (DOIs, ORCID) as default |
| Assessment | Early altmetrics alongside citation counts | Contributor-role taxonomies alongside altmetrics |
Where do preprints sit in the research cycle now?
In the 2015 dataset, preprint servers appeared in the publication phase as one option among many, used by a minority of respondents concentrated in physics, mathematics and, increasingly, biology. A decade later, scientific preprints have moved from a niche discovery-to-publication bridge to a structural checkpoint that funders and journals actively reference.
This shift is visible in funder policy rather than survey sentiment alone. cOAlition S’s Plan S rights-retention approach explicitly permits preprint deposit combined with a rights-retention licence as a route to immediate open-access compliance, and journals across the life sciences now routinely accept manuscripts that have already circulated on bioRxiv or medRxiv. The practical effect is that preprints no longer sit only in the “outreach” gap the original survey identified between writing and formal publication — they now function as a recognised, citable stage in their own right within the scholarly communication cycle.
What the 2015 baseline still gets right is the underlying diagnosis: tool adoption in scholarly communication rarely moves as a single wave. Preprints diffused unevenly by discipline then, and adoption of preprint-linked policies remains uneven by discipline and funder now — the survey’s disaggregated view of workflow stages is why that unevenness is still visible rather than averaged away.
How has the standards and identifier layer changed since 2015?
The 2015 survey’s vendor lock-in warning has a direct institutional answer that has matured substantially since the data was collected: the persistent-identifier and standards layer. ORCID identifiers for researchers, DataCite DOIs for datasets, and the Research Organization Registry (ROR) for institutional affiliation now interoperate across discovery, publication and assessment tools in a way that was still emergent in 2015.
Contributor attribution has followed the same trajectory. CASRAI originated the CRediT contributor role taxonomy in 2014, the same year the 101 Innovations survey was being designed; the standard is now stewarded by NISO as ANSI/NISO Z39.104-2022, giving the “assessment” phase of the original workflow model a formal, interoperable vocabulary for who-did-what that the survey’s respondents were, at the time, largely improvising without.
- ORCID: persistent identifiers for individual researchers, reducing name-disambiguation friction across discovery and assessment tools.
- DataCite: DOI registration for datasets and other non-article outputs, extending citability beyond the journal article.
- ROR: standardised institutional identifiers, used increasingly by funders and publishers for affiliation matching.
- ANSI/NISO Z39.104-2022 (CRediT): a controlled taxonomy of contributor roles, now stewarded by NISO.
Answer-first questions on the survey and its legacy
What is the 101 Innovations in Scholarly Communication survey?
It is a 2015–2016 survey and open dataset created by Bianca Kramer and Jeroen Bosman of Utrecht University Library, which asked over 20,000 researchers and research-support staff which digital tools they used across six phases of the research workflow, then published the anonymised results openly.
What are the six phases of the scholarly communication cycle it identified?
The survey defined the cycle as discovery, analysis, writing, publication, outreach and assessment. Each phase groups a distinct set of tools, from literature-alert services and reference managers through to altmetrics and peer-review platforms, giving institutions a structured way to audit workflow tool adoption.
Why did the survey warn about vendor lock-in?
The analysis found large commercial publishers acquiring tools spanning multiple workflow phases, meaning a researcher could end up dependent on a single company’s suite from discovery through assessment. That concentration risk is the reason interoperable, standards-based alternatives such as ORCID, DataCite and ROR have gained institutional priority since.
Are preprints still a minority practice in scholarly communication?
No. Preprints have shifted from a discipline-specific minority practice in 2015 to a funder-recognised compliance route under policies such as cOAlition S’s rights-retention strategy, with routine preprint-to-journal pipelines now standard in fields including biomedicine and physics.
What this means for institutions, publishers and developers
For research administrators and institutional leaders, the practical lesson from re-reading the 101 Innovations dataset in 2026 is that workflow fragmentation has not disappeared — it has been partially re-solved through standards rather than consolidation. Where the 2015 survey found researchers assembling a median of 22 tools with little interoperability, the current landscape increasingly links those tools through shared persistent identifiers and controlled vocabularies rather than through a single vendor suite.
For publishers and developers building on the scholarly communication cycle, the implication is that preprint infrastructure, contributor-role metadata, and organisational identifiers are no longer optional add-ons; they are the connective tissue the original survey’s respondents were missing. The next iteration of tool-landscape research will need to measure not just which tools researchers pick, but how well those tools speak to each other across the discovery-to-assessment pipeline — a question the original taxonomy was designed to surface, and one that remains unresolved a decade on.
Leave a Reply