Taking Analytics Up a Notch
Employers are discovering new uses for workforce-data-analytics products so they can detect problem areas and possible solutions far more easily and effectively than before.
By Larry Stevens
As HR departments have become increasingly digitized, especially over the last decade or so, the amount of data companies hold about their workforces has become prodigious -- and, too often, relatively useless.
While the reams or screens of data -- often in Excel tables -- inform experienced HR executives who use them to solve problems and develop corporate workforce strategies, the analytics capabilities have lagged behind the availability of the data.
For example, when Brad Church, HR business partner at The Williams Cos. Inc., an energy entity based in Tulsa, Okla., gave a presentation to his internal stakeholders (who -- unlike at many companies -- include board members, executives as well as HR managers), most of the narrative came from his expertise and not from statistical analysis of the data.
"We were flying by the seat of our pants," he says. "We had the data, we digested the tables the best we could; the rest was pretty much intuition and gut feeling."
Intuition and gut feel in the hands of an experienced HR professional are valuable assets. And Church, who has no data-analytics training or background, was able to provide insightful answers to some thorny questions. But, he says, he recognized that "most of the people we were talking to wanted something that could carry more weight than just my best judgment."
About three years ago, The Williams Cos. implemented data-analytics software from SuccessFactors, an SAP Company, based in San Francisco. This is one of a small handful of products that integrates workforce data from most or all corporate, and often online social media, sources and applies a statistical process called regression analysis, a method for determining the relationships among data points, finding causal relationship and making predictions about the future needs of the company. The products use graphical interfaces to insulate users from the underlying statistical process.
As a result of this software, Church says, his conversations with company stakeholders have become less descriptive and more predictive. For example, Williams has an aging workforce with a higher-than-average retirement risk -- something Church has had many discussions about using companywide statistics for. Now, with SuccessFactors implemented, he complements the standard companywide reports with retirement danger zones, areas where retirement -- while not a current risk -- may become a problem in the future.
Similarly, with discussions about overtime, Church was always able to make presentations on overtime levels and then drill down to problem departments, resulting in suggestions that some managers should consider adding full-time workers. But now, the report not only indicates departments where overtime has reached a preset trigger point, but also where overtime moves outside the average range for the entire organization, or when there is a significant change in the amount of overtime -- either higher or lower. These may indicate areas where overtime may be an issue in the future.
The new system has even caused a shift in the way managers and executives measure performance at the company. In the past, it used a very centrist approach -- recognizing the top producers in the entire company. Now, with access to statistical analysis, the company can view performance based on many different categories such as years of experience, amount of training taken, type of recognition given, time since last promotion and more. It can also recognize people who, while not yet top producers, have improved more quickly than their co-workers.
"What we're getting now," says Church, "is a view of performance life-cycle. The performance discussion has morphed from one in which we only want to recognize top producers to one where we also find those factors that create the top producers." Soon, the company may be able to identify hires that are most likely to become top producers, he adds.
Ironically, Church has found that, while his presentations are more data-driven, the conversations that accompany them are more of a narrative rather than dry and difficult-to-digest numbers. "When I give a presentation now, I feel I'm telling a story instead of pointing to a lot of figures on a chart," he says.
As companies implement products such as this one, they often find new and creative ways to provide stakeholders with an accurate view of the company's workforce, its risks and opportunities. Says Josh Bersin, principal and founder of Bersin by Deloitte in Oakland, Calif.: "Compared to only a few years ago, a lot more talent-related data is available to companies. Now, with these analytic tools, HR can help their companies turn that raw data into actionable information."
Separating Truth from Fiction
While a forward-looking view of the company can help it prepare for the future, data-analytic products can also enable companies to ferret out false positives and false negatives, which can sometimes lead companies to engage in unproductive and even maladaptive reorganizations.
Mark Sullivan, strategic vice president for human resources insights, analytics and operations at New York-based McGraw Hill Financials, and a data-analytics specialist, was made aware of this problem early on.
"We found that, not only were people coming up with the wrong answers, they were also asking the wrong questions," he says. For example, he received a frantic call from an HR executive who wanted him to use data analytics to find out why the company was losing its female population more rapidly than its male population. "He was expecting me to come back with facts and figures that showed that, say, women were being compensated at a lower rate than men, or that they were being promoted more slowly."
What Sullivan discovered surprised everyone. "The data showed that, over the last few years, we had done a great job of recruiting women; so, obviously, with a higher female population than we ever had and with the majority of that population being newly hired, that cohort would exhibit a higher turnover rate," he says. So, instead of a knee-jerk reaction to, for example, encourage managers to increase compensation or promotion of women, costly efforts that might have yielded little effect, the company improved its onboarding process and loyalty programs for all workers. But executives also realized they might simply have to wait a few years until the female population caught up with the male population in terms of seniority before the turnover disparity would disappear.
As well as this situation went, Sullivan knew he couldn't provide that level of data analysis to the entire company through his office. In fact, in most cases, HR managers believe they know the problem, so they won't even turn to him. The solution, he determined, was to make data analysis more self-service. "We had to provide data in a consumable way," he says.
Being a global-analytics company, McGraw Hill Financials executives have a lot of respect for data analytics. But when it came to HR, the shoemaker's children had no shoes. "It was long past time to bring HR into the data-analytics fold," says Sullivan. So he and others at the company decided to implement products from Visier in San Jose, Calif., which provides different ways of viewing data and understanding its significance without requiring users to have an understanding of the regression analysis that is under the product's hood.
HR professionals at the company had different levels of sophistication -- and tolerance for -- data analytics. So Sullivan uses the tool to provide analysis in three ways.
On the most basic level, he provides interactive slides that allow users to visualize data through graphs they can manipulate using onscreen slides, tabs and buttons. So if, for example, they want to look at the entire company by region, gender, age or some combination of categories, they can do so with only a few mouse clicks on self-explanatory screen icons. They can then click on buttons and tabs to drill down to specific parts of the company. Sullivan says more than 400 data points are now encompassed through these slides.
On a higher level, users can select from a preset series of questions, each one leading to another set of questions. For example, they can select "show me what my population of a particular type of worker is." The next set of questions asks about the same data, but by region, for instance.
Although Sullivan is not allowed to reveal many of the pre-set questions his system offers users, Visier reports that many of its customers typically use such questions as: Are we hiring the right people? Who is at risk of leaving? Are we keeping the right people? How many key employees will reach retirement age in the next five years?
At the highest level, used by only a few executives and managers at McGraw Hill Financials so far, users of the product are allowed to create their own questions from scratch.
Other data-analytics projects focus on helping managers know where to focus. For example, Michael Thurston, an HR analyst at Salt Lake City-based Intermountain Healthcare, uses products from Tableau Software, based in Seattle, to create "heat maps." These alert HR to problem areas, such as a department that has high first-year turnover, higher costs or longer-than-average time to hire. Once a problem area is uncovered, users of the software can drill down to determine what might be causing it.
Of course, much of the data was available before installing Tableau, but finding problem areas in a set of five or six Excel worksheets, while possible, was arduous and error prone. "The power of visualization is very strong. The heat maps give a means of discovering areas that need attention," Thurston says.
For example, like most organizations, Intermountain Healthcare wants to reduce its vacancy rates and its requisitions for new employees. Once the data became more graphical, Tableau users at the company were able to correlate two problem areas: departments where workers tended to have fewer hours and those that had higher turnover rates. The correlation of these two heat-map items revealed that the fewer hours people worked, the more likely they were to leave the organization. So, while onboarding and ongoing training were, of course, always important, this new analysis allowed the company to work on increasing the number of workers' hours, which has resulted in a gradual reduction in vacancies.
The visual data has also helped with goal tracking such as it relates to retention, participation in wellness programs, vacancy rates and so on. In the past, managers usually gauged how well they were doing relative to their goals through monthly reports, a time period that made it difficult for them to institute mid-course corrections. Now, dashboards display daily progress in a way that can be digested with only a glance. And, because departmental goals can be displayed on any dashboard, the system has led to some healthy competition. "It's easy to see how departments are doing compared to other parts of the organization," Thurston says.
Despite the obvious benefits of the tool, Thurston has had to do some evangelizing to get managers to use it. "People naturally navigate to the familiar; and in our case, that's Excel," he says. "Like all new things, the dashboard approach takes a little time to get comfortable with."
Also, as with any tool, if the software is not taken full advantage of, the results will reflect that. While even new users find the tool useful, managers willing to spend time and "get their hands dirty experimenting with the dashboards get the most out of it," Thurston says.
Mollie Lombardi, president of the workforce-management practice and principal analyst at talent-management company Brandon Hall Group in Delray Beach, Fla., agrees that part of the job of implementing workforce-data-analytics products is encouraging people to use them as fully as possible. She points out that one sticking point is a culture that associates problems with poor performance on the part of the manager.
"At some companies," she says, "managers are cautious about what they ask the systems for, for fear of what they might find." Companies that have successful implementations of these products are those that reward managers who uncover problems rather than use those problems as a cudgel, she adds.
Data-analytics products for HR are still in their infancy. Many managers are dipping their toes in the workforce-analytics tide to see what these products can do. The more they are used, the more ways managers and HR leaders may find to ferret out and then solve their unique business problems.