Managerial Buffering and the Sustainability Paradox: How Intermediary Organizations Shape the Social Sustainability of Platform Labor

Digital labor platforms have expanded employment in urban service work while shifting much of the operational and occupational risk of that work onto workers. Whether such employment can support decent work—the central labor commitment of SDG 8—remains unresolved. Algorithmic dispatching, real-time monitoring, and continuous performance evaluation have raised delivery efficiency, but they have also intensified the tension between operational speed and the conditions under which that work is performed. The prior literature has largely examined algorithmic control as a relationship between platforms and workers, yet has paid less attention to how intermediary organizations translate platform rules into enforceable—and at times protective—practices. What remains under-theorized is not the existence of intermediaries but the mechanisms through which they absorb, translate, and redistribute algorithmic pressure. Drawing upon a 16-week ethnographic investigation at Subcontracted Station J in Nanjing, China, this study integrates approximately 120 h of participant observation, semi-structured interviews with 24 relevant actors, and 827 digital trace data entries from three workplace WeChat groups to examine how subcontracted stations mediate algorithmic pressure and its impact on decent work. We find that station managers selectively absorb, translate, and redistribute algorithmic pressure through four practices we term penalty buffering, risk buffering, resource buffering, and knowledge buffering. We conceptualize this meso-level governance mechanism as “managerial buffering.” Our findings demonstrate that achieving the decent work championed by SDG 8 cannot rely solely on informal repairs by grassroots organizations; rather, protective practices within these mechanisms must be institutionalized into explicit, universal, and enforceable framework agreements. This study also acknowledges that the applicability of the proposed “managerial buffering” framework may depend on local institutional and socioeconomic contexts. Constrained by a single-case design, future research could conduct comparative studies across cities, platforms, and stations to further test the empirical manifestations of managerial buffering under diverse institutional environments.