BloombergBusinessweek Technology - By Caroline Winter on June 24, 2013
The rate at which you reply to e-mails may reveal your overall competence. At least according to Gigwalk, a company that helps businesses hire on-the-spot temporary workers. “If you reply within an hour, your success rate for completing a job ends up being about 97 percent,” says Gigwalk Chief Executive Officer Bob Bahramipour. “But if it takes longer than five hours, that number drops to 47 percent.”
Gigwalk, launched in 2011 by three former Yahoo! (YHOO) colleagues, has developed several mathematical models that shed light on employee behavior and follow-through. To hone its metrics, the company evaluated more than 300,000 workers who have used the GPS-based Gigwalks app to take nearby jobs doing everything from testing iPhone software to snapping photos of restaurants and nail salons for online-mapping companies.
Gigwalk uses what it calls its “Moneyball engine”—a reference to the book and film about a baseball team built by scrutinizing data to find undervalued talent—to find good workers without ever engaging in personal interactions. “We never meet our workers, never interview them, never look them in the eye,” says Bahramipour. After all, interactions cost time and money, and they don’t necessarily reveal much. “Selection at temp agencies is a somewhat random process,” he says. Gigwalk, by contrast, makes the claim: “We know more about our workers than anyone has ever known about workers.”
Asked what becomes of Gigwalk’s low-rated workers, Bahramipour suggests that his system is actually more forgiving than a human boss—a claim he stakes with reference, again and again, to the book about the Oakland A’s. “Our Moneyball machine manager is dispassionate—it deploys workers rationally based on their skills and observed behavior,” he wrote in an e-mail, explaining that slow but accurate “Gigwalkers” might be given jobs with distant deadlines, while those who are fast but inaccurate might land simple jobs with tight deadlines.
“In this way, Moneyball actually helps workers be more successful,” says Bahramipour in a telephone interview. “Where a human manager might pass up workers because of an apparent flaw, the Moneyball machine manager will pick up that worker and use him or her the best way possible. Just like the Moneyballconcept in baseball.”
Job candidates start out doing simple gigs in which the cost of failure is low. “If we hire someone to take a photograph of a stop sign at a specific intersection and they don’t show up, that sign will be there the next day,” says Bahramipour. Once Gigwalk evaluates qualities such as promptness, thoroughness, and efficiency, it matches workers with appropriate jobs.
The company measures each worker’s productivity by tracking his or her GPS location and the amount of time spent on the job. So if two workers do a similar gig—setting up a store display, for example—but one takes half the time, Gigwalk can easily tell who was more efficient. The company also evaluates how much complexity each worker can handle before running into trouble. The idea, says Bahramipour, is to measure “how well a worker truly can follow instructions—something you could never get from an intuitive interview.”
So far, Gigwalk is active in the U.S., Canada, and the U.K. The app is currently available only on the iPhone (AAPL), but Gigwalk plans to introduce an Android version soon. The service is similar to that offered by TaskRabbit, except that Gigwalk connects companies to workers, rather than connecting individuals. Members of Gigwalk’s “smartphone army” have completed more than four million jobs for 5,800 businesses, including Microsoft (MSFT), BMW (BMW), andEBay (EBAY).
Workers usually get $12 to $15 per hour and are paId via PayPal, says Bahramipour. Gigwalk normally keeps a cut of about 30 percent to 40 percent per job. The private company, which is not yet profitable, also sells analytics that help clients keep better track of work done by temporary hires.