As businesses evolve, their hiring needs change and the recruiting process must become smarter. As a result, HR functions are diverging from tried-and-true recruiting methods and embracing technology to make better hiring decisions. In fact, many HR leaders are turning to data, with leading companies like Google, NFL and LinkedIn constantly collecting and leveraging data - such as assessment scores for candidates and satisfaction and retention rates for employees - to make data-driven hiring decisions.
This is what we consider predictive analytics. Simply put, this is a form of advanced analytics that uses new and historical data to make predictions about future outcomes.
But what happens after you’ve made a decision based on your data? Whether you chose to hire or even terminate an employee, how do you know which of the data was accurate and which was misleading? Furthermore, how do you know if you’ve made the right decision, and how can you optimize your decision-making process in order to improve future hiring decisions?
This is where the scientific approach comes in.
Applying science to the recruiting process can help HR functions account for ongoing change and make continuous improvements based on the latest available performance data. And according to Talent Management and HR, shifting to a scientific model can significantly improve business results and improve the credibility of HR leaders among the executive team.
Science, in general, is the systematic study of a particular structure or behavior through observation and experiment. Scientists constantly hypothesize and test their predictions in order to accept or reject them.
Based on their findings, they will either repeat the experiment in different contexts to test the validity of their methods, or they will make corrections and iterate until they get it right.
And this same notion can be applied to the recruiting process to test your predictions. In fact, both Amazon and Google are committed to “becoming scientific” in order to make HR decision-making as precise and scientific as other major business functions, like engineering and finance.
To be scientific in any discipline requires you to continually challenge your assumptions and question your decisions. And this is especially important when it comes to the recruiting process, as the average cost of one bad hire is close to $15,000 USD, according to a recent CareerBuilder’s survey. Below I’ve shared five key principles to apply to your recruiting process in order to shift to a scientific model:
This is a fundamental step in applying science to your recruiting process, and it involves identifying a method or assumption that you would like to test for accuracy. For example, you may choose to test the underlying assumptions you have regarding the right people for a particular role. In this case, your hypothesis may be that a successful salesperson is extrinsically motivated (i.e., by money), tech-savvy and possesses key traits like empathy, focus and optimism.
Equally important, you also need to define how you will measure your hypothesis. Ask important questions such as, When will measurement take place within the recruiting process? What will it take to gather sufficient data? What tools will be needed to test the assumptions? In the example mentioned above, you may conduct a personality assessment to gauge a candidate’s soft skills, or test their proficiency in certain tools and software.
In order to effectively check your hypothesis, you will need quantifiable measures that you can compare to actual performance data, like interview and assessment scores. To ensure the validity of your methodology, you will also need to score all candidates in a standardized way in order to create reliable data.
As an example, consider Google’s sample interview rubric for a fictional role. The rubric utilizes a structured matrix to identify the key attributes evaluated in each interview, and outlines five predefined scores and indications for each trait - ranging from poor to excellent - that will allow recruiters to grade candidates in a consistent manner.
To determine whether your findings support - or oppose - your hypothesis, you must identify key performance metrics to track within the recruiting process. These criteria will help you decide whether to accept or reject your hypothesis, as well as what to measure in real life. For example, sales motivation may correlate with actual sales, while high empathy scores may translate into higher customer retention rates.
The ultimate goal of science-based hiring is to continually improve your decision-making process based on your findings. That’s why it’s important to compare the resulting data to actual performance data. Be sure to analyze which parts of the process worked and should be continued, as well as which parts didn’t work and should be changed or stopped completely.
Continuing with our example above, maybe you found a strong correlation between sales motivation score in the interview and actual sales performance. In this case, you should continue interviewing for this trait. Conversely, you may have discovered empathy score in the personality assessment had no connection to customer retention rate; therefore, you can stop measuring this trait or choose to assign less importance to it within the decision-making process.
Using predictive analytics to make hiring decisions is essential but insufficient. Rather, we need to take it a step further and begin using science to challenge our assumptions in terms of what does and doesn’t work in the recruiting process. Although taking a scientific approach to hiring may seem like a long and difficult process, you can start small by checking one hypothesis at a time. And fortunately, integrating science into your hiring process is easier now than ever due to advanced technologies like machine learning and cloud computing, which we’ll explore further in future posts.
Enter your Email to download the full Case Study