Data Overload

By Drew Robb Oct 1, 2003

HR Magazine, October 2003Analytics turn a mishmash of facts into a workforce blueprint.

For the past decade, HR professionals have dutifully entered employee data into HR management systems (HRMS). Every hire. Every transfer. Every departure. Within those databases lies a wealth of data that HR can use to guide workforce planning. But how can value be extracted from all of that raw information?

That was the challenge facing International Finance Corp. (IFC), part of the World Bank Group. Although IFC had some homegrown tools to pull data from its PeopleSoft HRMS system and import it into a spreadsheet, it didn’t have what it needed to extract useful intelligence. Even worse, when different people would run seemingly simple queries, they ended up with conflicting results.

“We lacked an integrated view of our workforce and a historical perspective for doing analysis,” says Joe Fucello, IFC’s human resources program officer. “We needed quick answers to inquiries about our workforce, but we weren’t able to provide those answers.”

But all that changed once Fucello implemented a workforce analytics software package.

“We get a lot of queries such as, ‘What is the average age of people who have left in the last five years?’ ” he explains. “It used to take overnight to run the numbers; now I can have the answer in five minutes.”

Retention Challenge

IFC works in developing countries to provide financing for private-sector projects, assists companies with obtaining financing in international financial markets and advises businesses and governments. Last year, IFC approved $3.1 billion in financing for 204 projects in 64 countries.

Two-thirds of IFC’s 2,000 employees work out of its Washington, D.C., headquarters, while the rest operate out of 80 branch offices. While IFC, like many companies, is concerned about workforce diversity, Fucello and his 40 HR colleagues must also ensure that the organization’s workers reflect more than just the demographics of a single nation.

“IFC has 175 member countries who serve as shareholders and collectively approve the investments IFC makes in emerging markets,” he says. “We have a very diverse workforce of 125 nationalities who are representative of the clients we serve in the developing world.”

Hiring the right mix of staff is one thing; retaining them is quite another. IFC was losing a huge number of its loan officers before they had been on the job for five years, the point at which they become most productive. Fucello knew that to remedy the situation he needed a better way to analyze who left and why.

This is where workforce analytics comes into play. Analytic tools take all the raw data, combine it, compare it and present it in a format that managers can use for planning. Besides presenting historical trends graphically, such products also allow users to run “what-if” scenarios.

“The most useful tools are ones that can show relationships between changes you can make in the workforce and how it affects the business,” says Jane Paradiso, national practice leader for workforce planning at Watson Wyatt Worldwide, a consulting firm based in Washington, D.C. This data must be presented at a high level. “If you get lost in the analysis of small details, that can be a problem.”

Companies have several options in setting up such a system. (See “Analytic Options.”) At IFC, both HR and information technology (IT) staffs conducted evaluations, but HR had the final say. “The IT department helped us with the technical requirements,” says Fucello. “But it was HR’s decision in the end.”

Nuts and Bolts

IFC considered several different products. One, from Cognos Inc., based in Ottawa, was a generic tool that would have required too much development to meet the organization’s needs. Analytic tools from PeopleSoft Inc. in Pleasanton, Calif., and SAP AG of Walldorf, Germany, lacked the full features and functionality that IFC demanded. IFC went with Human Capital Management (HCM) software from SAS Institute Inc. in Cary, N.C.

“We felt it better to go with a pre-built HR solution rather than a generic OLAP [online analytic processing] package that would require a lot of customization,” Fucello says. “This cut our implementation time and gave us a big head start.”

Cost was also a factor in making the decision, Fucello says, but he declines to say what IFC paid for the software. However, Betty Silver, SAS’s solutions strategist for HR, says the product is priced at about $5 per employee for the first year’s license. Implementation typically runs about three times the software fee, with total costs starting around $100,000, she says.

Before beginning full-scale deployment, IFC conducted a six-month pilot of the software to see if it would integrate with the company’s systems and deliver the information needed.

Although SAS’s HCM software comes with many pre-built reports and features, like any other software it required a fair amount of customization. SAS’s Silver acted as lead contact. “It started with us talking with the vice president of HR, Dorothy Berry, about her concerns about retaining loan officers,” Silver says. “They would hire MBAs out of college who ended up using IFC as a training ground before leaving.”

Silver worked with IFC on the strategic side of implementing the analytic software, and with programmers at SAS to write the necessary code. SAS consultants verified the data and transferred it into the SAS database before sending it to IFC for approval.

Fucello led the project at IFC, working with Berry and Rohit Gupta, the system administrator and developer for HRMS. The HR department at IFC doesn’t program PeopleSoft, however. Another part of the World Bank Group handles that, and IFC is an internal client.

Both the SAS software and the web server run on a dual processor Windows 2000 box with 1.5 gigabytes of RAM and 70 gigabytes of storage. Early each morning, a data integration tool from Informatica Corp. of Redwood City, Calif., extracts the data from the PeopleSoft database and refreshes the SAS data warehouse, a process that takes about 45 minutes.

Initially, IFC used a single server for both test and production purposes, but it has since purchased a separate development server, Gupta says. Based on his experience with this project, he strongly recommends two servers right from the start. The SAS software has been trouble-free, but there was a problem with the servlet engine—a Java application that provides server-side processing—obtained from a third party, Gupta says. For now, he has a script in place to work around the log-in failures caused by the servlet. He is planning to replace it with a different servlet.

Garbage in, Garbage out

Analytic results, of course, are only as good as the source data.

“You have to make sure the data is clean, that it is meaningful and that it will fill in all the dimensions that you need,” says Monica Barron, research director at AMR Research Inc. in Boston. “In some cases, what seems to be a straightforward project of implementing HR analytics turns out to be more multifaceted.”

IFC turned out to be just such a case.

“While we were testing the SAS product, we realized our data quality was not where it needed to be to leverage the analytical tool,” Fucello says. “When generating reports, we didn’t get the numbers we thought we would.”

As a result, HR has been working to continually upgrade the quality of the data. Though this is a slow process, Fucello reports that data quality improves every month. One problem, for example, concerns historical data. Corporate officers might have seven or eight different titles over the course of a decade or two because of changes in organizational structure. Although HR employees can’t go back and modify the data in the original transaction system, they can at least create new data in SAS that incorporates all the changes to assist in analysis.

Fucello also notes that IFC’s PeopleSoft HRMS did not provide all the data needed for performing the type of analyses he wanted. So he incorporated some web information sources and a succession-planning module into the analytic system.

“We needed to collect data from a wider body of sources to generate the reports we wanted,” he says. “We now have a more complete view of the information than we ever had before.”

Open for Business

After six months of development, IFC opened up its analytic software to end-users in July 2002. SAS conducted a two-day training program for HR analysts and demonstrated the system to management. Although SAS provided generic tip sheets for users, IFC tailored its own version based on its specific system design. If users want more detailed information, they can access a training warehouse online.

IFC extended use of the analytic program from HR to an additional 70 managers outside the department who needed direct access to personnel information in their own areas.

“We used to spend 80 percent of our time manipulating data and 20 percent on analysis,” says Fucello. “Now we spend almost no time on manipulation and so can accomplish a lot more analysis.”

Users access the system through IFC’s intranet. Each user has a customized home page affording access to HR metrics, reports—HTML or PDF—and tables of data for queries. Since the warehouse is refreshed every morning, users always have access to the latest available data.

“HCM comes with its pre-built reports, tables, queries and data cubes,” Gupta says, “but we have largely discarded the ones that came with the package and authored our own.”

IFC added a daily report for management that condenses 30 key indicators into a two-page report. “They don’t want to look through 30 pages of reports,” says Fucello. “This is in a nice format, it’s easy to read and it refreshes itself daily without human intervention.”

Information vs. Results

Having information available is one thing, but the real test is whether it helps decision-making and produces bottom-line results.

“It’s good to have data, but if you are not using it to change your strategy, it is a wasted effort,” Fucello says.

IFC has put the data to good use. For example, IFC has traced its loan officer retention program down to a compensation system that treated high producers and low producers the same way, as well as to a lack of a clear career path. IFC, therefore, adjusted its career management program to be more explicit about what it takes to move to the next level.

Further, IFC learned that although the organization recruits broadly, a large number of the MBAs hired come from the same eight or 10 schools. This resulted in a change in the recruitment strategy to produce a more diverse workforce. IFC also used the system to evaluate recruitment of mid-career professionals and to evaluate different scenarios for moving employees out of headquarters so they could be nearer to customers.

As a result of its use of hard data rather than just opinions, HR has grown in stature within IFC. “We can bring figures to the table showing the impact of changes that we have made, as well as the impact of potential changes, which gives us more credibility with our corporate partners,” Fucello says. “HR can be viewed as either a cost center or a strategic partner. We have made the leap to strategic partner.”

Drew Robb is a California-based freelance writer who specializes in technology, engineering and business.

Web Extras

HR Magazine: Analytic Options

Tips for Implementing Workforce Analytics

  • Focus on data quality. Good trend analysis typically requires a minimum of three to five years of historical data.

    • Master routine reporting. Having access to—or the ability to create—hundreds of reports doesn’t necessarily add value. Challenge your vendor to condense volumes of critical elements into one or two pages.

    • Understand your user base. Be realistic about employees’ willingness to adopt what can be a somewhat intimidating paradigm for analysis. A one-size-fits-all learning approach won’t work. (Tip sheets are great for certain user segments.)

    • Don’t forget support and maintenance. A fresh approach for analyzing your HR business will likely require a new support strategy. Look past implementation to assess what you will need to keep your workforce analytics system up and running.

    • Consider an HR-specific workforce analytics solution. The analytics marketplace can be confusing to the novice buyer. Deploying an HR-centric analytics solution will likely speed implementation and accelerate time-to-value.

    Job Finder

    Find an HR Job Near You
    Post a Job


    Find the Right Vendor for Your HR Needs

    SHRM’s HR Vendor Directory contains over 10,000 companies

    Search & Connect