What a single company-search result does not tell you: 106 Chinese manufacturers searched by registered English name

Search-layer outcomes for every one of the 106 Chinese manufacturers on the NHTSA vPIC trailer/vehicle frame, each submitted once by its registered English name from a mainland egress, with candidate rows recorded in upstream order before any re-ranking. Headline counts: 45 of 106 (42.5%) returned no candidate; 54 of 106 (50.9%) returned exactly one row; 7 of 106 (6.6%) returned more than one. The middle outcome is the one this study is about: zero results is honest about its own failure, several results at least show a choice exists, and one row shows neither. Within the 84 candidate rows returned, 11 (13.1%) were registered in Hong Kong rather than mainland China. Ten of those eleven carried no 18-character Unified Social Credit Code, since Hong Kong identifies companies by an eight-digit Business Registration Number adopted as the Unique Business Identifier on 27 December 2023. A further 6 rows (7.1%) were already deregistered. Among multi-result searches, 12 of the 23 rows beyond the first were branches of the anchor company rather than separate companies. A hypothesis that failed, reported as such: the study was designed to test whether search results are crowded by affiliated non-producing entities (investment, holding, supply-chain, consultancy, laboratory wording). That did not reproduce on this frame — zero of the 23 additional rows carried such wording. The observation that prompted it came from a single large industrial group and did not generalise. Files: a 14-row summary-metrics CSV and a JSON containing the per-input raw records. Method, limits and the landing page: https://currawongweb.com/verify/single-result-confirmation-study/