Reflections on a food delivery workshop that almost did not happen (part 2)

Niels van Doorn

Proceedings from the workshop

After I gave a general introduction, Uber was given the opportunity to briefly comment on the five discussion themes and address how it envisions its role in improving the wellbeing and income opportunities of its “delivery partners”. Uber's representatives explained that NYC is a big and important market for Uber Eats and, for the company to succeed on the long run, it will need to do better and be better partners. One way Uber Eats has been making improvements is by engaging its partners through roundtables, phone calls, and surveys. The new partner app that was recently rolled out has likewise been designed with structured input from drivers and bike couriers. Considering Uber within the larger context of how the realities of work are transforming, they raised four thematic points where the company can be of value to workers: 1) Access, in the sense that Uber lowers access barriers to work and income, but also to the extent that partners should have access to information so they can decide if the opportunity works for them; 2) Flexibility, in the sense that partners can decide if, when, and where they work without having to fear any consequences for changing their mind; 3) Protection, in the sense that Uber seeks to fill the gaps in existing social safety nets, which is why the company recently rolled out its “Partner Protection” program in Europe and is looking into similar initiatives elsewhere; and 4) Opportunity, in the sense that Uber is exploring how it can be a stepping stone to something else, for instance by supporting its partners as they seek an alternative job trajectory or career path. Uber also emphasized that partners have flexibility in terms of where and when to work, and are not required to reserve “blocks” of time during which they can work. (This is different from other companies’ models). Instead, Uber Eats offers “true flexibility” while also aiming to improve the transparency of its app with respect to providing better (real time) information about busy periods as well as earnings. After this statement, which loosely touched on all scheduled discussion themes except for Accountability, it was the couriers’ turn to offer their input. As I expected, the ensuing discussion veered in all kinds of directions and touched on a great variety of subjects, making it impossible to neatly stick to the planned thematic order of the workshop, but for clarity’s sake I will nevertheless use the five themes to structure my account of the proceedings.


One issue that repeatedly came up during different parts of the discussion was the discrepancy between Uber’s celebration of flexibility and the desire for more income stability among “fulltime” couriers for whom food delivery is the primary source of income and who mostly rely on the apps to make ends meet (as was the case for all couriers present that afternoon). It is not that working for Uber isn’t flexible, because in many ways it is – everyone agreed on this. But, as one courier noted, Uber’s image of itself as a service that provides quick and easy access to “extra cash”, or a supplementary income stream, does not align with the economic realities of many of its couriers. Even though, according Uber’s data, in most US cities about 80% of couriers work less that 10 hours per week with Uber Eats, the couriers in the room agreed that they’d love to use the app more if Uber Eats would make it worth their while. The reason for their relatively limited use of Uber Eats wasn’t that they were merely supplementing their income from some other, primary job, but because it wasn’t economically viable for them to put in more hours on this app – they couldn’t rely on Uber Eats as a main source of income so they had to work with a number of other companies to cobble together a livelihood. Most of them currently used three or more apps, having figured out a precarious routine that worked, for the time being, as they navigated what many felt to be a form of forced flexibility. Moreover, despite the fact that they could technically sign in and work whenever and wherever they wanted, there were clearly “pockets of time and areas with little work”, as one courier phrased it, such as the afternoon hours between the lunch and dinner rushes.

In response, Uber’s representatives first of all agreed that New York is different from most other food delivery markets, not only in terms of size, level of fragmentation, or its particular history, but also with respect to the number of “fulltime” couriers and the type of vehicle they use – primarily (e-)bikes in Manhattan. Uber Eats has been trying to increase demand outside of traditional meal times, by encouraging breakfast deliveries and expanding the number of corporate orders, with the aim of decreasing down time and spreading income opportunities throughout the day (and night). When someone suggested that Uber could create a tiered system that provides an option for partners who’d like to work fulltime (and then receive an hourly wage), Uber Eats repeated that the company’s model focuses on flexibility and providing low-barrier access to work. It does not want to control how many hours its partners have to be logged in and, given the market’s unpredictability, it would be infeasible to have couriers work set hours. Some couriers wondered, however, why Uber Eats wasn’t able to better predict how many couriers it would need on a given (segment of the) day, with all its advanced technology, and they also questioned the company’s “open access” policy: perhaps Uber Eats should stop recruiting new partners and increasing the number of couriers on the streets, given that this decreases the income opportunities of existing couriers (who, to be sure, are paid per order). One courier claimed this was the first time since he downloaded the Uber Eats app that he experienced extended periods during which he didn’t get a single order. Uber’s representatives did not offer a solution to this problem.


One other point that all couriers – especially those with years of experience – agreed on was that it used to be so much easier to make good money doing food delivery in New York City, especially in the early days of Postmates and (the now discontinued) Uber Rush. Even though Uber Eats noted that the company’s base rate had not changed since May 2017, many in the room noticed that it had become more difficult to reach their earnings goals compared to early 2017 (when Uber last reconfigured its pay structure), forcing them to become more resourceful juggling multiple apps. Uber Eats regretted to hear that this was the case and guaranteed that Uber is always working to improve its evolving pay structure so that it better rewards its delivery partners’ effort. The couriers’ response to this rote company line was rather lukewarm, if not incredulous.

Beyond its base rate, Uber Eats incentivizes its delivery partners through promotions, either in the form of opportunities to earn extra money (via so-called “Quests” or in “Boost zones”) or by offering an earnings guarantee, all of which are contingent on whether a courier meets a set of quantitative criteria (i.e. number of deliveries within a specific time frame and order acceptance rate). While the couriers generally liked Uber’s promotions – some claimed that promotions were the only reason they continued to work with Uber Eats at all – because these incentives enable them to potentially boost their earnings and compensate for the low base pay, they also complained about their inconsistency and lack of reliability. It wasn’t clear to these couriers what made them (in)eligible for particular promotions, as they had gathered by now that not every courier receives the same offers. Moreover, there seemed to be a shared sense of suspicion with respect to their chances of completing certain Quests and actually receiving the money: there were quite some stories doing the rounds about couriers having to wait a long time before receiving the final order needed to complete a Quest and claim the promotion, while others allegedly couldn’t complete theirs because that final order never appeared on their screen before the promotion expired. In response, Uber Eats ensured everyone that Uber’s system is not programmed to keep delivery partners from reaching their Quest goals and that there must have been less nefarious reasons why some partners experienced such problems. Couriers also complained about the relatively long distances of some order they had to complete during a Quest (when declining an order would disqualify them from the promotion), and about promotional payments not showing up in their account until a week later. Meanwhile, Boost zones (where earnings are boosted by a variable and dynamic multiplier) were much appreciated, but all couriers knew that these were exceedingly rare in Manhattan, where most of them worked – at least during the day. Boost zones were more common in new markets like the Bronx and parts of Brooklyn, but in these areas one needs a car to really make food delivery work. Generally, all couriers looked forward to the Fall and Winter seasons, when – they were sure – promotion-based income opportunities would increase as the weather would get worse and competition would decrease after those with less skin in the game have transferred to other, less arduous forms of work.

One other reason many couriers stick with Uber Eats is its “instant pay” feature, which allows partners to cash out whenever and however many times they want – at the cost of a small fee. All couriers at the table loved this feature, which some other companies now also started to offer, because they don’t like and often cannot afford to wait for a (bi-)weekly paycheck. With instant pay, they have access to the money they earned that same day, so they don’t have to worry about sudden expenses when these inevitably arise – for instance when their bike needs to be fixed. It also compensates, in a small way, for the lack of tips they receive from customers, either through the app or in cash. Everyone agreed that Uber’s customers are very poor tippers, probably because tipping wasn’t an option in Uber’s app until quite recently. Couriers suggested that there should be an option in the app to tip in advance, before the order is delivered, in addition to tipping afterwards – as is the case for Caviar’s app, for instance. Someone also proposed to modify the app so it allows couriers to rate customers who do not tip or tip very poorly. Uber did not directly respond to these suggestions.


During the workshop it became clear that Uber understanding of where transparency matters does not always align with how couriers and academics think about transparency and in what areas it should be improved. Uber Eats emphasized the importance of improving transparency with respect to information about earnings as well as earnings opportunities, so that Uber’s partners have a clear view on how their income is calculated/composed and how they can use the app to optimize their earning potential. For instance, by providing real-time information about which parts of the city are – or will likely become – busy, the company hopes to assist (as opposed to influence?) the decision-making processes of its delivery partners. In order to make its partner experience better, Uber regularly hosts roundtables and conducts surveys to consults drivers and couriers, and Uber Eats was keen to point out that its new app – called Carbon – incorporates many of this crowdsourced feedback. Some of the couriers attending the workshop were eager to know more about opportunities to act as an (unpaid, except for travel expenses) consultant, while others were more cynical about Uber’s motivations and the extent to which their voices would be heard in these settings. All, however, agreed that Uber’s “blind dispatch” – where couriers cannot see where they are expected to drop off the food until they have accepted the order and picked it up – attested to the fact that the company still had some improving to do when it comes to transparency. Those who already got the new app liked how it indeed gave a clearer earnings overview, but wondered why blind dispatch wasn’t discontinued. If Uber truly was committed to helping its partners make the best possible decisions, why not provide them with all the relevant information they need to decide whether or not to accept an incoming order?

Uber Eats responded to this question by explaining that the historical reason for not showing partners the drop-off address until they accept the job is because people used to decline orders based on location or on the particular restaurant, which Uber conceives as a form of discrimination and aims to avoid. When couriers countered that they should be allowed to decline any order they don’t want because they are independent contractors (something Uber agrees with and, it claims, does not penalize) and that other companies do not or no longer have blind dispatch, Uber’s representatives repeated the challenges with respect to this issue and said they were looking into what more information Uber Eats could usefully provide couriers before they accept a trip. The conversation then also turned to the issue of what one courier called “bonus transparency”. As mentioned earlier, there are quite some uncertainties among the couriers regarding the criteria that determine who receives which promotions at what time, and many felt that Uber could do more to clearly communicate what partners can do to optimize their eligibility. After all, the chance of getting good promotions is one big reason to keep using the Uber app at a time when its base rate is relatively low compared to some of its competitors. Another courier in the room compared Uber’s system of promotions to a video game; designed to keep you playing – or what was generally called hustling – while trying to get ahead and make sense of the rules as you go along.


I had imagined in advance that the theme of support would primarily revolve around the possibilities of improving Uber’s live support for its partners, which I heard quite some complaints about during my interviews, but as it turned out this theme emerged in a rather different context. What couriers really wanted Uber’s support with, first and foremost, was improving their safety when they were out on the streets delivering food. Nearly all of them had sustained some kind of injury on the job and could share stories about times when they either had serious accidents or managed to narrowly escape them. One courier told us how he broke both of his arms, when doing deliveries through another app, and had to move back home to live with his family until he recovered. When he tried to get courier insurance he learned that due to his high risk profile the premium would be far beyond what anyone of his means could possibly afford. Like his peers, he was therefore very interested to hear more about Uber’s rollout of its European Partner Protections program and its tentative exploration of ways to introduce similar insurances elsewhere. There was a clear sense of disappointment in the room that this did not yet extend to the US and couriers urged Uber to try harder and expedite their initiatives, as they desperately need accident insurance and other protections. There was a broad consensus that delivery companies as well as regulators need to step it up because the risks of doing food delivery in New York City are so ubiquitous that, as illustrated by the story mentioned above, buying private insurance is simply unaffordable for individual couriers.

Although the couriers welcomed the pending Fall and Winter seasons because they were likely to make more money on promotions, they also dreaded the poor weather conditions that instigate the proliferation of these same promotions. Wet, slippery streets, low temperatures, heavy rain, snow, and winds – such conditions all exacerbate and compound the everyday risks of doing food delivery work. Uber Eats empathized with the couriers’ situation and recognized the abundant risks involved in this line of work. They also discussed some product developments Uber is currently working on to improve courier safety. Although the couriers were glad to hear that Uber is working on concrete improvements, they also considered such initiatives as merely a small step in the right direction. Ultimately, according to many in the room, a much more comprehensive approach is needed and couriers expect more from Uber in terms of support. There was a worry that couriers are seen as replaceable by delivery platforms, which have access to a large pool of new delivery partners who are ready to work. This concern connects to another topic that often came up during the workshop: couriers want respect and dignity, which they often lack. Accordingly, they want their work to be properly appreciated, both in moral and economic terms. Being a courier entails “way more than just picking up food and delivering it”, as one participating courier phrased it. It is far from the “unskilled” labor it is often made out to be and in fact requires a host of abilities and talents, including so-called “soft skills” ostensibly valued by employers. In other words, the notion of support should be understood in much broader terms, when thinking about what food delivery platforms can do to improve the working conditions and income opportunities of their courier fleets. Moreover, it is important to recognize that couriers support Uber and other companies as well: as another courier argued, “delivery workers know this city, they don’t need a map and they make the service better as they do their work.” What these couriers expected, indeed demanded, was that this supportive “partnership” is more reciprocal and well balanced when it comes to each party’s rights and responsibilities.


The one theme that didn’t get much attention, mostly because there was so much to discuss with respect to the other themes and we had limited time available, was the question of accountability. I had intended to introduce this theme in order to debate the extent to which couriers should be able to hold Uber accountable for its actions and what organizational reforms and innovations would be needed to institutionalize the procedures by which such accountability would be practiced/governed. More specifically, and more modestly, I would have liked to discuss the possibility of creating a policy ensuring that no courier can be deactivated without a formal chance to appeal this decision and a subsequent process of (openly conducted) review. Many couriers I spoke with during my time in New York had been (temporarily) deactivated by one of the delivery platforms, usually without any form of recourse. This can have serious consequences for couriers’ livelihood and even their ability to keep a roof over their heads, especially when they cannot turn to other apps to compensate for their sudden loss of income. More ambitiously, I had imagined we could discuss a policy that prevents Uber from making sudden and/or major changes to its product and daily operations without due process and input from delivery partners. If you’re all in this together, as Uber likes to point out, then important decisions should also be made collectively, with such power extending beyond executive ranks. Unfortunately, however, the most ambitious discussion theme of the day didn’t make it to the table, which I regret as I make up this inventory of what happened during the workshop. All in all, I have mixed feelings about the event.

As I mentioned in part 1 of this extended research note, a workshop like this in no way offers a meaningful substitute for collective bargaining and organizing initiatives. It was merely intended as an experiment – a minor platform enabling food delivery workers to voice their complaints and concerns in a space where their voices are heard and (hopefully) taken seriously by representatives of a company that does not employ them but nevertheless shapes the conditions under which they do their work. While the resulting conversation was deeply engaging and laid bare a number of problems that urgently need to be addressed, Uber eventually made no commitments to significantly change how it runs its business. This was to be expected, of course, but secretly I had hoped for more – some concrete concession that directly responded to one of the many suggestions for improvement put forward by the couriers that day. Why can’t Uber not just end blind dispatch? Why can’t the company ease the worries of its partners by committing to not further lower its per order base rate, or even to raise it? Just because these questions are partly rhetorical does not mean they are not worth asking with some frequency and insistence. I want to close by expressing my special gratitude to the couriers who showed up and participated. Without the generosity with which they shared their knowledge and experiences the workshop surely would have been a failure. I also want to thank Emilie and Jordan at Uber, for staying on board when the other companies jumped ship and for their openness to criticism and suggestions for improvement. A number of couriers later told me they had enjoyed the opportunity to have a face to face discussion with Uber representatives and were disappointed that Postmates and Caviar didn’t show up. Still, while they appreciated Uber’s presence, none of them had high expectations regarding the company’s readiness to really improve their pay or their working conditions (aside from maybe introducing an affordable insurance program). This, everyone agreed, would require more comprehensive industry-wide commitments, which won’t be easy to attain. Not easy, but not impossible either. As I was evaluating the workshop with some colleagues at the end of the day, someone suggested that it would be a good idea to organize a follow-up workshop, this time including not only more companies but also relevant policymakers and legislators. I am currently exploring what it would take to make this happen.

PS: Many thanks also to Aiha Nguyen and Data & Society for all their support, and to the academic colleagues who freed time in their schedules to attend the workshop.