Until approximately three years ago, Néstor Ares, HR analytics and processes unit chief at the Inter-American Development Bank, would often muse about the irony of working at an organization whose very essence is analytics while the HR department provided only “seat-of-the-pants” answers. The Washington-based bank, which provides development loans and grants to Latin American and Caribbean countries for economic and social-development projects, depends for its success on correctly analyzing financial factors such as loan performance and making predictions about the viability of proposed capital projects. Accordingly, every business unit at the bank had an analytics team … except HR.
When it came to questions about such things as employee performance and engagement and future human capital requirements, Ares says, HR might ask a few people who worked in the area for their best guess. “They may have been correct,” he says. “But without numbers to back them up, no one could be certain.”
Three years ago, however, Ares was given the go-ahead to develop a data-analytics unit within HR. The unit could leverage the vast amount of data the organization collects in order to provide quantitative analysis about trends, future requirements, the performance of HR initiatives such as training and engagement efforts, and other strategic issues.
That unit, which is still evolving, currently includes nine people, each with different experiences and skills. It contains two IT people, who create the analytics programs and the interfaces that allow business managers, executives and others to create their own reports. There is also a person experienced with database management systems who is reformatting and restructuring data to enable different systems to share data.
The unit also contains statisticians, who develop the back-end formulas that power the analytics programs. Other members are skilled at creating readable and understandable charts and tables for prepackaged and ad hoc reports.
And, importantly, there is one person whose background is communications. “We needed someone to explain to people outside HR what we can do for them as well as someone who can help our team collaborate,” Ares says.
IADB’s analytics unit is gradually moving HR’s function in the company from one solely focused on tactical issues, such as recruiting, training and people management, to that of being an important partner in the organization’s business process. And, in terms of analytics, it is moving from strictly qualitative judgments to quantitative reports.
Developing an internal HR-analytics team is not an easy process. It requires finding and hiring people whose skills are entirely new to HR. But many experts believe the ability to do in-house analysis is well worth the effort.
Matt Stevenson, workforce strategy principal and partner at New York-based Mercer, says that for some smaller companies, farming out data-analytics functions may make financial sense. But when feasible, an internal team has the advantage of understanding the organization’s business and its specific HR requirements.
“The internal HR group knows how the organization operates better than an outside group could,” Stevenson says.
“Democratization of Data”
A second option, purchasing an HR dashboard, will speed up the reporting process, since business managers can do their own report generation. But dashboards generally focus on current conditions, not on analytics. Says Stevenson, “The questions [that managers want answered] change so frequently that to use a dashboard for that purpose is a bit like building a plane while you’re flying in it.”
Companies that create an analytics team will usually include a dashboard so managers can have a view of the current situation. But the advantage of analytics is that it can identify problems early on in any process, before they would be apparent in a pre-packaged report. To cite just one example of this, the data-analytics team at IADB, in creating an analysis of hiring bias in the 26 countries the bank serves, discovered to everyone’s surprised that in one country, if current trends continued, employees from that country would be underrepresented. That led to HR working with the executives at that country’s office to rectify the situation.
In developing his team, Ares says, he used a phased approach, gradually adding personnel and functions over the last three years. He attributes the relatively smooth rollout to the fact that, “we were comfortable and confident with one set of capabilities before we added new ones.”
Al Adamsen, executive director of the Talent Strategy Institute in Santa Cruz, Calif., agrees that a phased approach can often help companies avoid pitfalls in the development of an HR-analytics unit. He says that, very generally, there are three levels of HR data analytics that companies can move through. Version 1.0, as he calls it, is the ability to create prepackaged reports, such as charts and graphs, on an event-driven basis. In this version, “when a department is having a problem–say, with retention–it calls HR, which creates the relevant data analytics,” he explains.
Version 2.0 is what Adamsen calls “the democratization of data.” It includes user applications that allow departments to create their own reports and even capture new data sets. Version 3.0 adds more predictive elements, determining, for example, what types of people are likely to leave the company or what types of engagement programs are most likely to be effective.
As HR-analytics teams mature and companies move through the versions, says Adamsen, the skill sets of the unit will begin to look less like a traditional HR team and more like number crunchers, with team members composed of statisticians, database professionals and data analysts.
It might also be helpful to add to the unit a person or small group that specializes in surveying and industrial and organizational psychology. Doing so will allow companies to gather more data to add to its analytics mix, he says.
The more data the company generates, the more accurate the reports. And adding new employee surveys will usually improve the analyses. However, everyone interviewed for this article says that the problem for most companies lies in organizing and analyzing data, not in gathering more. “We’ve found that in virtually every implementation project, organizations have plenty of unused data on hand,” says Mick Collins, global vice president of workforce analytics and planning at SAP America Inc. in Newtown Square, Pa. “The primary challenge is mining it, not finding more of it.”
Analytics Are Everywhere
Obviously, an essential factor in the success of an internal HR data-analytics team is the quality of its people.
Because IADB is a financial institution, Ares had a large pool of internal employees with analytics experience from which to seek potential members for the HR team, although he also sought people from outside the organization.
The rising number of companies that depend on business analytics means that even companies that aren’t in industries traditionally considered analytics heavy–such as manufacturing and retail–may have large numbers of data analysts on hand who could serve as an internal talent pool for HR-analytics teams, says Collins. Internal candidates, he adds, are often the best choice.
“Ideally, people on your HR- analytics team will have intimate knowledge of the company’s business,” he says. “So internal people should be prime candidates.”
In addition to having experience with the organization, or at least the industry, team members should also have HR experience, says Rishi Agarwal, principal at the people-analytics group at PwC in San Jose, Calif. Although he admits that recruiting people with industry, analytics and HR experience can be a tall order, having such people on your team can bring invaluable benefits.
“Analysis in HR is very specific to that function,” says Agarwal. “It’s very different from financial analysis, for example. I believe it’s easier to train an HR person with some analytics background to be a full-time HR data analyst than to train a business analyst to be an HR professional.”
Another quality Ares sought when he was hiring for his unit was the ability to collaborate. In his experience, he says, many younger workers who haven’t spent years in the older silo system of business functions are better able to learn to speak one another’s language and work together, although Ares acknowledges that’s a generality that is far from true in all cases. However, when he’s interviewing someone strong in data analysis, he says, he’ll try to assess how willing and capable the candidate will be to learn at least the language and basic functions of software engineering.
Finding people is one problem, but convincing them to work in HR may be even more challenging. Ares says recruiting unit members requires a bit of explaining and selling. “A lot of the people in our unit–the analysts, the programmers–never expected to work in HR,” he says. He had to convince them that they’d be doing the same type of work they would expect to do if they took a position in other business units. They also had to be assured that their careers would be advanced by their stints in HR. “We showed them that we’re moving from traditional HR functions to becoming strategic, and that the organization respects the department’s ability to provide that type of analysis,” says Ares.
Many experts say that respect can be key in making the difference between a successful implementation or chaotic failure in an analytics unit. “If you’re going to develop this team, you have to take the project seriously,” says Agarwal. “That means getting buy-in from executive and business units and committing the planning and resources it needs to succeed.”
Centralized or Decentralized?
Although finding and hiring the best people for an HR-analytics team is essential for success, another consideration is how the company will organize them. “Having HR analysts in-house will be of little value if they’re not used by the business functions the team was created to serve,” says Agarwal.
And while under-use can doom an HR-analytics team to irrelevancy, overuse can strain the team’s resources to the breaking point, especially when it’s still young. For example, if anyone at the company can request a report, the team may be burdened with redundant projects. The solution, he says, is to create the organization and work-flows that enable the team to work efficiently.
Agarwal says the first factor to determine is the operational model. The HR-analytics unit can be centralized within the HR department; decentralized with HR people working in the business units they serve; or a hybrid model, in which there’s a centralized HR-analytics team that supports HR analysts working within the business units. The decentralized and hybrid models are most easily developed when the analyst began in a business unit and was moved to HR. On the other hand, if the analyst was hired by HR and has no experience with the business units, the decentralized model may require a cultural shift that could complicate implementation of the HR-analytics project.
The second consideration is the governing structure. Which departments within the company can ask HR data analysts for reports and analysis? Does HR create reports it expects business units will need, or does it wait for the units to request reports? A new HR data-analytics team may prefer to take the lead in initiating reports, since its capabilities may be evolving and it may not be initially capable of serving everyone in the organization–although there is no general rule of thumb that will fit all organizations, Agarwal says.
Finally, there’s the organization chart: No matter which departments are allowed to request reports from HR, as a matter of simple logistics, the company needs to develop work flows of how the requests are routed and who has to sign off on them.
The job of creating an internal data-analytics team cannot be taken lightly. Careful planning and choosing among the many options can increase the chances of a successful implementation.
