53.6% of access-eligible SMB websites in our sample failed the extraction-readiness threshold. That's the finding that surprised me most from a study of 428 business websites across four countries. The paper asks a deceptively simple question: being reachable by a retrieval system is one thing — but is the website actually structured so that the system can extract from it reliably? I documented the protocol, data and limitations here: https://vsourcecode.com/reports/book/measuring-the-citation-gap-research-paper DOI: 10.2139/ssrn.7324978