When faced with the challenge of adding 250 people to its 600-person team in a year’s time, Opower—a customer engagement platform for utility companies that’s since become a subsidiary of Oracle—knew it needed data on its side.
To predict the number of recruiters needed to fill these roles, Opower crunched the numbers and presented company executives with three possible scenarios:
- Continue with its current team of seven recruiters at no additional cost, but risk missing its goal significantly for 30% of roles.
- Invest $350K into a combination of tech sourcing sites, additional recruiters, recruiter bonus programs, and referral bonuses, and cut the risk of missing its goals to 15% of roles.
- Invest $750K into the heavy use of employment agencies, and reduce the risk of missing its goal significantly to just 5% of roles.
Without these numbers, the risk of continuing with the status quo would have jeopardized 75 of its needed jobs. But because the company had concrete data to work with, it was able to reduce risk to an acceptable level, while minimizing its investment through the cost-effective $350K alternative.
Not only did Opower decrease its average time-to-fill from 93 days to 67 days through data-driven decision-making, it met its recruitment goal with a total of 237 new hires.
HR, Meet Data
In the Opower example above, the team used a combination of predictive and prescriptive analytics. First, the team extrapolated from the data available to them to determine what the likely outcomes would be if different variables were changed (in this case, the amount and method of investment). Then, the company’s executives made an informed decision about how to move forward.
However, predictive and prescriptive analytics aren’t the only applications of big data for HR. In fact, there are four analytics practices HR professionals can adopt, including descriptive, diagnostic, predictive, and prescriptive analytics.
Michael Corcoran, senior vice president and CMO of Information Builders, suggests an easy way to understand the differences, based on the key question behind each type of analytics:
- Descriptive analytics: What is happening?
- Diagnostic analytics: Why did it happen?
- Predictive analytics: What is likely to happen?
- Prescriptive analytics: What should I do about it?
Don’t worry. Your company doesn’t have to master all four disciplines at once. Whether you have an in-house data team or zero employees dedicated to analytics, your HR team can reap insights from data by starting small and evolving as your data skills grow.
Here’s how to get started.
Think of the four types of analytics as building on one another. Descriptive analytics is your foundation. You can’t use diagnostic, predictive, or prescriptive analytics practices until you have a solid understanding of what’s happening within your company.
In a SlideShare presentation, People Team director Akshay Raje suggests seven HR functional areas that can—and should—be translated into measurable descriptive analytics:
- Performance and career management
- Compensation and benefits
- Organization effectiveness
For instance, data points for recruitment analytics could be broken down into the subcategories of recruitment, internal movement, and staffing effectiveness:
Image Source: SlideShare
The entire presentation is worth a review if you’re new to the use of big data for HR. However, don’t be overwhelmed by the number of metrics suggested there. Select five to 10 to start with that most closely align with your department’s biggest challenges or definitions of success. You can always add to your descriptive analytics efforts later on.
Diagnostic analytics provides context to the information gathered by descriptive analytics, suggesting not just what is happening, but why specific trends are being seen. An article by Neocase Software presents the example of recruiting and retaining a large sales organization.
“In our example, it could be a graphic report that that shows the ranked reasons why salespeople have departed,” the author states. “The reasons may range from low quota attainment to higher base salaries offered by competitors.”
Diagnostic analytics isn’t something you extrapolate after gathering your descriptive data. The two processes can occur concurrently, depending on your organization’s needs.
According to Greta Roberts, CEO and co-founder of Talent Analytics, “You don’t need to complete analytics projects in a linear fashion. It’s different for each organization. Two levels can blend at times, occurring simultaneously. Some levels may take longer to complete. Other levels may have smaller amounts of data, but be enough.”
Though multiple levels of analytics can operate simultaneously, it’s hard to advance to using predictive analytics in an HR capacity if you don’t have that solid foundation of descriptive and diagnostic data.
Once you have a good feel for what’s happening and why, however, you can begin to be more strategic about leveraging your metrics to predict future outcomes.
Prescriptive analytics—essentially the ability to make informed decisions through data analysis—is the gold standard of big data in HR. Yet, as Data: The New Language of HR reveals, only 4% of 480 large companies studied by Deloitte have reached the point where they’re able to perform predictive, let alone prescriptive, analytics.
A Pixentia article gives an example of what prescriptive analytics looks like in practice. Imagine that a manufacturing company’s descriptive data reveals that machinist attrition is rising. From there, imagine that:
- Diagnostic analytics finds that overtime increased by 12% over the last year, and that, while machinists want some overtime, they don’t want too much.
- Predictive analytics posits that increasing the number of machinists will reduce overtime, though doing so will cause FTE spending to increase.
- Finally, prescriptive analytics might suggest that the optimum amount of overtime is 2.25 hours per week, and that limiting overtime to this amount will optimize overtime vs FTE spending.
With this information at the ready, HR team members can confidently suggest a plan with the potential to reduce machinist attrition without incurring unnecessary costs.
Launching an Analytics Program
Pulling together a program that incorporates the four levels of analytics described here requires an investment, but it may be less than you expect. For example, though there are expensive programs that automate most of these analyses, Opower conducted its projections primarily using an Excel spreadsheet.
Further, you don’t need a mountain of data to get started with each type of analytics. Simply start somewhere, and enlist the help needed to get your program off the ground. Tap into freelance data analyst talent if your current team lacks these skills (and if you aren’t ready to invest fully in an in-house data specialist).
For more guidance on launching an analytics program, check out the free Data: The New Language of HR ebook. Not only will it help you understand your company’s current level of analytics maturity, it’ll give you all the tools needed to build and sustain the culture needed to use data to inform your current and future business decisions.
Download Data: The New Language of HR now to get started.