Gigged Page 3
The only hurdle that might have stopped him was a long interview, the very thought of which made Curtis cringe. In college he had spent his free time on coding projects, like building scripts to formulate minute-song-snippet playlists for “power hours,” a drinking game in which participants drink a shot of beer every minute for an hour. But he had a hard time focusing on subjects for which he could find no real-world applications, which included most of his classes and many of the questions tech companies asked during interviews.
When Curtis had interviewed at Google—which Fortune magazine named the best workplace for millennials in 2015—it had been a five-hour-long process. He had stood nervously in front of a whiteboard as various managers filtered into the room, asked him esoteric questions unrelated to the work he would actually do on the job, and watched him draw and explain his answers. He performed so poorly in his daylong interview that shortly after he had begun, he already knew he’d failed. It felt terrible.
Gigster’s interview, he was relieved to find out, would follow a completely different process. It would be conducted via a typed chat. Rather than esoteric mind games designed to test theoretical knowledge, like the ones he’d completed while interviewing at other tech companies, all of Gigster’s questions related directly to whether or not Curtis would be able to do the job. The company had no obvious reason to care if Curtis was a “culture fit,” had growth potential, or worked well on a team. If he worked for Gigster, he would complete tasks alone. Only his current skills would matter.
Curtis answered questions like “If you had to implement [a particular piece of code], how would you do that?” “Ok, what if that didn’t work?” This was the kind of problem solving at which Curtis thrived. He nailed the interview. Gigster invited only 7.7% of applicants to take gigs, and Curtis was one of them. Now he had to decide whether to take the non-job.
Though programmers on Gigster worked on projects for startups, freelance work didn’t come with the same chance to hit it big that often lures people to early-stage companies. That chance belonged to employees who were paid partially in equity.16 But joining Gigster did seem like a more interesting opportunity than sitting in a corporate office trying to pass the time.
Curtis wasn’t one to quit his job immediately based on the promises of an unknown freelance service. Before he joined Gigster, he needed to do some due diligence. He’d already saved enough for a year’s worth of expenses. Now he consulted an accountant. He researched health insurance plans and found that COBRA, a US program that would allow him to continue his current insurance plan, was too expensive. Purchasing the same plan that he’d enjoyed as an employee would, without his employer’s contributions, cost him about $600 per month. Through the health exchange that had been set up as part of the Obama administration’s Affordable Care Act, he found plans that would cost only $200 to $300 per month. On a spreadsheet, he laid out this cost, what he wanted to contribute monthly to his retirement savings account, and his expected taxes, which would double once he made the switch from employee to independent contractor. Then he browsed through Gigster’s website, which listed available jobs and their compensation, to estimate how much he would need to work in order for Gigster to be viable.
At the end of the equation, he realized he could take home almost as much pay working as an independent contractor for Gigster as he did at his full-time job, around $10,000 each month. The gig economy struck him as a logical step between starting a company, which he wasn’t ready to do, and working for the man, which he hated.
On a Friday in September, Curtis poked his head into his boss’s office and told her that he was quitting to join the gig economy. His manager asked if there was anything that could make Curtis stay. There wasn’t. The company gave him a choice about whether to work for two more weeks. Curtis decided he’d rather leave immediately. On his way out, he stopped at the company cafeteria and filled his backpack with free peanut butter bars, jalapeño chips, and a “shit-ton” of oatmeal packets. Liberal snack offerings had been the best part of his job.
The next day, a Saturday, Curtis ordered his usual Starbucks coffee and got to work.
He was free.
CHAPTER 3
THE BEST OF BAD OPTIONS
Kristy Milland, a mother of one living in Toronto, turned to the gig economy not out of a desire to become a millionaire or to leave her full-time job, but out of desperation.
While Kristy didn’t have a college degree or access to capital, she did have a genuine entrepreneurial spirit: a willingness to try new things, to hustle, to start a business, and, when that didn’t work, to start another business.
Many years ago, after dropping out of high school just before the birth of her daughter, she had worked to stretch a welfare check as her husband looked for jobs: She’d made spreadsheets showing which stores offered the best prices on every household item, repurposed leftovers, shopped with coupons, and bought baby clothes at the dollar store. Eventually she’d signed up for one of the first online education programs in Canada and finished her high school degree from home.
When her husband finally found a job at a temp agency, she taught herself how to build websites. She opened a daycare center. In addition to these ventures, she invented odd jobs like selling Winnie the Pooh Beanie Babies and Swarovski crystals—salvaged from garage sales—online. More unusually, she built fan websites for everything from kids’ toys to television shows.
The most successful of these sites was a reality television fan forum she built from scratch. Kristy stayed up nights to watch live video feeds from the house where Big Brother was filmed (available online as part of the promotion for the show) and reported new developments among the house’s residents before the edited episodes aired on television. To monetize the site, she taught herself how to sell ads and sold subscriptions.
Getting by became easier when a temp position Kristy’s husband had landed at a Nestlé factory turned into a full-time job. (According to Kristy, he’d been making $11 per hour while, to his best knowledge, the temp agency collected around $17 per hour for his work.) Nothing changed about the work itself: The temp work was so similar to the full-time work that most managers didn’t know which workers fell into which category. But his new hourly wage, nearly double what he’d made as a temp worker, combined with Kristy’s odd-job entrepreneurship, allowed them to live more or less comfortably for 11 years.
And then, in 2007, the recession hit. Nestlé sold the factory where Kristy’s husband worked, and Kristy knew she’d have to figure out finances all over again.
A survey commissioned by the Freelancers Union and Upwork in 2016 found that 20% of full-time freelancers in general don’t have health insurance, compared to 10.3% of the non-elderly general population in the United States that went without insurance at the end of 2016.1
But in Canada, where Kristy and her family lived, losing her husband’s health insurance wasn’t quite as dire—healthcare was publicly funded. Kristy and her husband weren’t going to miss out on seeing a doctor regularly because he’d lost his job. Still, prescriptions weren’t fully covered, and they had ongoing health issues that, without his additional insurance, would cost them $250 per month (Canada’s supplemental support for prescription drugs would only cover what they spent after the first $1,500).
Kristy’s husband would be paid 11 months of severance—1 for each of the 11 years he had worked at Nestlé. But after that the family would need to have another source of income.
The Big Brother website did not make enough money to live on, and though Kristy was about to sell it for a modest sum (a wise move, considering that Twitter and Facebook were about to take over as platforms on which strangers discussed television shows), it wouldn’t be the kind of sale that could support a family. Kristy had closed her daycare when she moved into a new neighborhood. And selling garage sale finds on eBay wasn’t going to cut it.
In the few customer service jobs Kristy had taken at call centers years earlier, she had not enjo
yed working under someone else’s rules. Slowly, this preference for working at home had turned into a lack of experience working outside of the home. She hadn’t waited on tables, had no experience in fast food, and had not learned any skills that might be particularly useful in a factory. She’d once applied for a job at McDonald’s. Nobody had called her for an interview. Jobs were more difficult to find after the recession, and Kristy felt unqualified or too inexperienced for most of them.
But Kristy was someone who always figured it out. When I met her in 2015, she was in her late thirties, with long, blond hair that she sometimes wore tied in a scrunchie. She was not at all physically imposing—she owned a pair of baby pink crocs—but she had the sort of determination and strong belief in her ideas that would later inspire an academic researcher to describe her to me as intimidating. As a website moderator, she had picked fights with television channels that wanted to shut her down for spoiling show secrets. And as a daycare provider, when she’d been stiffed by most of her clients the week before Christmas (because they were also struggling to make ends meet), she had marched to each of their homes with a letter explaining that she could not afford her own Christmas dinner and demanding payment. If Kristy couldn’t figure out how to make a living, it wouldn’t be because she hadn’t tried, and it wouldn’t be because she wasn’t a fighter.
Her first idea was to work more hours on Mechanical Turk.
Founded in 2005, Mechanical Turk is an online “crowdsourcing” marketplace run by Amazon. Its clients post work tasks on a dashboard that a “crowd” of workers can choose to complete. The process doesn’t work that much differently than Gigster’s process. But the tasks on Mechanical Turk are often simple and pay just cents each. They’re jobs like adding tags to images, filling out spreadsheets with contact information, or writing product descriptions for websites. While programmers who sign up for gigs on Gigster are called “remote talent,” workers who take small gigs on Mechanical Turk are called “crowd workers.”
Because individual tasks on Mechanical Turk are often so simple and low paid, many assume that they’re completed mostly by foreign workers who have a lower cost of living than people in the United States or Canada. But a UN International Labour Office report found that often crowd workers are more like Kristy. In the survey, which included workers on Mechanical Turk and a similar site called Crowdflower, 85% of respondents were Americans, 36.7% had a college degree, and 16.9% had a post-graduate degree. Almost 60% said they were unemployed before starting their gigs as crowd workers.
Kristy had started using Mechanical Turk when it launched in 2005 as a way to earn extra income in her downtime. She’d used the money to buy birthday presents and prizes for competitions she ran on her Big Brother forum. Now, in 2007, she would join the 28% of independent workers in the United States and 32% of independent workers in Europe who, according to the McKinsey Global Institute, have forgone the traditional job out of necessity rather than choice.2
* * *
To understand why Mechanical Turk exists, it helps to understand that the way technology “learns” is a bit like how a child learns. Teaching a young child to identify a cat by describing one isn’t a great strategy. A cat has a tail and four legs. It has two small ears. But that could just as easily be a bear. If you really want a child to pick a cat out of a lineup, you need to find some cats and point them out. The child will start to build a mental blueprint for “cat,” and eventually she’ll recognize a cat when she sees one, even if it differs slightly from examples of “cat” that she’s already experienced. It’s more or less the same with machines. You could try to describe, say, a shoe, in a piece of code, but it’s much more effective to say “Computer, here are 10,000 pictures of shoes. Now build a model for identifying a shoe.”
One of Kristy’s early Mechanical Turk jobs was to label thousands of pictures of garments and shoes with their colors. If presented with a photo of a blue shoe, she added a label that said “blue.” If presented with a photo of a gray sweater, she added a label that said “gray.” In this case, Mechanical Turk was Amazon’s way of finding thousands of examples of each color so that it could train its algorithms to automatically sort searches for “blue shoes” and “gray sweaters.” Improving technology in this way was one reason that Amazon had created Mechanical Turk.
Another reason was to compensate for technology’s shortcomings with human intelligence. Amazon created one of its first applications that used Mechanical Turk in the days when people could email via cell phones, but couldn’t yet access the internet. The idea was that people on their “mobile email” could send questions, such as “Where is the best restaurant near me?” and receive an answer almost immediately. It seemed like magic, but workers like Kristy were at the other end, Googling and answering for a penny per question.
Amazon had not launched the Mechanical Turk platform with a promise to create jobs, the way that Uber had early on bragged about adding “20,000 new driver jobs” to the economy every month. Rather, it had built the website as a way to integrate human intelligence with code—as a service for programmers. TechCrunch’s founder wrote shortly after the product launched that. “Amazon’s new Mechanical Turk product is brilliant because it will help application developers overcome certain types of problems (resulting in the possibility for new kinds of applications) and somewhat scary because I can’t get the Matrix-we-are-all-plugged-into-a-machine vision out of my head.” He called the workers who would be doing the work “Volunteers.”3
Early characterizations of Mechanical “Turkers” often portrayed them as people who were playing a game or passing the time while they watched television, which echoed the way that staffing agencies had portrayed temporary workers in the 1950s and 60s. As the UN International Labour Office report pointed out:
[The staffing industry] chose to promote the view that the work was done by middle-class housewives who were looking to earn “extra” money while still fulfilling their household duties. In 1958 the Executive Vice-President of Kelly Girl described “the typical ‘Kelly Girl’” as someone who “[doesn’t] want full-time work, but she’s bored with strictly keeping house. Or maybe she just wants to take a job until she pays for a davenport or a new fur coat.” Similarly, the temporary agency Manpower wrote in 1957 that temp work is “ideal for a married women with responsibilities that do not permit her absence from home every day of the week.”4
This idea that early temp workers were just working for fun, to pass the time, or to buy frivolous items wasn’t accurate. In a study of the temp industry from the early 1960s, 75% of women cited “to earn money” as the primary reason they worked, the majority saying they needed this money to meet daily needs. Decades later, in a tech-fueled era, the same portrayal wasn’t quite accurate when it came to Mechanical Turk workers, either. In one survey, 35% of Mechanical Turk workers—not an insignificant portion—said they used the platform as their primary income.
Kristy hadn’t ever attempted to make substantial income on Mechanical Turk before her husband lost his job. But she had found a forum for Mechanical Turk workers that would help her learn how to be more efficient while working on the site. In fact, she was already the lead administrator—which wasn’t surprising to anyone who knew her well.
Since she had been a teenager, Kristy had gravitated toward pockets of online chatter. Her father had been one of the first computer engineers, and he’d introduced her to the internet while she was a high school student in the 90s. By today’s standards, the early internet was comically slow. Kristy used a modem to dial into a local BBS (bulletin board system) and leave messages or take a turn in a game. The BBS only had one phone line—which meant only one person could be on the site at a time—and so she would log out and periodically check back to see if someone had responded.
Kristy had been a goth kid in high school—purple hair, piercings, black clothes—and she didn’t have many friends in her class. But the forums she dialed into had local numbers, which meant that
the people who she met online lived nearby. She liked them. They were all at least weird enough to seek out this odd new technology, and they were all at least smart enough to figure out how to use it. Kristy started attending GTs (“get-togethers”) to meet members of her message boards offline. They played laser tag and paintball and went bowling together. It was her social scene.
Now the Mechanical Turk forum that she moderated, called Turker Nation, had a similar collaborative, friendly vibe. Workers, many of whom by this point she considered friends, shared tips and helped each other get the best work. Through them, she had found the short computer code that turned the Amazon product classification task—what color is this item?—from two clicks of the mouse (one to select a color, one to hit “submit”) to a much faster single tap of the keyboard. She could hit the “y” key when an item was yellow, and the submission of the answer would be automatic.
A novice could not turn tasks like labeling images into a meaningful income. But with the help of the forum, it might be possible. And, Kristy figured, Mechanical Turk was available immediately, while looking for a job might have taken months. It didn’t require her to have traditional experience, and it allowed her to make a living the way that she always had preferred—by figuring it out on her own at home.
Kristy started reading up on how to succeed on Mechanical Turk. In the meantime, she took two buses to a pharmacy across town where the filling fee was $4 instead of $14 and bought bigger-sized, cheaper pills that she cut down to the correct dosage with a knife. She stopped going to the dentist. When she noticed that the local Rib Festival was handing out samples of her husband’s heartburn medication, she went back for the free medication the next day. And then the next three days after that.
* * *
When Terrence Davenport first heard the gig economy gospel of independence and flexibility in 2014, it sounded like the answer to his literal prayers.