Beijing Tells Food Delivery Platforms: Red-Light Waits No Longer Count Against Riders

Beijing's New Rule for Delivery Couriers: The Clock Stops at Red Lights

China has up to an estimated ten million food-delivery riders, and their working day is largely controlled by algorithms that assign orders, calculate routes and set delivery deadlines. Until now, those deadlines continued ticking while riders waited at red lights. Some riders responded by speeding or breaking traffic rules to recover lost time.

Beijing's market supervision authority now says three large delivery platforms agreed, in talks with regulators and rider representatives, to automatically exclude red-light waiting time from the delivery time used to evaluate drivers. The platforms will receive traffic-light data from the transport authorities, and a rider who waits 90 seconds at a signal will be credited 90 seconds of extra time. The system is due to be tested in parts of the city in September.

The platforms also agreed to recalculate delivery times using an average speed of no more than 15 kilometres per hour, to grant more time in bad weather or difficult traffic, and to extend the available time when an order is transferred to a new rider. In China's large cities, millions of people deliver food and other goods, and earnings depend heavily on the number of completed orders. Direct deductions for late deliveries have already been largely abolished or reduced, but riders still report strong time pressure and fear of poorer ratings.

Why Beijing Is Rewriting Delivery Platform Algorithms

Traffic-light data becomes a scheduling variable

The agreement described by Beijing's market supervision authority is not a simple extension: the platforms are to receive traffic-signal data and automatically credit any red-light wait against the evaluation window. The stated logic is direct — a 90-second stop creates 90 seconds of extra time. That turns municipal road data into an operational input for gig-economy dispatching.

This is a regulatory acknowledgment that some time pressure was designed into the system. The fix also shifts responsibility: instead of asking riders to manage risk individually, the platforms must adjust the metric that drives their behaviour.

Why the 15 km/h speed assumption changes delivery windows

The second commitment is more consequential for operations. Recalculating delivery times around a maximum average speed of 15 km/h, with additional allowances for bad weather, heavy traffic and transferred orders, removes much of the incentive to ride faster than city traffic allows. For e-bikes in dense urban areas, 15 km/h is closer to safe mixed-traffic speed than the implicit speeds created by tight deadlines.

The likely trade-off is that platforms will absorb longer planned delivery windows and possibly lower order throughput per rider in the pilot zones. That may raise operating costs or require more couriers to maintain service levels, but the agreement does not yet specify how the platforms will manage that balance.

Ratings, not penalties, remain the open pressure point

The source notes that direct deductions for late deliveries have already been largely abolished or curtailed, yet riders continue to report intense time pressure and fear of lower ratings. The red-light credit therefore addresses one source of pressure but not the entire incentive structure: if customer ratings still reward the fastest possible delivery, riders may still push speed where they can.

For the September pilot to deliver its safety objective, it will need to show whether adjusted delivery windows actually change rider behaviour — and whether platforms treat the new credits as real time rather than offsetting them by tightening other parameters.

What the Beijing Pilot Means for Platforms, Riders and Investors

• Before the September launch, the three platforms should confirm that Beijing's traffic-signal data feed can be matched automatically to active orders; any manual fallback would recreate the old pressure on riders.
• Test whether the 15 km/h maximum average speed and bad-weather allowances produce clearly longer windows in pilot districts, and measure rider-reported pressure and moving violations, not only on-time performance.
• Watch whether credited red-light time is offset by tighter routing or higher delivery volume per shift, since the agreement also covers weather and transferred orders but does not cap workload.
• For investors in China's delivery platforms, the pilot's cost-per-order and courier-utilisation figures will signal whether a citywide rollout implies a structural increase in delivery time assumptions and operating costs.