Deliver Us from Tip Theft
Delivery drivers need clear & fair communication with platforms, including pay transparency
Cases of platform-mediated tip theft are on the rise in the United States. This summer, Doordash drivers scrambled to download a third-party app called Para to allow them to preview tips upfront before accepting a delivery, thus providing workers with tip transparency that Doordash itself does not provide. Motherboard reported that the tip transparency app was downloaded 250,000 times as of July 24th, despite the fact that Doordash rewrote its code to stonewall the app in early July.
The rapid uptake of Para by delivery drivers illustrates the need for greater pay transparency within the gig economy. Platform-mediated tip theft — often a combination of pay obfuscation and deceptive marketing — reportedly occurred at other delivery services including Instacart, a popular pandemic consumer option that relies on independent contractors to source, pack, and deliver grocery goods and services.
In another corner of the logistics landscape, Amazon will pay $61.7 million in stolen tips to Amazon Flex drivers as part of a February settlement with the FTC. Back in May of 2019, the FTC launched an investigation after discovering that Amazon was secretly using customer tips to subsidize its delivery labor costs. As part of Amazon’s massive logistics ecosystem, the company relies on thousands of gig workers — known as Flex drivers — to deliver packages to customers in their own vehicles. Despite promising drivers that they would receive the standard base pay and 100% of customer tips for grocery deliveries such as Whole Foods, Amazon instead reportedly used tips to reach the minimum base pay, effectively cutting their labor costs without notifying drivers or customers of this practice. The tip theft began in late 2016, and lasted for two and a half years according to the FTC’s complaint.
In this particular case, Amazon’s intentional lumping of tips and base pay initially prevented drivers from realizing that they were receiving less money than they were offered. As the FTC complaint documents, drivers eventually noticed a pattern of lower tips, and sent Amazon questions and complaints. Despite this, the company responded to drivers only on an individual basis with canned responses that repeated the initial misleading messaging that drivers were receiving “100% of customer tips.”
In August of 2019, upon learning of the FTC’s investigation, Amazon changed its practice, and now breaks down base pay and customer tips in the app, and gives drivers the full base pay and, separately, passes along all customer tips. While investigations and retroactive settlements are preferable to nothing, it’s worth examining how platform-mediated companies are able to get away with extended periods of tip theft.
…it’s worth examining how platform-mediated companies are able to get away with extended periods of tip theft.
Through our interviews with Flex drivers, the Labor Futures Team at Data & Society observed elements of the Flex app that helped to obfuscate, and likely prolong, Amazon’s tip theft activity. These observations complement the prevailing analysis of the intrinsic precarity of a gig marketplace in which drivers’ classification as “independent contractors” bars them from labor protections such as wage floors, forcing them to effectively cede control to their bosses.
Flex drivers are managed entirely by an app; delivery responsibilities such as shifts, routes, customer specifications, as well as support, training, ratings, and earnings are all maintained within the Flex app. As one driver told us, “the app is our supervisor and boss.” This is distinctly different from Amazon warehouse workers or “delivery service partners” (Amazon drivers who are employed by third party delivery companies) who, despite being subject to automated decision-making processes, also have human managers they can turn to.
In the Flex app, timely support is contingent on subject matter; issues concerning the successful delivery of a package are handled quickly, while worker concerns over pay are postponed. This subject-contingent support makes it difficult for drivers who have concerns about their pay or ratings to get clear answers. At least one degree removed from human support, Flex drivers routinely encounter issues the app is not programmed to account for, and when this happens, Amazon has no efficient pathway for resolution.
For example, one driver described reaching out to Amazon support after a shift to try and dispute a drop in her ratings. The driver, who had been forced to return packages after a warehouse slowdown, tried to reach out to support through the email and phone number provided in the app to explain what happened and get her ratings changed. “I had reached out to like four different people,” she said. “And I swear, it was no help.” Despite her efforts, her ratings had dropped through no fault of her own, and she feared being deactivated (the gig equivalent of firing). Recalling the incident, she said, “I felt like that was unfair. Like that was a setup for failure.”
When drivers are accustomed to unresponsive and unhelpful support, they are less likely to report an issue or concern in the first place. As Data & Society’s gig worker report Beyond Disruption argues, “the ability to correct unfair negative reviews, suspensions, and other sanctions often requires persistence, considerable free time… and resources to make up for lost income.”
This sentiment was echoed in our interviews. Participants described engaging in a cost-benefit analysis when deciding whether to reach out to support. Knowing that they could waste an hour trying to get paid for an extra 30 minutes of driving, drivers might decide it’s not worth it. One interviewee said that a typical driver would reach out to driver support “only if you really do need something, and it has to involve your packages… other than that, everything that happens while you’re on the road is on you.”
Interview participants reported mostly positive experiences with support when they called for requests specific to package delivery, such as trouble entering an apartment complex or dealing with an extra package.
He told us that some drivers take their own photos with timestamps of the package delivery in case a package is stolen.
However, for issues distinct from package delivery, support offered little relief. One driver described the difficulty of clearing his name when a package thief swipes a box after a driver has dropped it off. He told us that some drivers take their own photos with timestamps of the package delivery in case a package is stolen. Even in this case, he expressed frustration with the process for vindication. “A lot of people just say, the worker doesn’t matter,” he said. “It’s always, the customer is right. And there’s nothing we can do about it. I don’t feel like anybody has anything other than taking your own pictures and emailing support. There’s no other suggestion.”
As a recent Bloomberg article described, many Flex drivers struggle to understand and dispute drops in their ratings — which range from “at risk” to “fantastic,” and are reported across two metrics of delivery quality and reliability. As a result, Flex drivers are sometimes deactivated suddenly and without specific reasons.
Returning to the FTC case, the initial tip theft was possible because, as is typical in platform-mediated gig work, the company has full control over what they choose to withhold from workers. However, the tip theft was likely prolonged by the fact that Amazon designed a system without pathways for reciprocal communication or clear processes for conflict resolution for driver concerns, outside the successful delivery of a package. Workers, activists, researchers, and journalists have long been talking and writing about the precarious world of gig work and its tenets which make it ripe for exploitation. Labor platforms built by companies like Amazon should not make systematized deception easier to maintain by putting gates around communication and support pathways.