Follow Datanami:
January 28, 2014

Hiring with Hadoop: How Big Data is Helping HR

Alex Woodie

Talk long enough with someone in charge of hiring, and you’ll hear the importance of gut instincts. After all the forms have been filled out and tests have been taken, the hiring manager will often dismiss them and go with his or her gut instinct. Thanks to new data analytics services, such as those offered by Evolv and PeopleAnswers, data-driven measures are displacing the subjective approach.

Evolv’s flagship software as a service (SaaS) offering, called Evolv Selection, is a Hadoop-MapReduce application that uses big data and machine learning algorithms to help large consumer-facing companies hire the right people for a given position. Carl Tsukahara, the company’s executive vice president of marketing and product, compares the product to the Sabermetrics approach that Oakland A’s GM Billy Beane made famous in “Moneyball.”

“That’s what Evolv does, but for corporate America,” he says. “We use high power predictive analytics to build the best team at the best budget. It helps companies ask themselves, ‘Am I acquiring the right players and putting them in the right positions for them to succeed? Do I have the right way to mange them and incent them?'”

Evolv uses MapReduce to crunch terabytes of structured and unstructured data on behalf of its customers every night. The company runs one giant data reservoir on its Hadoop cluster that includes data from all of its clients, including data from enterprise systems (HCM, CRM, and ERP); the latest compensation data from the Bureau of Labor Statistics; and sometimes even data from social networks. With hundreds of millions of data points, the company is able to draw broad conclusions about American workers in general, and yet make very targeted predictions about likely employee performance by region, industry, and individual company.

Attention from Big ERP

Another company putting HR analytics on the map is Infor, a $2.8-billion provider of ERP software for midsize and large companies. Earlier this month, the company acquired PeopleAnswers, which develops software that helps organizations identify which employees are more likely to succeed.

PeopleAnswers uses a test to identify 39 behavioral attributes, such as discipline, ambition, energy, and empathy. The results are then weighed against particular job requirements, and presented to the hiring manager.

PeopleAnswers says using analytics to match people to jobs leads to lower turnover and higher revenue among its customer. One large retailer that used the software in some of its stores reported a 30 percent increase in sales and lower turnover compared to the stores that didn’t use it, the software company says.

“Evaluating talent is one of the most important decisions for a manager but these decisions are hindered by uncertainty and a lack of data,” Infor CEO Charles Phillips says in a press release. Through this acquisition, we will not only expand our HR offering, but we will add a cadre of scientists to move beyond simply automating processes to push the envelope on big data and HCM innovation.”

Manage for Exceptionalism

Today, Evolv officially unveiled its second major Hadoop-powered SaaS offering, called Evolv Insights. The new offering is designed to provide data-powered recommendations about how to manage employees to reach a business goal.

“For a given employee there are hundreds of potential knobs you can turn: How much to pay them, who they should report to, what shifts they work, what product line should they be assigned to,” Tsukahara tells Datanami. “Our conjecture is historically companies try to turn those knobs by guessing.”

For example, say a large grocery chain was looking for a way to improve their employee attrition rates for stores in the Southeast region. After all the data is loaded, Evolv Insights would present the customer with the 10 most important factors affecting attrition in that region.

“We’re not going to tell them to raise their salary by $5 an hour,” Tsukahara says. “We’re going to tell them, ‘Salary is the number two predictor in that region because you’re not paying people the right rate to get the best performance on customer satisfaction.'”

Minimizing employee attrition is an important business goal because of the high cost of finding replacements. Other possible HR goals that Evolv helps customers reach include driving higher customer satisfaction (CSAT), higher profits and revenues, and maximizing customer net promoter scores.

Follow the Numbers

Hiring and managing employees to reach one goal will sometimes help them reach several goals. For example, Evolv has clusters of customers in the telecommunications and cable TV business, which are notorious for iffy customer service. A cable TV technician who misses a scheduled appointment by five hours may help the company more if he’s transitioned into a less visible role.

“We believe that customers very much want to figure out how to hire and manage their workforce in a manner that will improve the ultimate business outcome,” Tsukahara says. “They want to hear customers say ‘Wow that guy sure did a great job and I’m really happy with the service I’m getting and I’m going to refer him to my friends.'”

Following the numbers sometimes leads to head-scratching recommendations. For example, Evolv has found that, in some situations, ex-convicts make excellent hires who have the potential to stay with a company for years. What’s more, it’s usually a very bad idea to re-hire former employees (at least in the consumer-facing companies that Evolv works with). And in some customer service jobs, high turnover is actually a good thing.

“It’s extremely counter-intuitive, but it is the data and the predictive analytics that are providing these insights,” Tsukahara says. “This is based on each of our clients’ business performance and HR data that clearly says, by the way, these are the most important predictors that will drive your ultimate business outcome.”

Datanami