HR Analytics Trailblazers
Some companies are leading the way in doing HR analytics right by turning data into action.
By Tom Starner
When Leonard Dandurand arrived at Electronic Arts Inc. eight years ago, the odds of seeing data analytics within EA HR was about as likely as having a placekicker grace the cover of the company's wildly popular Madden NFL video game.
Within the past few years, however, data analytics has come alive at Redwood City, Calif.-based EA, a global provider of interactive entertainment software. Dandurand, group director of HR information services and workforce planning, says EA had been slogging along with a combination of an old human resource management system and a custom front end to provide basic transactional services to users.
"While we had no shortage of data about our workforce, reports were limited and difficult to produce," he says.
As the developer of many of the world's most well-known digital video games, such as The Sims, EA SPORTS FIFA and the aforementioned Madden NFL, innovation holds a much esteemed spot within the EA culture. Eventually the HR department needed to catch up. Today, EA has morphed into being a serious HR data analytics innovator in managing its highly intelligent, talented -- and some might say quirky -- workforce.
"We in HR have transformed from a central-services team with a support function to a key strategic adviser," Dandurand says. "And a good part of it is due to using analytics."
EA isn't alone in carving out a place among data-analytics superstars. BAE Systems Inc., a Mt. Laurel, N.J., defense contractor, has in the last several years become increasingly focused on infusing data analytics into its workforce-planning processes. It currently has a dedicated workforce-intelligence team leading the effort, according to Carol Darling, vice president of workforce analytics and HR compliance, talent acquisition ops and global workforce planning.
Darling says the BAE Systems team uses employee data to track critical roles in each of the company's business areas, and delivers comprehensive data on hiring, attrition, retention and demographics -- all of which help business teams assess specific talent strengths and weaknesses.
"Our dashboard is incredibly robust with drill downs and just about anything you can imagine," she says. "That helps our business lines manage their workforces. For instance, we can tell them generationally what our workforce looks like and how it will change over the next few years."
EA and BAE serve as prime examples of how successful HR organizations are taking advantage of data analytics that go beyond the basics, turning themselves into analytics pacesetters in the process.
Helen Friedman, global leader of human capital analytics at Willis Towers Watson, says there are many more ways that organizations can get talent analytics "wrong than right." She says that, in her experience, trailblazing employers across industries getting it right follow a strategic focus, use a scaled approach to data, develop clear ownership, deploy fundamental analytical capabilities and, most of all, keep it simple in the early stages. By following that formula, an organization can both support business strategies and effectively manage talent, she says.
"The key is to start where value can be added and data issues can be managed or, better yet, minimized," Friedman says.
For example, rather than trying to capture every possible data source for a turnover analysis, employers might consider the top 10 perceived factors driving turnover that would come from a single data source and see to what degree turnover can be explained using only a subset of data. By careful framing of the issue and the applicable group being considered, HR can hone in on an issue without getting trapped by what many call "analysis paralysis."
"Just one word of caution," Friedman says, "there is a tipping point at which the dataset or group becomes too small and you can no longer see the story -- you do need some scale to make the approach statistically meaningful."
Matt Stevenson, a partner in Mercer's Workforce Strategy and Analytics group, divides effective data analytics into two primary buckets. The first, more basic approach requires collecting and reporting data and then using benchmarks to make decisions. The second is making the data actionable -- predicting, for example, turnover risk across the workforce or how to decide how much overtime to pay versus hiring new people, or measuring performance-based pay.
"The real HR-analytics superstars make data actionable, using it to predict any number of outcomes to improve the culture, keep attrition down, meet business strategies, etc.," Stevenson says.
"The classic function we see is around employee turnover," he adds. "Using analytics, you can, for instance, predict an individual's turnover risk. And there are some factors you can do something about -- training, intervention, how many people [a manager manages]. Whether they are remote or not, can manage turnover . . . rather than blundering around in the dark."
Fiona Jamison, CEO at Spring International, a research-based HR consulting firm in Conshohocken, Pa., says what differentiates her firm's clients that are doing some of the more sophisticated analytics is they have an executive who says: "I want to make more fact-based decisions. We are seriously looking to eliminate some of the subjectivity [involved in] business-critical talent decisions." For example, Jamison says, it's a common notion that people leave employers because they don't like their manager, but can it be proven?
"Executives are starting to demand the 'why' when it comes to outcomes," she says, "That's a primary thing that separates our most savvy clients; where a senior leader has said he or she wants more accurate data to back up the talent-based decisions they are making."
Jamison adds that it also helps to have someone inside the HR organization who truly understands analytics, someone with a statistics background who can connect the dots.
"For many in HR, asking them to manage analytics is like asking them to bake a soufflé, when all they've ever baked is bread," she says, adding that her firm says professionals with both analytics and HR savvy are known as "purple unicorns" because the blending of those two areas of expertise are so rare.
What separates HR-analytics trailblazers from the pack? Jamison cites visionary leadership, disruptive thinking, clear business alignment, being comfortable with exploration and failure, and collaboration among the top traits. Stevenson adds factors such as investing in people and processes (not just tech tools) to move from data processing to data inquiry, the ability to differentiate what is useful from what is simply interesting and, finally, having a clear set of critical questions to be answered, along the lines of: "We don't get data unless we know what decision we will inform with it."
Taking It to the Next Level
To launch its journey towards data-driven workforce decision making, EA's workforce-analytics team began delivering extensive, hard-coded, monthly workforce reports to HR and business leaders, prompting both to go through extensive data clean-up efforts and gain a better appreciation of HR data.
According to Dandurand, the lengthy reports were a hit with leaders who were looking for basic workforce facts to help them understand and plan for their talent needs. Over time, however, the number of complex weekly, monthly and quarterly workforce reports that the team produced grew exponentially -- encompassing a variety of trends for hiring, talent movement, terminations, demographic shifts in the workforce and other employment factors.
EA's workforce-analytics team could not keep up with the data demand, and manually building reports was a reactive exercise. Also, the team's lack of a flexible reporting environment limited its ability to slice, dice and drill down to the right level of insight needed to move from workforce observation to diagnosis, according to Dandurand.
EA's search for a flexible solution to optimize workforce data came on the heels of implementing Workday's new HR management system. EA chose Visier, a Vancouver, B.C.-based provider of cloud-based HR analytics, as a complementary tool that would help HR produce deep, trended analytics about its workforce. It proved a decisive move, Dandurand says.
Visier was chosen because it offered all the predefined HR metrics required, and because it utilized Software-as-a-Service. According to Dandurand, the gaming business is a rapidly evolving industry, one that has had to keep pace as a "packaged goods" economy has migrated to a digital one. For EA, that meant moving from shrink-wrapped software to online gaming.
Apart from using Visier to deliver workforce reports to business leaders and HR business partners, EA, as part of the onboarding process, set up data from Workday to load automatically into Visier, allowing the HR analytics team to begin digging into the workforce data immediately. EA and Visier collaborated to replace all the standard, manually generated monthly and quarterly workforce reports with online presentations that leaders could access on demand.
Dandurand conservatively estimates that Visier has saved the company about 24 weeks of labor per year -- half of a senior analyst's work year. The time savings, he says, stems from the platform's ability to automate manual report generation and quickly deliver analytics that would otherwise take much more time to produce.
In another quick, well-received initial win, Dandurand says, EA HR analytics built detailed sensitivity analyses related to average hiring and termination trends -- showing the probability of either exceeding or meeting fiscal-year headcount plans. The EA analytics team also began delivering headcount-management reports across company leaders, HR business partners, talent-acquisition partners and finance/planning partners to better integrate cross-functional headcount-planning processes.
"The value of workforce intelligence grows as organizations mature in their use of it," says Lexy Martin, former Sierra-Cedar data analytics guru and principal of research and customer value at Visier.
Martin says initial value is often calculated in terms of cost efficiencies gained through the adoption of a new technology solution. The next step is typically measuring the improvement of key HR metrics -- for instance, reducing unwanted turnover. Finally, organizations that are the most advanced in the use of workforce intelligence can connect their people strategies and decisions to the improvement of key business outcomes, such as revenue, profits, customer satisfaction, margins or industry-specific metrics.
Building a Better Dashboard
BAE's Darling says her company's HR predictive-analytics effort grew out of its compliance team, because compliance data is, by its nature, very "clean" (read accurate), so it's a logical starting point. BAE formed a core group of HR people who were very familiar with the data.
"At that point, the analytics space kind of expanded within the HR industry and we were asked to start tackling more complex issues," Darling says. BAE expanded its core team and then upgraded some of the team's skill sets over time.
"We have a couple of Ph.D. statisticians on board, along with some business-intelligence professionals," she says. "In a really short period of time we were able to start delivering products to our user community, tools that they could gain insights from and make better decisions with." For example, BAE Systems uses predictive analytics to determine flight risk in the employee population, creating flight-risk profiles. With that, one group was able to decrease attrition by 20 percent using tactics based on the analysis the HR team provided.
BAE started off with a relatively simple dashboard and then built up from there using a combination of Alteryx, a self-service analytics platform, and Tableaux, a visual analytics platform -- a successful combination.
"Our team is often asked to give briefings at the Tableaux conference or the Alteryx conference [based on] what they've done with the data," she says. BAE has even won awards at both user conferences.
Being in the defense industry means BAE has an aging workforce compared to other sectors (due to the nature of the skill sets among its workforce), so HR analytics are used to forecast what its workforce might look like five years out.
"We triangulate the data to figure out the average retirement ages for our employees. Next, we develop heat maps to identify what areas might be at most risk," she says. "We also added programs to encourage people to stay with us longer."
Darling's department is focused on helping BAE businesses with respect to segmentation of their critical workforce skills. It can identify what is already in the talent pipeline and what their organization could potentially look like in the future if it does nothing differently. It also overlays data with a macroeconomic analysis, checking into the demographics of specific geographic locations, as well as skill sets available in those locations and the predicted future availability of such skill sets.
"It helps inform if they need to 'buy or build' [their] way out of potential gaps or holes related to talent," she says. "We also are getting a better grip on the cost of attrition. We have a variety of different types of people doing different types of work, whether that means engineers inventing things or service people who work on-site with our government customers.
"One ultimate goal is to eliminate guessing when it comes to talent planning," she says.
It's no surprise that IBM is among the elite in using predictive analytics within its talent-management strategy, according to Elissa Tucker, research program manager for human capital management at American Productivity & Quality Center (APQC), a member-based nonprofit HR research firm in Houston. Tucker collaborated with Talent Analytics on a research report, Getting Started with Predictive Workforce Analytics, which focused on five employers, including Armonk, N.Y.-based IBM, in their efforts to creatively take advantage of predictive analytics.
"IBM is doing a number of innovative things," Tucker says.
IBM's HR function includes a dedicated predictive social-analytics team. One of the team's first projects was to use social media to get a real-time understanding of employee engagement. As part of its report, APQC interviewed N. Sadat Shami, manager of IBM's Center for Engagement and Social Analytics.
IBM's social-analytics team found that it could account for 48 percent of the variability in employee-engagement scores by analyzing social-media-data use among employees. Shami told APQC that the number represented a "really strong effect," so this result gave IBM confidence that there is significant internal social-media data related to employee engagement.
The IBM team also created an analytics tool called Social Pulse, which uses IBM employees' "social-media sentiment" to predict if engagement is increasing or decreasing as a result of IBM HR initiatives.
"To get action, we try to make our analytics as consumable as possible. One way we do this is by communicating analytics results visually," Shami said in the interview. "We realize that, if business leaders can quickly understand analytics, then it helps them drive action. Behind our visuals are a lot of advanced analytics, but these don't need to get displayed and maybe don't even need to be mentioned."
One small but meaningful Social Pulse success story singled out in the APQC report resulted in changing an unpopular IBM HR policy. Driven by concern for safety and reduced insurance risk, IBM would not reimburse employers for use of popular ride-sharing services such as Uber and Lyft. Unhappy with the policy, IBM employees posted a petition to change it on Connections, the company's internal social-media platform.
Within 24 hours, the petition received more than 100 "likes" and 50 comments, and the Social Pulse tool alerted the analytics team, which then notified the company's chief HR officer. IBM's CHRO convened a meeting of relevant people who decided that IBM would reverse the policy and reimburse employees using ride-sharing services -- again, fairly inconsequential, but proof that accurate data in any form can affect an HR policy or address a challenge quickly.
"Whether assessing Sierra-Cedar survey data from early adopters hearkening back to 2000, or analyzing the results of Visier customers such as EA today, it's clear that quantified organizations outperform others," says Martin.