Skip to main content
Illustration for Findable, Not Found: Backend Relevance-Model Capability and the Retrieval Gap in the Library Discovery System

Findable, Not Found

Backend Relevance-Model Capability and the Retrieval Gap in the Library Discovery System

A 20-month, four-release deployment tracking a university library discovery portal's offline relevance quality (nDCG@10, rising from 0.54 to 0.82) against session-level retrieval success, which held within a 40-43% band throughout (r = 0.14, n.s.). A relevance-explanation badge intended to help patrons exploit the improved ranking changed session success by 0.6 percentage points (n.s.), and a nine-month extended follow-up found success plateauing rather than converging toward the offline capability gain.

Download PDF

Cite as

@misc{slop_lwklbz,
  author       = {Verity Marris and Petra Umbile},
  title        = {Findable, Not Found: Backend Relevance-Model Capability and the Retrieval Gap in the Library Discovery System},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.lwklbz},
  url          = {https://slop.university/outputs/slop-paper-testing-whether-upgrading-lwklbz/},
  version      = {1.0},
  note         = {Research paper},
}
First page of Findable, Not Found: Backend Relevance-Model Capability and the Retrieval Gap in the Library Discovery System