Hiring managers use data analytics to increase employee retention

For years, employers have relied on gut instinct, behavioral interviewing and plain old résumés to determine a candidate’s fit for a particular job. And while those tactics are still very much in use, a new brand of analytics -- dubbed 'talent science' or 'workforce science' -- aims to bring more discipline to the recruiting process.

As a component of talent management -- the process of creating a success profile for a role within an organization, making sure the profile is well-communicated to everybody within the organization and outside of it, and hiring and promoting to that success profile -- workforce science shows promise in helping reduce turnover and recruiting costs, with some employers reporting that those candidates who are hired using talent science stay at the job up to 20% longer than those who are not.

In a nutshell, workforce science uses data gleaned from a number of sources -- pre-employment personality assessments, employer data (everything from how long a person’s been in the job to how many vacation days they’ve taken to where they’re located and who their manager is) and publicly available macroeconomic data from government agencies such as the Bureau of Labor Statistics -- to help determine which applicants are more likely to stay on the job, which in turn lowers attrition and recruiting costs.

Also see: Financial services companies renew focus on talent management

Vendors in the space include Infor (with its PeopleAnswers product), SAP (with SuccessFactors), evolv (Evolv Selection), IBM (Kenexa products), Oracle (Taleo Recruiting Cloud Service) and others. Some operate on a pay-per-use basis, while others charge a flat fee based on positions, departments and geography the system is used for.

Retention challenges

Organizations such as The Limited, The Results Companies and many others are using talent science to get better candidates in available jobs -- candidates who stay longer and are a better cultural fit for the organization. For The Limited, a national chain of women’s apparel stores, retention is a constant challenge, especially at the sales associate level. The company has about 5,000 employees spread across several locations, including 260 retail stores across the U.S.

"As the competition for talent remains tight, we continually evaluate our acquisition process to ensure we're selecting candidates that possess the skills and experience necessary to execute in their role but who will also be a match for our environment in the long term," says Tara Plazaran, manager of staffing, with The Limited in Columbus, Ohio.

The Limited began using PeopleAnswers, owned by business software company Infor, a number of years ago to support its recruitment and retention efforts. The cloud-based talent science platform, developed by a team of PhDs in behavioral science, analyzes 39 behavioral traits for job candidates and produces what it calls a fit rating. Attributes include ambition, discipline, energy, acceptance of authority, attention to detail, flexibility, conscientiousness, and empathy, to name a few.

The software categorizes candidates into four groups. At The Limited, candidates hired from the top two groups "turn over far less frequently than candidates in the bottom two recommendation categories," says Plazaran. "We are definitely seeing a correlation between recommended hires and retention."

Also see: Top 5 HR talent trends in 2014

And while the assessment test itself is not customizable, what does change is the role-specific profile a candidate’s answers get compared to. For example, someone applying to The Limited and Macy’s would take the assessment once, which would capture the candidate’s behaviors, but each company would see a unique fit score because they would have different profiles for what makes a best-fit employee for their organization in that role. The fit score indicates how strongly a candidate’s characteristics match the traits that distinguish a company’s high performers.

Trade off with payoff

Pre-employment personality testing can be useful, but some caution it does slow down the recruiting process. "There’s a trade off that companies have to make," says Tony Marzulli, vice president of product management for talent solutions, with ADP, the payroll vendor.

For The Limited, it’s a modest trade off with a big pay off. Ami Lane, senior recruiter with The Limited, says that while the PeopleAnswers assessment does add about 30 minutes to the recruiting process for applicants, it in no way affects the length of the interview process.

Moreover, proponents say, this type of data analysis, dubbed "robot recruiting" by some, does not remove the "human" from "human resources." "It’s one part of the hiring process, one aspect," says Plazaran. "You still need to take into account the human element and cultural aspects we continue to screen for."

Amy L. Kaufman, vice president of global talent and HR for The Results Companies, agrees, noting that her organization still uses behavioral interviews and, in fact, its talent science vendor, evolv, helps The Results Company tailor behavioral interview questions to the job profile. "The human still decides but they’re presented with fact-based, tested data on the likelihood of this person to succeed and stay," she says. "It gives a lot more ammunition to the recruiter to be able to make the strongest selection possible."

After implementing a workforce science approach to its recruiting process in 2011, The Results Company -- a call center organization with about 2,000 U.S. employees -- found that candidates who underwent the online pre-employment screening process stayed 20% longer than those employees who did not. And with the cost of replacing an employee who leaves hovering around 1.5 times an employee’s salary, cost savings can be significant.

"Even though we were following the traditional interview techniques, we struggled with finding people who considered working at our company a career versus a job," says Kaufman.

The company partnered with evolv, a data analytics firm in San Francisco. Evolv works with large companies that have turnover rates in excess of 100%. It’s a configurable cloud services platform, and is delivered through software-as-a-service in a browser. The firm generates data on employees and collects and integrates data from many sources (from a company’s existing systems, plus unstructured data) then uses its cloud-based system to perform large-scale analysis on that data. It also pulls from public data from the Bureau of Labor Statistics. The data reside in evolv's secure infrastructure, and any personal information is made anonymous and encrypted.

Also see: Apply Mad Men to talent management

Evolv sends industrial organizational psychologists into companies to develop unique online assessments for a given role. Typically 45 minutes to an hour long, the assessments present candidates with a number of different questions about their personality, their behavioral attributes and the type of person they are. Questions are deliberately opaque so it’s not clear there’s a right or wrong answer.

For example, candidates might be given a choice between "I am good at sensing what others are feeling" and "I like asking a lot of questions."

"We’ll ask them things like 'do you like to try new things and meet new people?' or 'are you someone who sticks to the rules?'" says Mike Housman, chief analytics officer at evolv. "It’s not as clear what the right answer is in there."

Candidates are then scored as green, yellow or red. "We present that score to recruiters and hiring managers and suggest they hire the greens, vet the yellows closely and, ideally, stay away from the reds," explains Housman. "And when they do that, we find that there are typically double-digit improvements in attrition."

For The Results Companies, the assessment also gives candidates a realistic preview of the specific job they’re applying for and also assesses the candidate’s customer service orientation, their likelihood for call handling efficiency, their technical aptitude, their dependability and their aptitude for sales. The questions "are giving us indicators as to whether this person is likely to be good at the job, is likely to learn quickly, is going to enjoy this job and is going to stay," says Kaufman.

Performance, attendance, retention, attrition and employee training data are constantly being fed back to evolv, resulting in a continuous feedback loop that is used to fine-tune the algorithms and make the assessments more predictive and shorter.

This constant fine-tuning enables hiring managers to challenge long-held beliefs, such as job hoppers don’t make good employees or prior experience in a field dictates success. For example, "one of the things we found is that work experience is not necessarily a predictor of success in our industry," says Kaufman. "Having specific call center experience is not as important as having that natural inclination for service orientation."

Economic reality

Throughout the most recent recession, employee retention took a back seat to corporate survival. Employees, relieved to have a job at all, stayed where they were. Many employers, meanwhile, focused on staying afloat, paying little attention to, or not having the resources for, traditional retention tactics such as merit pay increases or employee training and development programs. But as the U.S. economy continues to recover, retaining high performing employees is again becoming a top challenge for employers.

"By last year in North America the percentage of companies indicating that they were having challenges retaining, in particular, critical-skill employees, was pretty much back up where it was pre-recession," says Laury Sejen, global practice leader, rewards, with Towers Watson, a consulting firm that conducts an annual research study into the top talent management and rewards challenges facing employers in North America and globally. According to its latest study, 41% of organizations reported problems retaining critical-skill employees, and the percentages have been trending upward the last four years.

Still, Sejen says using this level of data analytics in recruitment is not yet the norm. "I just don’t think it’s a majority of companies that are leveraging the available analytics to support an enhanced recruiting process, yet," she says. "I think it’s just a matter of time, but I don’t think it’s a majority yet."

Evolv's Housman agrees. "Not a lot of organizations have really gotten into using predictive analytics and big data, but it sounds like the prevailing mentality is that this is coming, and this is going to be the way people do HR in a number of years."

In time, says Kaufman, "I’ll get to a point that I can look at a single site and I can look at my risk factors for that site -- culture makes a difference, managers make a difference, leaders make a difference. If something changes, I can watch the data and see whether my risk factors have gone up. I can see whether I need to plan some additional cultural intervention, for example. It’s where HR, instead of just processing benefits, gets strategic in looking at organizational behavior and solving problems before they happen, rather than analyzing them after they’ve happened. That’s the next generation of this."

For reprint and licensing requests for this article, click here.