On Reflection: Working the Phone

October 20, 2015

Patrick Marshall


Amazons Mechanical Turk was the first service to allow businesses to connect with hundreds of thousands of anonymous workers around the world and have them handle small tasks for small sums of money. Seattles Spare5 has improved on that model by offering screened experts to perform micro-tasks on a mobile device for micro-payments.

Spare5, which in August received an infusion of $10 million from Foundry Group, Madrona Venture Group and New Enterprise Associates, targets client businesses that have large amounts of unstructured data such as images and videos that need to be tagged, described, classified or keyworded and finds people able and willing to handle the tasks in short segments of time for a relatively low fee.

The cool thing about the human brain is that we are all superfast at synthetic reactions to rich media, says Spare5 cofounder and CEO Matt Bencke. Were all good at pattern recognition. Its really easy for someone in line at a coffee shop or on a bus to look at a photograph and have a reaction.

Those wanting to work for Spare5 in their spare moments download the iPhone app a version is also being developed for Android and respond to some personal questions. We ask for permission to look at some of your Facebook information basic stuff like your age, gender, education level, job history, says Bencke. Then we ask you to take a short survey to tell us a little bit more about yourself. We get to know [people] more based on their behaviors in the app. And we push more and more tasks to people who do well.

Spare5 has employed interior designers and architects to rate photographs of remodels for Porch.com. Getty Images has used Spare5 to find people good with photography to add keywords and title tags for photographs. Spare5 also enlists workers called Fives to take surveys commissioned by clients as well as to conduct surveys, calling businesses to gather information.

Another type of task Spare5 farms out to Fives targets the training of machine intelligence.

Machines are getting smarter and smarter, says Bencke. But if a machine is going to act like a human, it needs a human to teach it. One client is training its computer-vision and machine-learning algorithm to recognize similar styles of shoes in order to improve the shopping experience for consumers. With responses from Spare5s network of Fives, the algorithm gets better at predicting which shoes to offer a given consumer.

Bencke declined to reveal just how many people have signed up, but he says, We have more than we need and more joining every day. He adds that in its first 10 months in business, individual Fives earned everything from a few bucks to a few thousand bucks.

Spare5 may be blazing a trail into a new type of relationship between employers and workers. Its approach to processing and collecting big data of interest to businesses takes advantage of a confluence of trends including the growing power of mobile devices to make it possible for people to work anywhere, anytime; the search among people for new ways to supplement their incomes; and the ability of businesses, through such platforms as Spare5, to use the ubiquity of smartphones to reach a wide array of service providers instantaneously.