Digital migration history increasingly relies on keyword-based searches within large-scale digitised archives, yet a fundamental question remains underexamined: under what technical conditions do migrants actually become searchable subjects, and how does platform design mediate that visibility? This paper reports work in progress on a platform-search methodology for measuring romanisation-induced invisibility in European digital archives.The study originates from a concrete archival difficulty. A Chinese migrant recorded as Wang in the Archives nationales cannot be retrieved through keyword search because the same individual appears under Ouang, Vang, and Won across records from different administrative offices. The record is present; it is simply unsearchable. To measure this systematically, the study searches for the surname 王 and its romanisation variants across six discovery platforms in France, the Netherlands, the United Kingdom, Belgium, and Norway, cross-validating returns against each platform's catalogue holdings.Three findings structure the argument. Platform returns overcount: FranceArchives returns 637 results for Wang against a verified count of 163, an inflation of 274%. Platform returns undercount: Norway's Digitalarkivet returns 5,599 results of which only one is a confirmed Chinese individual. And platform returns are nationally bounded: Ouang appears exclusively on French platforms, invisible to any non-French search. Together, these results suggest that discovery systems retrieve administrative romanisation conventions rather than migrants. This reframes search auditing as a replicable methodological tool for digital migration history.The paper invites discussion on whether computational name-matching or archival metadata enrichment offers a viable path toward recovering what keyword search cannot reach.