Crowdsourcing Jobs to a Worldwide Mobile Workforce
An ambitious startup strives to create a business based on paying poor people to do microtasks on their phones.
A few years ago, Nathan Eagle had a big idea. What if millions of people in poor countries—people who couldn’t find work in their local economies—could become a remote workforce for organizations all over the world? And what if, instead of traveling to do such jobs at call centers or other outsourcing offices in big cities, they could do their work quickly, reliably, and easily through text messages on their mobile phones?
Eagle founded a small startup, Txteagle, in Cambridge, Massachusetts, to put the idea to the test. It has struck deals with mobile-phone carriers around the world to pay workers in credits for mobile airtime. In many places, that’s as good as cash.
But while the concept sounds promising, expanding the business has proved difficult. Eagle told Technology Review this summer that his venture is “going to be binary—a huge hit or a spectacular failure.”
One big challenge is to find valuable tasks that can be completed through text messages and phone calls. Eagle got the idea for the company after he created a service that let nurses in the coastal Kenyan village of Kilifi send text messages to tell central blood banks how much blood their hospital had on hand, so its supplies could be refilled more efficiently. Simply compensating the nurses for the cost of their text messages turned out to be the key to its success.
He launched Txteagle in Kenya and eventually had 10,000 people doing part-time tasks such as filling out surveys for international agencies, translating text, or collecting address data for business directories. One of his first partners was Nokia, which paid local people to translate mobile-phone menu functions into the 60 languages used in the country. But that task was quickly exhausted.
Now Txteagle needs to form several solid partnerships with multinational corporations that could supply a steady stream of small tasks. Eagle believes one promising idea is to use Txteagle as a market-research tool: workers could be paid to help companies learn what sorts of products would be desired in their rural corners of the world.
Txteagle recently announced a collaboration with the United Nations, which will use the mobile-phone platform to survey up to 500,000 people in 70 countries about their local governance. That brings the number of countries with Txteagle workers up to 80. The U.N.’s goal is to lay the foundation for future disaster-response efforts by learning how well communities and their governments communicate with each other. People who complete the survey will be paid about $1 and reimbursed for the cost of the text message.
For the U.N. initiative, Txteagle is working with the Global Network for Disaster Reduction, a nonprofit organization that influences policy in more than 90 countries. Most nonprofits operate on a relatively small scale, says Terry Gibson, a project manager at GNDR, but Txteagle allows them to reach a significantly larger audience.
Txteagle isn’t the only company exploring ways to crowdsource small tasks to people all over the world. In 2005, Amazon launched its Mechanical Turk project, which sets up a way for a large group of distributed workers to participate in jobs like identifying elements in a set of photographs or performing data entry and transcription. A San Francisco-based startup, CrowdFlower, collaborated with nonprofit organizations this year to have people translate and map text messages that were sent from victims of floods in Pakistan and the earthquake in Haiti. Lukas Beiwald, CEO of CrowdFlower, says his company compensates its workers through PayPal and, in some cases, with virtual currency like the money used in Second Life.
The fundamental technology behind Txteagle includes algorithms for quality control, so that people who do consistently accurate work make higher wages. Workers who recruit others are paid small bonuses. To generate revenue, the company takes a tiny fraction of certain paid transactions.
To make real money with this business model, however, Eagle will need millions of workers using the platform. For now, he estimates, about 100,000 people will be using Txteagle to make money by the end of U.N. survey. And he hopes to find enough partners, with enough of the right sort of small tasks, to push those numbers even higher. “We’d like to be the largest knowledge workforce in the world,” he says.
Kate Greene and Nathan Eagle are coauthoring Reality Mining: Using Big Data to Engineer a Better World, to be published by MIT Press.