Resilient Data Futures
EvidenceE-0006draft

191 research data repositories shut down 2012-2023; median age 12; 47% no migration

Exec Summary, §1.2, §3.42026-05-034 out · 0 in

Strecker, Pampel, Schabinger, and Weisweiler 2023 analyzed the re3data directory of research data repositories and identified 191 closures since 2012, with a median operational age of 12 years at closure. 47% of the closures gave no indication of data migration or continued limited access.

The number is not "platforms that became less reliable" or "platforms that raised prices." It is platforms that ceased to exist as preservation services. Of those, nearly half terminated without a successor — the data they held was either lost, fragmented across multiple ad-hoc destinations, or transferred under conditions Strecker et al. could not document.

The 12-year median is the clearest available number for how long Tier 1 hosted infrastructure persists in practice. It is shorter than a researcher's career, shorter than the citation tail of most papers, and shorter than the timescales NIH/NSF preservation requirements implicitly assume. It is the empirical input that establishes Tier 1 as one organizational decision away from the same outcome as Tier 0.

The 47% no-migration figure is the rate at which Tier 1 closures convert to permanent loss rather than transition. The architectural implication: the survival of data on Tier 1 depends on the continued business viability of one provider, and the realized churn rate at the sector level is documented at this magnitude.