Just now
In January 2025, Seattle became the first — and only — jurisdiction in the U.S. to prohibit deactivations based solely on customer ratings.
The App-Based Worker Deactivation Rights Ordinance may be well-intended, but it has directly caused unintended consequences by significantly restricting platforms’ ability to act swiftly, including when there’s a pattern of problematic behavior.
In the first 8 months since the Ordinance came into effect, we have observed some early impacts:
- Problematic couriers stayed on the platform longer. Many low-rated couriers, whose accounts would have been previously deactivated, remained active for two extra months until they were deactivated for later, more serious violations, including fraud and identity-related misconduct.
- Thousands of orders were negatively affected. These low-rated couriers were involved in over 2,000 problematic deliveries, including delays, missing items, or incomplete orders.
- Reduced accountability in maintaining service quality. Since the ordinance took effect, the share of Seattle couriers falling below the previous minimum rating threshold has increased by 25%, compared to the same period in 2024. This shows couriers are less motivated to maintain their ratings when there is no perceived consequence.
Why ratings matter
Customer ratings have existed long before delivery apps, and they are a widely accepted tool to promote trust and quality at scale.
At Uber, ratings are more than a reflection of satisfaction — rather, they contain early and crucial signals to help platforms identify problematic behaviors and intervene or mitigate before issues escalate.
Uber alerts drivers and couriers when their ratings are declining, and shares useful tips and education to help them avoid account deactivation. A growing body of research shows that this leads to meaningful change — for instance, rideshare drivers maintained steadier speeds and took more efficient routes after receiving low-rating alerts.
Designing a fair ratings system
Banning ratings-based deactivation fails to solve the problem it was intended to address. Instead, it has created a new set of problems.
Rather than eliminating one of the most effective early indicators of safety and quality, efforts should focus on ensuring ratings create a win-win for workers, consumers, and platforms.
We believe in a fair and responsible approach to ratings:
- Transparency and opportunities to improve. Platforms should notify workers when their ratings decline, and offer clear feedback or learning resources.
- Trends and patterns over single incidents. Deactivation decisions should be based on a sustained pattern of low ratings — not single incidents. On Uber Eats, for example, ratings are calculated as the average of their last 100 thumbs-up/thumbs-down ratings from customers and restaurants.
By strengthening safeguards, ratings and ratings-based deactivations can uphold fairness for workers, improve consumer experience, and protect the integrity of the platform.