Loading page…
The dispute is over more than a declining fee record. Riders want to know how automated systems distribute work, price orders and affect access to the apps that determine their income.
Listen to this story
Food-delivery platforms face a transparency challenge over automated allocation, pricing, and account-access decisions. One Edinburgh rider’s logged average fee fell from £3.67 per order in 2023 to £3.42 in the first half of 2026, despite a stable 3.6–3.8 orders-per-hour pace. Deliveroo, Uber, and Just Eat describe human safeguards and efficiency-focused matching, but do not disclose the rules behind individual offers.
The rider’s average fee declined annually: £3.67 in 2023, £3.63 in 2024, £3.51 in 2025, and £3.42 in early 2026.
Deliveroo calls its order-allocation system Frank, which predicts timing and assigns work; the company does not explain individual fee calculations.
Platforms’ living-wage comparisons exclude unpaid waiting time, limiting what those headline earnings claims capture.
One Deliveroo rider’s average fee per order fell from £3.67 in 2023 to £3.42 in the first half of 2026, even as his delivery rate stayed between 3.6 and 3.8 orders an hour. Edinburgh food-delivery riders are urging Deliveroo, Uber Eats and Just Eat to explain how the automated systems that govern their work set offers and pay.
The workers say pay and conditions have worsened as the platforms increased automation. Their complaint is not that software helps run delivery networks; it is that they cannot see the data and rules shaping the jobs offered to them, the fees attached to those jobs, and decisions affecting app access.
The rider’s records track a gradual decline: £3.63 per order in 2024 and £3.51 in 2025, before reaching £3.42 in the first half of this year. The log is one worker’s experience, rather than a measure of earnings across every courier, but it frames the transparency fight around a visible output and an opaque payment decision.
Deliveroo calls its order-allocation system Frank. The company says it uses machine learning to predict the timing of each order, assign work to riders, and make deliveries more efficient and reliable. That describes Frank’s operating role, but not how a particular rider’s fee or offer is calculated.
All three companies say riders earn more than the national living wage while on an order. That calculation excludes the time riders spend waiting for offers, leaving a key part of a shift outside the stated measure.
The Workers’ Observatory, created by gig-economy workers and academics at St Andrews and Edinburgh universities, has tested variable automated offers. In Dunfermline, some riders rejected low-paying jobs while others logged on at the same time; some workers then received higher short-term rates. One rider was deactivated soon after, although the reason was unclear. The major platforms say riders are not deactivated for declining offers.
Automated access checks are also part of the dispute. Graeme Frances said Deliveroo’s facial-recognition system blocked him from logging in after an accident gave him a black eye. Deliveroo said blurry images caused the failures and that he could work after a manual review; Frances said he did not realize the review had cleared him until he next logged in.
Drivers from the UK, Netherlands and other countries have filed a class action in Amsterdam alleging Uber’s AI-powered system breaches data-protection law and suppresses earnings based on what an individual will accept. Uber denies setting trip prices according to individual driver behavior. A 2023 study by Oxford and Columbia Business School researchers found Uber drivers earned substantially less per hour after dynamic pricing was introduced, though that result does not resolve the food-delivery riders’ claims.
Loading discussion...
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.