Selling Live: Metrics Architectures and Labor Stratification in China’s Live-Commerce

• In Event: Market Codes: Bureaucrats, Brokers, and Algorithms

Abstract

How does the same selling labor produce millionaire entrepreneurs and disposable shift workers within a single industry? I investigate this puzzle by comparing two labor regimes in China’s trillion-dollar live-commerce industry: influencer streaming, where workers build loyal audiences through personal accounts, and brand streaming, where workers sell through brand accounts in rotating shifts. Drawing on multi-year fieldwork in both regimes, 141 interviews with streamers, platform professionals, and intermediaries, and internal platform documents, I develop the concept of "metrics architecture" to specify how platforms structure the customer-worker relationship through algorithmic and organizational design. Two ideal-typical architectures produce opposite outcomes. A “bowl” enables workers to accumulate durable, portable customer ties. A “filter” attaches customer ties to the brand while workers carry nothing forward. Intermediaries institutionalize these designs, channeling similar labor into either entrepreneurial wealth or mass precarity. Yet these outcomes rarely appear as the result of design choices. Workers read the same interface metrics and come to narrate success and failure as personal performance. This study extends service work scholarship by showing that the accumulation of customer ties remains central to labor stratification, but is increasingly determined by platform infrastructure rather than managerial authority. It advances platform labor research by identifying a shared mechanism for explaining divergent outcomes within and across platforms. It also specifies how power operates in platform capitalism through infrastructural designs that generate inequality while concealing it.

Author

• Jun Zhou, University of Michigan-Ann Arbor