Opaque Shipt pay algorithm slashes employee pay
Opaque Shipt pay algorithm slashes employee pay
Occurred: October 2020
Report incident 🔥 | Improve page 💁 | Access database 🔢
A new pay structure introduced by Target-owner grocery delivery service Shipt prompted a backlash after it cut employee earnings by up to 50 percent.
Shipt rolled out a black box, "effort-based" algorithm, dubbed "V2", to replace its previous flat-fee model, claiming it would better account for order effort, including number of items, peak shopping times, and location, and provide fairer compensation to workers.
However, many workers reported pay cuts of 30-50 percent and highly unpredictable earnings, and complained about a lack of transparency about how pay was calculated. Shipt's previous system had allowed workers to choose jobs based on known pay rates, but the new algorithm removed this ability, reducing workers' sense of autonomy and causing widespread anger and stress.
Classified as independent contractors, Shipt employees retaliated by organising strikes and walkouts, boycotted the platform and asked customers to do the same, and collaborated with researchers to analyse their pay data.
They also created a "Shopper Transparency Calculator" to collect and examine pay information, finding that 40 percent of workers were earning at least 10 percent less under the new system, with around 33 percent earning less than their state's minimum wage.
The fracas was seen to highlight the power imbalance between platforms and workers regarding algorithmic management, the lack of data rights and transparency for gig workers, and the need for regulation requiring companies to disclose information about AI systems used in the workplace.
V2
Operator: Target/Shipt employees
Developer: Target/Shipt
Country: USA
Sector: Transport/logistics
Purpose: Calculate pay
Technology: Pay algorithm
Issue: Ethics/values; Fairness
Calaci D., Pentland A. The Shipt Calculator
Calaci D. Bargaining with the algorithm: pooling worker data to estimate gig economy worker pay
Page info
Type: Incident
Published: August 2024