Workforce analytics experts share secrets for finding ROI in talent
Senior managers from three major brands outlined the organizational structures of their analytics teams and shared best practices in a panel at the HR Technology Conference.
Predictive analytics has been the rage in human resources, with executives using it to anticipate turnover and formulate remedies. Now some companies are taking their workforce analytics to the next stage and analyzing the economic returns on workforce-related investments to identify the highest-yielding talent-management strategies.
That was the premise of a panel at the recent HR Technology Conference & Exposition in Las Vegas. Workforce analytics leaders from Chevron, Morgan Stanley and Wal-Mart revealed their experiences and best practices, including how they structure their analytics and data-science teams, and which methods work best with certain problems. Moderator Brian Kelly is president of Vestrics, a maker of cloud-based, workforce analytics software based in Chapel Hill, N.C.
Establishing a workforce analytics function
RJ Milnor, manager of planning, analytics and reporting at Chevron Corp., an energy provider based in San Ramon, Calif., said he got into workforce analytics after serving as an investment banker. He became fascinated with how small adjustments in ways managers interact with their workforces could change financial outcomes. “It’s very much like changing a financial portfolio: We change the portfolio of our people,” he said.
In 2009, Milnor launched a team at Chevron to deliver key performance indicators (KPIs) and tools requested by HR, such as headcounts, attrition rates and dashboards. The team has since more than doubled in size and shifted its orientation from basic reporting and KPIs toward business questions faced by executives.
That led to creation of a workforce analytics center of excellence (CoE) consisting largely of project managers and analysts — generalists who gather data and build dashboards. Nowadays, the 15-person CoE maintains a “workstream” on workforce planning and advanced analytics staffed by specialists in attrition, hiring and employee surveys. “The very narrow focus allows them to knock out reports when they need to, and also have the bandwidth to be able to explore deep into that terrain,” Milnor said.
The CoE is not part of HR. “We have conversations with HR all the time, but we’re actually a separate business unit,” he said.
Elpida Ormanidou, vice president of global people analytics at Wal-Mart Stores Inc., the giant retailer based in Bentonville, Ark., said she moved into HR in 2008 after working in customer operations. An 11-year employee, Ormanidou was asked to build a workforce analytics team from scratch.
The department, now called people analytics, is “a pretty autonomous function,” she said. “You can lift it and put it with any discipline and it could work just like that, because all the people are technical experts in one analytical discipline or another.”
Wal-Mart also has a CoE that Ormanidou heads, managing a staff of around 30. The CoE provides analytics, including social media analytics, visualization and data modeling and mining services. Ormanidou also manages the point-of-sale analytics team, also around 30 people. She reports to the senior vice president of people strategy, who reports to the chief people officer.
For Jeremy Shapiro, executive director in HR and head of global talent analytics at New York-based securities firm Morgan Stanley, workforce data crunching centers on learning things about employee populations can help senior managers make more informed decisions.
“When you have a good day, you can actually influence them in a way that’s positive,” said Shapiro, who reports to the global heads of talent and compensation reporting.
Five years ago, when he started the analytics effort, Shapiro learned from the large department that researches financial services and other financial instruments — the “primary storytellers” inside Morgan Stanley. “If we couldn’t tell stories in the way that our executives were used to hearing stories, we knew that we wouldn’t have as much of a credible argument,” he said.
A sizable group in Morgan Stanley had used business intelligence and data warehouses for more than a decade, but the workforce analytics team is separate and consists of statisticians and business analysts who handle project-based research into specific problems. “Storytelling is with our generalists, who help build the skill of both data-driven decision making and working with clients on making the numbers on the page jump to life,” he said.
Reporting vs. analytics: What’s the difference?
Shapiro, who co-wrote a 2010 article on talent analytics in the Harvard Business Review, said he walked around the conference exhibit hall and found that most vendors claiming to have analytics really had reporting. “There was some analytics that you would think of as either some type of advanced statistical methodology, some big data, some machine learning,” he said. “There was a bit of that sprinkled in, but the analytics was being used as a marketing phrase.”
Reporting nonetheless has its uses. “There’s nothing wrong with reporting,” Shapiro said. “Reporting tells great stories and can tell people why in lots of different ways.” But statistical manipulation is often needed to get at why.
Milnor said it’s more important to know whether a particular action aligns with the business strategy. “For so long in HR, we’ve struggled with what the [return on investment] is of HR programs,” he said. “We have, in many ways, the methods and now the technology to really get at that, and because we have that, we’ve lost sight of what we really need to be doing.”
He said many of the best questions ask: What, so what, and now what? “When I look at analytics maturity models, sometimes there’s an implication that we should be predictive or prescriptive all the time,” Milnor said. There’s a need for those two, but many times the question is simply, What?
In workforce planning, the what could be simply headcount, attrition rate or traditional financials. But true optimization tries to answer more difficult questions, such as which skills are undergoing a transition or that can help replace skills gaps that arise from employee movements and are usually addressed by hiring.
Ormanidou said her take on the reporting, analytics and optimization breakdown is that business managers tend to want reports: “If you let them, they will drown in an ocean of reports and they will keep asking for more,” said.
For example, when David Scott became Wal-Mart’s head of HR, he wanted the exact count of the roughly half million out of the 2.2 million workers who leave every year. She told Scott she could run such a report for every payroll period, but what he would do differently with the information?
Analysts can get lost in building models, Ormanidou said, but an optimized workforce is more likely to come from connecting data points in new ways. “You create new information that people haven’t thought about before, and that’s where the magic happens,” she said.