This qualitative research explores how and why unintended decoupling emerges in organizations under institutional pressure. I investigated five cases of mid-size Japanese-listed B2B manufacturing companies in their adoption and implementation of global talent management (GTM) at their headquarters during 2004-2015, through an interview-based comparative case study. Three firms had slow implementation and symptoms of unintended decoupling, and the other two had fast implementation and no characteristics of unintended decoupling. The case choice provided an exemplar context in which organizations in a field faced institutional pressure to adopt a new practice (GTM) over the traditional practice (Japanese HRM) which contradicts with the new in many aspects. The study provided three findings. First, there was a lack of consensus among internal stakeholders (the top management and GTM task force team versus business units and HR in overseas offices) about whether GTM was relevant to their current operations and would bring practical returns, when unintended decoupling was observed. In those organizations, GTM task force team could not resolve the practical value dispute and thus failed to persuade business units and overseas HR to accept its institutional legitimacy. Second, the emergence of unintended decoupling correlated with the nature of the organization’s adoption decision making process. Unintended decoupling emerged in cases in which GTM adoption decision was made largely due to institutional considerations. Those organizations, without strong sources of internal legitimacy such as founder family CEO and shared problem solving protocols, proactively searched for sources of legitimacy outside themselves. Conversely, non-decoupling case firms adopted GTM purely due to practical considerations. These firms had founder family CEO and their own explicit problem solving protocols which were developed through the history of the founder and firm. Third, GTM implementation advanced when the practice received official organizational endorsement, such as the clear link with the firm’s mid-term strategy, cross-functional projects, and large resource investments in the tasks force team. The will and efforts of individuals did not explain differences of results across cases. The present study contributes to the debate about organizational responses to institutional pressure, by empirically showing (a) the mechanism of why and how unintended decoupling emerges and (b) the understudied link between the way a practice is adopted and the way it is implemented.