How One Company Learned to Use Analytics to Make Smarter Hires

By Aliah D. Wright Dec 1, 2015
Reuse Permissions

TYSONS CORNER, Va.—Want to hire people faster, and have them stay longer and be more engaged?

Use analytics.

That was the message delivered recently during RecruitDC, a conference for recruiters held at USA Today headquarters here.

Representatives from Opower, which provides cloud-based software for the utility industry, told recruiters how the company started using analytics to track, evaluate and interpret data to improve recruiting.

“When Opower began deploying analytics to measure how we recruit and hire people, amazing things happened,” said Dawn Mitchell, Opower’s director of talent acquisition.

This led Scott Walker, senior people analyst, to crunch the numbers. He said he was overwhelmed at first.

“When I joined Opower as a sourcer, I knew little about recruitment metrics,” he said, adding that he could barely “sort columns in Excel.”

His experience mirrors that of most people entering the field of data analytics for the first time, he said. According to research firm Bersin by Deloitte’s report, Talent Analytics: From Small Data to Big Data, 75 percent of HR executives say they are aware that using analytics is an important driver in the success of their organizations. Yet, 51 percent have no formal talent analytics plan in place. What’s more, close to 40 percent say they don’t have the resources to perform analytics, and 56 percent rate their own skills in workforce analytics as poor.

As Josh Bersin, founder of Bersin by Deloitte, wrote for HR Magazine earlier this year, “if you don’t start, you’ll never get there. This level of investment initially feels like a big expense, but in a very short time the analytics team may pay for itself.”

Citing Deloitte research, Walker told attendees that most organizations are at one of four levels in their workforce analytics journey. He said 56 percent of organizations are at level one and that “it’s very basic … they’re just getting general reports.” Thirty percent are at the level two, which is advanced reporting. They’re “pulling metrics, benchmarking … keeping track of performance and trying to put things in context, but not solving a problem.” For example, they may be pondering such questions as: “How has our time-to-fill [jobs] changed over time?”

Level three is “proactive analytics,” Walker said. In this phase, “they’re looking at the root cause of an issue and solving talent challenges through data and statistical analyses. For example, they may be asking, ‘How do we staff our team for constantly shifting hiring needs?’ So they’re using data to make decisions.”

Level four is predictive analytics. Only 4 percent of recruiters are at this stage, Walker said. They’re using data to “forecast future talent outcomes.” In the course of doing so, they’re asking such questions as: “How much attrition will we experience next year, and how much money do we need to eliminate time in empty seats?”

What Opower Discovered

One of the main things Opower discovered was that it was important to first know what problems the company wanted to solve, and not to crunch data just for crunching data’s sake.

However, those using analytics to make talent decisions need to understand that “level one and level two never, ever go away,” Walker said. “People will always have a need for basic reports.”

Citing research from insight and technology company CEB, Walker added that, “Organizations with mature talent analytics functions will see a 12 percent improvement in talent metrics overall, 6 percent improvement in gross profit margins, 12 percent increase in employee performance, 10 percent increase in quality-of-hire ratings and a 30 percent higher stock than the Standard & Poor’s 500 index over the last three years.

“You won’t really see the value of your analytics until you’ve mastered the reporting,” Walker said.

“Our first year, we created dashboards for everything,” he added, noting that the dashboards examined passive conversion rates, offer declines, source of hires, time-to-fill, referrals, candidate experience, cost-per-hire, career site traffic, click-through rates and a number of other metrics. Most of the metrics were calculated in isolation without looking at trends, forecasts or benchmarks, he said. “Ninety percent of the time was spent scrubbing the data, and the remaining 10 percent of time was spent trying to make pretty charts in company colors,” he said.

However, he advised conference attendees, “You won’t see value until you use the data to solve problems and make predictions. You need direction from the company” and to know what problems you’re trying to solve. “When you have no direction from the management team at your company about how to use data and what they want to see, you create dashboards that are basically irrelevant.”

One of the things Opower did was devise recruiter scorecards to see how long hiring managers were sitting with requisitions (or job requests). This also gave recruiters an incentive to fill requisitions faster. Recruiters were awarded 0.5 percent of all their new hires’ salaries for the second quarter and “as a result of the bonus program, everyone’s performance went up. [Recruiters] went from 5.2 hires in one quarter to 7.1 hires,” said Opower’s Manager of Talent Acquisition Alan Henshaw.

Another thing Opower discovered was that interviews that included a panel of five or more individuals ultimately yielded candidates who stayed on the job longer.

“When we get to five interviewers, we can significantly predict who will be a top performer based on having a full interview panel,” Walker said.

He added that referrals and former interns “are twice as likely to stay past two years than agency or passive candidates” because “they get the most realistic job preview.”

Do’s and Don’ts of Recruitment Analytics


  • Have a hiring plan, spreadsheet, one system of record and a main dashboard.
  • Always be setting goals, “and track them over time,” Opower’s Mitchell said.
  • Make your analyst an insider. “Empower your analyst. Include them in management, team and strategy meetings. The more they know, the more they can help. If [the analyst] understands what the end goal is, he can come up with a variety of options,” Mitchell said.
  • Construct narratives and ask “why.” Added Walker, “You don’t want to show a graph with no explanation and context.”
  • Summarize takeaways, caveats and relevance. Don’t accept the data as is: Dig, segment and identify causes.
  • Beg, borrow and steal. If you lack expertise and a budget, borrow from finance, sales or operations. “Get their opinion,” Walker said—and their help. Ask to compare your dashboards with the ones they may have in their departments.


  • Waste time on things that don’t matter. Don’t construct “so what?” metrics or excessive dashboards. Teach the entire team to pull basic reports.
  • “[L]et perfect be the enemy of good,” Mitchell said. “The data is never, ever going to be 100-percent accurate.” Consider the impact of the data even if it’s just 95-percent accurate. “Some metrics need to be perfect, others don’t.”
  • Get comfortable. Re-evaluate which metrics are still valuable.
  • Forget to start by using data you already have, Mitchell said.
  • Get discouraged. “Analytics equals delayed gratification,” Mitchell said. “It gets better.”

Aliah D. Wright is an online editor/manager for SHRM.

Reuse Permissions


Choose from dozens of free webcasts on the most timely HR topics.

Register Today

Job Finder

Find an HR Job Near You


Find the Right Vendor for Your HR Needs

SHRM’s HR Vendor Directory contains over 3,200 companies

Search & Connect