Week 5: Worker resistance

How riders route around the system that routes them — GPS spoofing, order refusal patterns, and WeChat groups that coordinate informal slowdowns

Individual workarounds and the patch that follows#

Fake-GPS apps that report a rider further along their route than they actually are; deliberately declining low-value orders to protect an acceptance-rate score, even though the platform can read that refusal as bad behaviour — every individual tactic riders build starts an arms race with the next platform update. That race is itself data: a workaround only gets patched once it’s common enough to threaten a metric the platform cares about, so which tactics survive tells you which numbers actually matter to the business.

Ping Sun’s 2019 study, “Your Order, Their Labor: An Exploration of Algorithms and Neoliberalism in the Food Delivery Platform Economy in China” (Chinese Journal of Communication), documents this arms race from the rider’s side directly — riders describing which workarounds they trust to survive a month versus which ones they expect a patch to close within weeks. Sun’s fieldwork is worth citing precisely because it treats riders as strategic actors responding rationally to an incentive structure, not as passive subjects the system happens to.

Alex Rosenblat’s 2018 book Uberland: How Algorithms Are Rewriting the Rules of Work documents the same dynamic on Uber’s own driver forums — drivers trading notes on which routes trigger surge pricing, which cancellation patterns avoid a deactivation flag, and how “multi-apping” (running Uber and a rival driver app at once, accepting whichever pays better) turns the platform’s own incentive structure back on it. Same arms race, same rational-actor reading, a different continent.

Collective tactics change the shape of the risk#

Regional WeChat and Telegram groups that share which restaurants are running long kitchen delays are, functionally, riders crowd-sourcing the information the platform’s own ETA model doesn’t have — and sharing it faster than the platform can. This is a materially different kind of resistance from an individual workaround: it’s coordinated, it’s visible if anyone looks for it, and unlike a single rider spoofing their GPS, it can’t be patched by changing one parameter in a model. Some of these groups go further than information-sharing and coordinate an informal, undeclared slowdown on a specific day — riders in a region collectively declining orders below a set price for an hour, without any formal strike structure or recognised bargaining unit behind it.

Why the platform’s response tells you what it’s afraid of#

Individual workarounds tend to get patched quietly, in a routine update nobody announces. Collective coordination gets escalated — flagged internally, sometimes cited in exactly the kind of policy debate this course studies in its second half. That asymmetry in response is itself evidence: it tells you which kind of resistance the platform actually treats as a threat to its model, versus which kind is just noise to route around.

Before the session#

Find one documented example (reporting, a forum thread, a patch note) of a platform responding to a rider workaround. Bring what it tells you about what the platform was actually worried about.