According to Braden-Harder, the bulk of Appen’s business is with two key customers.
“We’re basically talking about Facebook and Google here,” she says. And now those companies are taking a hit as the global economy slows the growth of digital advertising and Apple’s new privacy features also hurt Facebook’s ad revenue.
Put simply, Appen’s major customers sneezed and Appen caught a cold.
Analysts covering the company have reacted brutally, and none more so than JP Morgan’s Bob Chen, who cut Appen’s valuation to just $3 late last month. Appen has not traded at these levels since 2017.
“The largest clients are beginning to feel the impact of the weaker macro [economic conditions] and have started to reduce capital expenditures, and this has resulted in a significant decline in Appen’s core revenue, and we have limited opportunities for when this could improve,” Chen said.
Meanwhile, Macquarie analysts have cited another potential downside risk from competitive pricing pressures, as well as the risk of Big Tech reducing its reliance on outsourcers like Appen.
And Appen’s troubles obscure another crucial issue of its future: the ethics of the crowdsourcing in which it is involved.
The problem came to a head earlier this year when the company featured prominently in an MIT Technology Review series. The series explored the idea that the AI sector is creating a new colonial world order using crowdsourcing platforms in a race to the bottom to find – and exploit – low-paid workers around the world. It was titled: How the AI Industry Benefits from the Disaster.
It focused on data labeling platforms like Appen and the millions they amass for doing this work – the so-called “ghost workers”. These workers flag data for the tech giants on small work packages that earn equally modest payments. The viability of apps and competing platforms depends on their ability to allocate and pay for this work with as little human intervention as possible.
It pits appen against workers for a share of every dollar earned. In fiscal 2021, Appen had total revenues of $447.3 million ($671.2 million). It paid out $268.4 million for crowd-labeling services, but the average wage for its million-plus workers that year was about $268 (US$391).
Appen also competes against other data labeling platforms that scour the globe for the cheapest labor. If it’s general work that can be done anywhere, then “definitely a race to the bottom,” says former Appen boss Braden-Harder.
It was one of the reasons she left Appen shortly after the IPO, as pressure had increased to maximize investor returns.
“I kind of knew it was going to be bad. There was pressure,” she says.
“I knew that with this business model, there weren’t too many options for a CEO to give Australian investors what they seemed to be looking for.”
The MIT series examined how these platforms invaded after Venezuela’s economy collapsed, which pushed the middle class into poverty and spurred demand for job opportunities. Venezuela’s economic meltdown created the magical combination of a desperate but well-educated workforce and an internet connection.
Oskarina Fuentes Anaya was one of many forced to turn to Appen as their only source of work. She fled Venezuela to Colombia. Her situation was made worse by a chronic illness that limited her job opportunities, but Fuentes soon learned what it was like to have her life ruled by the platform algorithms that ensured the most economical distribution of work to Appen’s million-plus employees.
“We all help each other,” Fuentes told MIT of the support these workers gave each other to share what little work was available.
The MIT story chronicles the wage cuts, the desperation to grab the dwindling available work, and the suspension of accounts — which also prompted wage locks with limited recourse to a human operator from the platforms.
“What began in Venezuela has created an expectation among AI industry players of how little they should pay for such services, and has created a handbook on how to meet the prices customers are counting on,” it said in MIT history.
While data labeling has offered workers like Anaya a lifeline, it has also exposed them to Darwinian levels of exploitation as platforms slashed their wages and locked accounts – and livelihoods – in an ongoing race to the bottom.
Dangers include rigorous customer reviews that can result in account suspension, as well as ambiguous tasks and administrative errors that can result in an account being suspended for months.
Julian Posada, an associate professor at Yale who has studied these crowdsourcing services in South America, says there is a huge power imbalance that favors platforms that have the power to set their own rules. You can literally scour the globe for cheap labor to do these menial jobs.
But Venezuela’s educated population, great infrastructure from before the oil collapse — provided a rare combination of ingredients that made it perfect for these outsourcers, Posada says.
“So on the one hand you have the infrastructure for the work. On the other hand, you have people who are in a crisis with the highest inflation, so you can pay them as low as possible,” says Posada.
In the beginning it was good work.
To build a viable network of contributors, these platforms offered bonuses and, in one case, even paid these outsourced workers an hourly rate. But once they reached critical mass, many of those payments disappeared and wage rates fell.
In one case, a platform Posada investigated inadvertently left their payment details for thousands of workers in a public Google spreadsheet.
He says it gave a clear picture of the relationship between rising viewership and falling wages.
“The more people that came in, the less people made,” he says.
As Venezuela’s situation slowly improves with higher oil prices, the trick will be to find the next low-wage job market with enough people desperate for work.
“The next time there’s a country in crisis, they’ll probably be there as long as there are computers and desperate people,” says Posada.
According to the MIT story, Appen has begun to highlight its handling of its crowdsourced workforce, which includes the company’s code of ethics.
It cites an internal survey of 7,000 workers late last year that showed 17 percent were long-term unemployed before joining Appen and 16 percent lived below the global poverty line. 63 percent used appen earnings to support their household or to pay for education.
But another character said so. In its annual report, Appen reported on the survey, which showed that 67 percent identified appen as their main source of income.
In response to questions, Appen said, “We are committed to fair pay and ethical treatment of our crowd. Our Crowd Code of Ethics specifically states that it is our goal to pay our Crowd above minimum wage in every market around the world in which we operate. To support our customers, we have a fair pay feature available on our platform.”
Appen also adjusts its per-task compensation to the minimum wage for the worker’s location. This means that workers from a poor country are paid less than someone from a wealthier country for the same job. In the MIT story, Appen said it had seen a rise in scams where users had used VPNs to access higher salary offers in other countries.
Braden-Harder, for example, is not impressed by talk of the minimum wage, which is set by each state in the US and tends to be very low.
“You can pay the legal minimum wage and still pay poverty wages,” she says.
Posada cited a recent fair work project that examined working conditions across all crowdsourcing platforms and found that none of them met minimum standards. But Appen was the best of an evil gang.
“It’s like the best of the worst. They have some standards, they have some rules,” he says.
Braden-Harder has retired from senior roles and currently serves on the Advisory Board of Santa Clara University’s Global Social Benefit Institute.
She supports global start-ups like one in Kenya led by an Australian university graduate who provides school lunches.
“I think we all, myself included, believe that companies can do something good, but you have to have the right business model,” she says.
When it comes to solving the crowdsourcing problem, Braden-Harder says big companies need to change their mindset when it comes to procuring these services.
“In my experience, sourcing is the bad side of any business because the same guy who buys toilet paper for big businesses is also buying those services.”
Get news and reviews on technology, gadgets and games every Friday with our technology newsletter. Sign up here.