Seoul Window
The short answer
An app gives the customer an arrival time. It rarely shows how waiting, travel and safety are allocated behind that number. Do not mistake a count of all platform workers for a count of delivery riders.
Delivery begins before a rider reaches the street. A kitchen prepares the food, an order is assigned, someone waits for pickup, then travels. To examine speed, first ask which part of that chain is being compressed.
How many platform workers are there in Korea?
A government survey estimated 883,000 in 2023. Delivery and driving are only part of the group; IT and professional services are included. In that same survey, delivery and driving declined from the previous year even as the total rose. The overall increase cannot be recast as a rise in riders.
Where might the cost of speed land?
Survey respondents reported concerns including health and safety and requests for work beyond contracts. The release cannot tell us the economics of one order or identify one cause of pressure. Customer messages, app design and restaurant preparation times need separate questions.
What the evidence cannot show
The 883,000 figure covers all platform occupations, not just riders. Surveyed safety concerns do not measure one order's waiting time or earnings. A 2026 safety agreement is a commitment, not an impact result.
The point is not to condemn convenience. It is to put all the time that makes convenience possible on the same page.
What does the platform-work survey actually say?
The Ministry of Employment and Labor’s release put the 2023 platform-work estimate at 883,000, up 11.1% from 2022. Yet it also reported that the delivery-and-driving segment declined. Presenting the national total as evidence of a delivery boom would reverse that distinction. It is a survey-based estimate across occupations, not a registry of everyone carrying food on a particular night.
The same release listed respondents’ reported difficulties: requests for work outside a contract (12.2%), health and safety risks or anxiety (11.9%), unilateral contract changes (10.5%) and delayed pay (9.5%). These are survey response shares to particular questions, not incident rates among delivery riders. They help identify what to ask next, not calculate the risk of one ride.
Where does one delivery's time go?
| Stage | Visible to the customer? | Question for a fairer account |
|---|---|---|
| Restaurant preparation | Usually only as an ETA | Is waiting time assigned to anyone’s pay? |
| Pickup and travel | Partly, on a map | What does the route require in traffic and weather? |
| Handoff and return | Often disappears after delivery | Is unpaid repositioning part of the job? |
Can a safety agreement prove delivery became safer?
No. A 2026 ministry notice described a platform-delivery safety agreement. An agreement establishes a commitment and perhaps a starting point for accountability. To judge impact, one would need implementation details, a comparison period and outcomes such as injuries or protected rest time, with a suitable denominator. Neither the agreement nor the 883,000 total supplies that evaluation.
The convenience of an ETA is real for customers. So is the need to account for the unseen stages that make the ETA possible. A rigorous story would follow an order with consent, separate restaurant delay from travel, include time between jobs and avoid portraying one rider’s day as every rider’s experience. This article has not conducted that observation; its claim is about what the published data can and cannot establish.
SeoulWindow's evidence check. All platform workers → riders' waiting and travel → a measured change in injury risk are distinct measures. The cited sources alone do not establish the final stage. This is desk analysis, not on-site reporting.
Read alongside: Why Do Seoul's Shopfronts Look Alike?.