Gigarbetare spelar på rörlig plan när Uber och Lyft efterliknar mobilspel

Gigarbetare har länge fått höra att delningstjänster som Uber och Lyft ger dem en chans att själva bestämma när, var och hur de vill arbeta. I verkligen är dock förarna hårt styrda. Inte av en chef, men av en algoritm, skriver Bloomberg. För förarna är apparna fyllda av gamification och de uppmuntras att ta olika turer för att nå nästa level. Men information som är nödvändig för att plocka poäng, och framförallt tjäna pengar, är ofta dold. Och spelreglerna för förarnas försörjning förändras hela tiden. Men som en motreaktion kommer nu en marknad för tjänster som ska hjälpa gigarbetarna att syna – och klå – algoritmen.

Gamification Took Over the Gig Economy. Who’s Really Winning?

Ride-share drivers say that the pandemic has exacerbated the imbalance with their overlords.

Jackie Davalos and Drake Bennett, Bloomberg Businessweek, 27 May 2022

In 2008, University of Chicago economist Richard Thaler and Harvard Law School professor Cass Sunstein published  Nudge: Improving Decisions About Health, Wealth, and Happiness. It’s a breezy tour through behavioral science, especially the ways in which decision-making can be shaped by what the authors call “choice architecture.” When a store puts sugary snacks at eye level so a customer is enticed to buy them, that’s choice architecture, and when credit card companies set low minimum payments to get people to pay interest, that’s choice architecture, too. The book’s argument is that we should enlist so-called nudges to socially beneficial ends, such as automatically enrolling workers in 401(k) plans. In the years since the book’s release, policymakers have occasionally adopted the ideas, with mixed results.

Where nudges have really caught on is in the part of the tech world that uses algorithms to mobilize and manage freelance workforces. Armed with troves of data and the quickly improving capabilities of machine learning, online platforms such as  Uber Technologies Inc. and  Lyft Inc. use nudges to coordinate millions of independent workers and extract maximum productivity. This model let Uber scale quickly, with the force of a command-and-control structure, even while corporate framed its drivers as independent, self-directed, and entrepreneurial. The company’s rapid growth was seen as proof that the human-machine relationship could work, and it became a template for a generation of startups looking to take a cut of people’s labor without putting them on staff. The concept of directly connecting customers to contractors paved the way for marketplaces including home-cleaning (Handy), pet-sitting (Rover), and, of course, food-delivering (Grubhub, DoorDash, and Uber again with Uber Eats).

For more than a decade now, gig economy companies have told workers that signing up for ride-share or delivery side hustles will allow them to be their own boss, to choose how and when to work in a way that fits their lifestyles. Instead, workers have found themselves managed by incentives and code. Sophisticated software holds real-time auctions every second, matching prospective customers with drivers or couriers in a vast marketplace. This system, the businesses tell investors, lets them grow — and if necessary shrink — faster and with less friction than a company reliant on old-economy ideas such as salaries, employees, and managers. If the practical applications of nudge theory sound like a win-win, then the companies accomplished their mission.

To many drivers, the flexibility of their contractor status is indeed a benefit, but many also describe an experience plagued by uncertainty and arbitrariness. The business model they participate in depends on an asymmetry of power and information: Typically, for example, they have to decide whether to accept a ride without knowing where the pickup or drop-off will be. Unable to assess how much money a trip will bring in, they’re at a disadvantage in the great labor-bidding scrum. “They call us independent workers, as if we have control over what goes on in the app,” says Anthony Arnold, who drives in Las Vegas. “Sure, you can make a bit of money, but it’s not going to be on your terms.”

Even before Covid-19, in other words, being managed by a ride-hailing app felt like competing with a wily, well-funded operation that knew how to keep drivers driving, tip the odds away from them, and take a bit more of their money. And today the number of variables has only grown. Drivers say rates fluctuate more and more now and that making an informed decision about what rides to accept is predicated on a game whose rules are constantly changing. It’s as if the app that governs their work life has taken the uncertainty of the current moment, fed off it, and supercharged it. In the process, the whole intricately choreographed system built on the software is being put to an unprecedented test.

Once drivers log on, the platforms keep tabs on their every move: the percentage of available pickups that they accept; the time of day they typically work; the neighborhoods they prefer; the ratings passengers assign after a ride. For their part, drivers see a dashboard with bar charts of weekly earnings, trips completed, and fare breakdowns. But their visibility is limited.

Take surge pricing. If demand in an area outpaces supply, the apps light up with flashing spots on a heat map, and fares climb to draw cars. The platforms send push notifications to drivers who are about to log off, telling them how close they are to advancing to the next performance tier, which unlocks benefits such as the ability to see a trip’s duration and direction. Or, as in a video game, bonus offers pop up: $30, say, for completing 10 consecutive rides.

What should be a straightforward chance to make more money can be a Kafka-esque ritual benefiting the platforms more than the drivers. Insistent pings prompt them to accept a fare before they’ve completed the one they’re on. Drivers are also part of rewards programs. They accumulate points by being behind the wheel at certain hours, maintaining high customer approval ratings, and keeping cancellation rates low. The points help them reach higher tiers and are redeemable for discounts on fuel and repairs. But the thresholds for qualifying for various perks can be stringent, such as needing to accept 9 out of 10 rides or keeping an almost-perfect customer rating. And after a “qualifying” period (Uber’s is three months), the score resets.

These measures, Uber and Lyft say, prevent drivers from cherry-picking rides and ignoring potentially less profitable customers. (The companies also say that they’re focused on increasing transparency by piloting new programs to improve the driver experience, though an Uber spokesperson says “we fundamentally disagree with the premise” that driver pain points have gotten worse.) But the lack of clarity also contributes to a sense of futility. “Even with all the data we can see and track, it’s hard to know what’s going on inside the algorithm,” says Surgeet Singh, who drove for Uber and Lyft for four years before quitting in February to work as a courier for Grubhub and DoorDash. A December Pew Research Center report found that fewer than half of gig platform workers said they understood how their pay was calculated. “These workers have very few ways of knowing whether they can count over the long-term on the level of pay they’ve achieved,” says Drew Ambrogi, a senior campaigns strategist at advocacy organization Coworker.org.

The resulting resentment helps explain the difficulties ride-hailing companies have had getting drivers back on the road. Two years ago, lockdowns crippled demand for ride-hailing and triggered a driver exodus. Many found other work that didn’t involve sitting in confined spaces with strangers. Some were better off collecting unemployment benefits. More recently, as life has fitfully returned to normal, ride-hailing demand has climbed back. To meet it, Uber and Lyft have done what the basic laws of economics would suggest, spending hundreds of millions of dollars in the past year on incentives and bonuses. Before expenses, driver earnings averaged a little more than $24 an hour in March, a roughly 56% increase from pre-pandemic pay. That was, in large part, because of these perks, according to Gridwise, a gig worker productivity app.

But wait times and fares remain elevated, suggesting constrained supply. This situation has boosted revenue for the companies but also uncertainty about how reliably they can manage contractors, raising the possibility that the mechanism that once subsidized cheap rides for customers is no longer tenable. Meanwhile, drivers continue to leave the industry. A’Maya Davidson was a long-haul trucker before starting to drive for Uber and Lyft four years ago. Now she’s going back to trucking. “I want more security,” she says. “I’m just tired of all the games. There’s really no upside to this job anymore.” Uber, though its spokesperson, says its US drivers recently had the highest 28-day retention rate since the onset of Covid.

The labor shortage has given the remaining drivers new leverage. “The power dynamic is rebalancing,” says Mareike Möhlmann, assistant professor of information and process management at Bentley University, who co-authored a 2017 study on drivers and algorithmic management. It’s a curious moment, she says. Drivers feel emboldened even as they’re not getting a commensurate cut of the inflated prices that customers are paying, and they’re working for companies that seem unwilling to disrupt the model that’s helped them acquire millions of customers. The shift may be temporary. But if it’s not, the companies may have to rethink the management model that’s been central to their growth and without which they may not be viable.

As the platforms contemplate their chess moves, so do the drivers. Prices for gas — long one of drivers’ biggest out-of-pocket expenses — have spiked almost 50%, devouring take-home pay. That, along with the stubborn persistence of Covid itself, has only heightened the furious gamesmanship that happens behind the scenes every time a passenger hails a ride.

Even the new bonuses offered to entice them, drivers say, are gamified to make them harder to claim. They say Uber has shortened the time to accept a ride from 20 seconds to about 5. (Uber disputes this. Its spokesperson says drivers have as much time now — more than 10 seconds — as they’ve always had.) In September, Lyft tweaked its surge pricing model by replacing “personal power zone” heat maps, which rewarded drivers for flocking to high-demand areas by offering rewards, with dollar amounts that increased the longer a driver stayed in the zone. Then came “bonus zones,” and the dollar amounts stopped increasing. Bonus zones “may update based on how busy it is in real-time,” according to Lyft’s website.

The companies have also made subtle policy changes that, in combination, can affect income. Last year, Lyft changed the fee drivers collect when a rider cancels from a flat $5 to a variable amount — often less than $5 — based on how quickly the cancellation was made. Uber tweaked its surge pricing incentives, which vary by market, by replacing a multiplier on fares with a set dollar amount: A $20 fare with a “2x” bonus was $40; now it’s $22. Recently, Uber rolled out a feature called upfront fares that lets drivers see more details of a prospective ride before accepting it. The company says this lets drivers exercise more control. But the update came with a trade-off. Uber removed the ability to see how pay — typically based on fixed time and distance rates — is calculated. Where before drivers had two data points, they now have only one: the flat fare.

The difficulty of keeping abreast of these changes is feeding a niche industry of bloggers, YouTubers, consultants, and even venture-capital-backed apps — a kind of gig economy economy. One of the first was the Rideshare Guy blog, which longtime driver Harry Campbell started in 2015. It’s since exploded into a media operation with newsletters, podcasts, and videos offering tips for maximizing income across the gig economy. Other driver-influencers have emerged as well, such as Torsten Kunert of Los Angeles, also known as the Rideshare Professor, whose YouTube channel has more than 53,000 subscribers. Kunert teaches courses on identifying earnings patterns and mastering surge bonuses. He says sign-up rates have doubled in the past six months.

Some startups are offering to simplify the calculations behind take-home pay. Solo, founded in Seattle in 2020 by former Uber executives, announced last year that it had raised $5.5 million. Its app gives drivers hour-by-hour earnings predictions, and the company has promised to make up the difference if their projections mislead drivers.

It’s possible to look at all this and see signs of hope for drivers. They now have a slew of options, says Gad Allon, a professor at the Wharton School who studies the gig economy. His research estimates that about 86% of drivers work across multiple apps: driving for Uber and Lyft, for example, while delivering for DoorDash and Instacart. Forty percent of those contractors toggle between platforms in the course of a day. “It cannot just be ‘We will offer you more flexibility and will pay you per transaction,’ ” Allon says. “Eve Amazon has realized this is going to be the issue of the next decade,” he says, citing its move to provide free college tuition for select front-line employees. Walmart Inc. adopted a similar program.

For their part, Uber and Chief Executive Officer Dara Khosrowshahi now talk about being an “earner-centric” company, tending to the concerns of drivers and couriers. That pledge builds off an effort that started last year when Uber introduced a team dedicated to boosting retention. “Coming out of the pandemic, we knew it was going to be hard to win drivers back, and we took that moment to introspect and reflect as a company,” says Carrol Chang, senior director and global head of driver and courier operations. Chang says that in addition to more transparency around pricing, the team has tackled issues that have long rankled drivers, such as outdated maps, payment glitches, and slow chat support. “The driver experience has to be at the center of what we do,” she says. A Lyft spokesperson says the company is testing upfront pay in select markets and that, in April, it started letting drivers take 15-minute breaks without losing their bonuses on back-to-back rides.

Of course, there are some for whom the abstruse cat-and-mouse game is almost the point. “I’m not an Uber hater or Lyft hater,” says Sergio Avedian, who’s been driving for five-plus years and completed more than 5,000 rides combined for Lyft and Uber in the LA area. “I like the game. If there’s a way to beat them, I’m going to figure it out.” The retired derivatives trader tracks data from his trips in a spreadsheet that includes pickup locations, destinations, mileage, time of day, tips, and other earnings. As a ride-hailing coach, he charges $75 for a 45-minute call and co-hosts a YouTube show called  Show Me the Money Club. He says it takes about six months to get good at mastering nudges. But these days, people don’t always stick around that long.

More stories like this are available on bloomberg.com.