The Death of Guess: Using Data to Make Better Hiring Decisions

As featured on HotelExecutive.com: http://hotelexecutive.com/business_review/5103/using-data-to-make-better-hiring-decisions

As new properties explode on the scene and traveler choices abound, hotels know they have to pull out all the stops to make every guest experience a positive one. Are staff friendly are courteous? Are rooms clean? Are meals excellent? Are bills accurate? We rely on our employees to execute their jobs, not just correctly, but with enthusiasm. And, if they don’t, business suffers. We do our best to hire good people (in a competitive market), we give them a little training, and then we HOPE they create raving fans.

Ever heard the expression “hope is not a strategy”? This phrase often pops into my mind when talking with HR practitioners about their hiring processes. Lurking beneath administratively burdensome screening systems that create an appearance of rigor, unstructured interviews, “gut feel,” and other subjective criteria continue to weigh heavily in hiring decisions-we take our best guess and hope things will work out. However, while strong intuition and a good ability to “read” people are attributes that can prove beneficial in many contexts, they should not be the linchpin of employee selection decisions.

A profound change is occurring in the HR profession: business leaders are calling for HR to adopt a more evidence-based approach to decision-making. The frequency with which the terms metrics, analytics, and big data are creeping into HR circles is a testament to this. “We see CEOs and others wanting better data and not just a headcount report, but how is talent driving business results?” says Scott Pollak, a principal at PwC Saratoga in a Harvard Business Review report. With 57% of companies reporting their intention to have integrated, multi-source analytics in place in the next two years (according to another Harvard Business Review study), there is a push to incorporate more scientific, evidence-based practices in the people-functions in our businesses.

Based on some informal research, I’ve determined that exactly 98.64% of HR practitioners have, at best, a mild distaste for statistics. Therefore, there’s an excellent chance that you may be experiencing some anxiety because of this demand for a more data-oriented approach to executing HR. Perhaps, too, this is why only 14% of businesses currently have data to show the business impact of their assessment strategy, according to an Aberdeen research study. With payroll and benefits representing one of the largest line items on virtually every company’s operating statement, effective selection is one of the principal areas where HR can have a significant impact on the bottom line. But what kind of assessment instrument should you use in order to systematically select the best employees? The answer is a firm it depends. Follow the steps below to create a highly predictive, evidence-based, and quantifiably valuable selection system that works for your organization.

The 4 V’s of Hiring

1. Vision. If companies had an unlimited budget (and candidates had endless time and patience) we could assess virtually anything-skills, knowledge, personality, values, attitudes, and the list goes on. But what are you really trying to accomplish? “We want to hire better people” is not nearly a clear enough goal. It is imperative to take the time, perhaps doing some research with internal stakeholders, to hone in on the ultimate goal for the organization.

Whether you’re trying to impact staff retention, sales volume, early hire failure rate, guest satisfaction, customer referral rates, employee engagement, productivity, theft, absenteeism, safety incidents, or drug use in the workplace, there are different assessment instruments designed specifically to measure constructs that can directly impact these, and countless other, organizational issues or goals. Once your objective is clear, you can determine what you can measure that will help to predict that outcome specifically.

2. Validity. Take a moment and think about the best hiring decision you ever made. Jot down the steps in the process that moved that individual from a faceless applicant in the crowd to the super-star employee that exceeded all expectations. Now, think about the worst hiring decision you ever made-the “wolf in sheep’s clothing” who turned disastrous. Did the hiring process that allowed the bad hire to get through the door differ from the one used when hiring the star? In all likelihood, the steps in both cases were nearly identical. After all, we develop our hiring processes with the goal of hiring the best employees every time-it’s just that sometimes it works and sometimes it doesn’t. Why? Simply because we’re basing our hiring decisions on a combination of data points that lacks the ability to predict future job performance reliably.

Extensive research has been conducted on the predictive validity-the overall ability to predict job performance-of different hiring methods and measures, and some are just flat-out better predictors than others. Unfortunately, many practitioners continue to rely heavily on some of the least predictive measures, including interviews, reference checks, and four-quadrant personality assessments (not to mention “gut feel”). Well-developed assessments, including integrity tests, mental ability tests, and multi-measure tests (which incorporate a variety of constructs) can greatly increase your odds of making outstanding hiring decisions more often. The key is to find an instrument that measures the most appropriate combination of constructs to predict the well-defined outcomes you identified in the first (Vision) step.

3. Verification. Once you know what you’re trying to accomplish (Vision) and the kinds of instruments that have a high degree of accuracy (Validity) in getting you there, you still need to choose an instrument from the hundreds (if not thousands) that are available in the marketplace. Considering the aforementioned fact that most HR people don’t choose their profession due to their love of statistics, this part can be daunting. However, it’s also necessary if you don’t want to fall victim to the supreme sales skills of a vendor. If necessary, ask an Industrial-Organizational Psychologist from your local university to assist you in sifting through the technical validation documents provided by the vendor. Any tools used pre-hire must meet certain criteria as it relates to reliability, validity, adverse impact, and a number of other factors. Test publishers should be able to provide ample data showing how rigorous they were in developing their instruments.

4. Value. Now that you’ve developed a highly predictive selection process that is virtually guaranteed to move the needle on the metrics most essential to the business, it’s time to quantify the impact that your expertise is having on the bottom line! You do this by demonstrating that the use of a particular tool is statistically correlated with what you’re trying to predict. In other words, as test scores go up, turnover goes down, or as test scores go up, upsell conversion rates increases. Often, this can be achieved through either a concurrent or predictive validation study. Can you show a statistical correlation between how people score on the test you’re using and how they perform on the job? If not, you may want to revisit the value you’re getting from that tool.

It can’t be stated strongly enough: your hiring process is destined for failure if the Vision step is not the bedrock for every decision subsequently made. As food for thought, let’s contemplate an example we’re all familiar with: pizza. Imagine your organization employs 20,000 pizza delivery drivers with an average annualized turnover of 170%. You know you can do a better job of “hiring better,” but what specifically does that mean? Consider each of the following lines of thought:

A. As HR people, we understand the costs associated with turnover, so we begin to screen for factors such as reliability, work ethic, or other constructs that might enable us to hire longer-term employees. We’d measure our success by tracking reduction in turnover.

B. On second thought, since training for pizza delivery drivers is fairly minimal, perhaps turnover isn’t actually that big of a deal. After all, all we really need are people who can zip to an address and hand someone a box, right? So let’s screen for sense of urgency and sense of direction! We’d measure our success by tracking the rate of on-time deliveries.

C. But wait! Our drivers shouldn’t zip too fast. After all, we need them to obey speed limits and demonstrate cautious driving practices. And, pizza delivery drivers handle a lot of cash, so we might also want to screen for integrity and safety. We’d measure our success by tracking cash discrepancies and driving incident rates.

D. And, gee, wouldn’t it really differentiate our company (especially in the rather commoditized pizza market) if our drivers were also really friendly and engaging with customers? So, let’s screen for friendly personalities and a passion for customer service. We’d measure our success by tracking customer satisfaction.
Each of the above is a completely valid and important strategy. Understanding which outcomes are going to have the greatest impact on the business will enable you to determine which combination of constructs is most important to measure.

The Death of Guess
Are you feeling pressure to incorporate more data-supported or evidence-based methods in your job? Is there anything you can do to increase the predictive validity of your hiring process? If pressed, could you tell your C-Suite exactly how accurate your selection system is and quantify the return on investment (ROI) for your efforts? The time has come for HR to embrace more data-driven, scientific methods for making key people decisions. Will you rise to the challenge? Well-placed and appropriately leveraged assessment instruments can be a relatively simple, cost-effective way for organizations of any size to infuse more objective data into people decisions-and to quantify the impact that the HR function makes on the metrics that matter most to the business.