PurposeTechnological firms increasingly depend on open innovation to compete in hypercompetitive markets. To openly engage the creativity of a multitude of private actors, firms can rely on crowdsourcing. Crowdsourcing challenges global companies as they span organizational boundaries to attract multiple local partners. Global companies must engage in boundary spanning to successfully communicate and create a sense of community with smaller local partners despite status and cultural differences. The collaboration between Google and developers in China deserves to be studied in particular, because it operates within a restricted market.Design/methodology/approachThis paper argues that crowdsourcing for innovation on a global scale requires effective boundary spanning capabilities. These boundary-spanning practices ensure smooth cooperation with the crowd and solve problems relating to differences in status and organizational contexts. This study applies Bourdieu's theory of practice including the concept of capital (economic, intellectual, social and symbolic) to understand the social relationships between Google and a growing community of Chinese developers. It also draws on a case study including ten semistructured interviews, which have been triangulated with internal documents and data from selected websites.FindingsFour types of capital (symbolic, intellectual, social and economic) have been identified as important devices to understand the sources of power and the stakes of Googlers and developers in the joint field. These types of capital contribute to structure the social fields in which developers and Google cooperate and their practice. The success of the collaboration between Google and Chinese developers can arguably be attributed to Google's ability to create boundary-spanning activities in order to reduce the endowment differential in the four types of capital and improve their communication. Therefore, this research provides a deep and conceptualized description of boundary-spanning practices, as well as providing a useful contribution for managers involved in crowdsourcing via platform in culturally different markets.Research limitations/implicationsThe main limitation of this study is methodological in nature, relating to the absence of interviews with board members of Google China who are reluctant to speak about Google activities in China for political raisons. This restriction is partly counterbalanced by the analysis of publicly available secondary data such as news and communications.Practical implicationsThis research has generated practical recommendations for managers of organizations, which require optimal boundary spanning for crowdsourcing. Managers must understand the different sources of social boundaries between their organization and the crowd. The crowd should be segmented into smaller groups with distinctive identities, and organizations should systematically design boundary-spanning activities to address each boundary of each segment. The boundary-spanning activities involve a specific set of tools, programs and platforms to address the target group. Efficient boundary spanning depends on the necessity to select boundary spanners with high cultural intelligence and communication skills.Social implicationsThis paper draws on Bourdieu's theory of practice to investigate the role of boundary spanning in crowdsourcing for innovation, specifically in the joint field between Google and Chinese developers. This research reveals how boundary objects such as developer documents, websites, programs and events are essential for developers to be able to participate on Google's platform. Companies should be prepared to invest in the design and delivery of boundary-spanning activities and objects, knowing that these are also a locus for negotiation with local partners.Originality/valueThis research contributes to the literature by applying the boundary-spanner theory to Google crowdsourcing practices within a restricted market. Bourdieu's theory of practice has proven to be a potent perspective with which to better understand the positive role of boundary spanners in the joint field between Google and Chinese developers. Moreover, this practice perspective has not been used in prior research to highlight power relations in crowdsourcing for innovation. This study has shown that, in addition to boundary objects, boundary spanners can also contribute in the transfer of intellectual capital, which is the pivotal resource for boundary spanning in this field.