Who is to blame? How order dispatch information shapes consumer responses to worker-caused service failures

Purpose Online service platforms (e.g. Meituan and UberEats) rely on order dispatch systems – either algorithm-directed (orders automatically assigned to workers) or worker-directed (workers self-select orders from an order pool) – to coordinate daily transactions. While prior work has examined consumer reactions to service failures based on agent or failure type, little is known about how order dispatch cues shape consumer responses to failures caused by human workers. To fill this gap, the current research investigates how framing the order dispatch process as algorithm-directed versus worker-directed shapes consumer responses to worker-caused failures. Design/methodology/approach Four experimental studies were conducted with 839 participants from China and the US. Participants imagined placing an order on a platform which either assigns orders to workers via an algorithm or allows workers to select orders from a pool, and indicated their platform reuse intention (Studies 1A, 1B, 2 and 3), intention to engage in negative word-of-mouth and to demand compensation (Study 2) following a worker-caused failure. They also indicated responsibility attribution to the platform in the failure incidents (Studies 2 and 3) and their power distance belief (Study 3). Findings First, consumers respond more negatively to a platform (e.g. lower intention to reuse the platform, higher intention to engage in negative word-of-mouth and to demand compensation) after a worker-caused service failure when the order dispatch is framed as algorithm-directed rather than worker-directed. Second, this effect is mediated by responsibility attribution to the platform. Third, the effect of order dispatch information on consumer responses to service failures is moderated by power distance belief (PDB): it is stronger among low-PDB consumers and attenuates among high-PDB consumers. Originality/value The findings enrich the service failure literature in both the AI-enabled service contexts and the on-demand service contexts by illuminating the nuanced role of algorithmic mediation in shaping responsibility attribution for human-caused failures. It also contributes to power distance belief research by revealing its moderating effect on consumer responses to service failures in the context of online service platforms.