The decision of who to hire is the most important and one of the most difficult decisions companies make about employees. It is important because every other decision about an employee is a consequence of the initial decision to bring them into the organization. It is difficult because it requires predicting the future based on very limited data. Answering questions like “based on a two-page resume and responses to a few interview questions, how is this person likely to behave in six months when we put them into a new environment and ask them to do things they have never done before with people they have never met?”
Predicting the future is uncertain and no hiring process will ever be close to 100% accurate. Nevertheless, companies must do what they can to increase their odds of hiring the right people. The best way to optimize hiring decisions is to approach them as a measurement problem. Accurate measurement requires three things. First, clearly defining what outcomes you want to learn or achieve through measurement. Second, systematically collecting and consistently interpreting data to ensure you are taking accurate measurements. And third, evaluating and improving the accuracy of your measures by comparing them with your desired outcomes. To illustrate, imagine you want to replace several rotten boards on your deck. First, you would define how long the new boards need to be to replace the rotten boards. Second, you would measure and cut new boards to replace the old ones. Third, you would place the boards into the deck and see if they fit. If they did not fit, you would go back and examine the first two steps to see where you made a measurement mistake.
When it comes to hiring, the first step is to define what data you can collect about candidates that will predict how they will perform on the job if they are hired. The second step is to collect this data from candidates and use it to guide hiring decisions. The third step is to compare the data collected during the hiring process with post-hire performance of the candidates to ensure the hiring process is predicting job outcomes. This might seem like a pretty basic concept, but the reality is very few companies actually do it. Instead of focusing on using data to hire the best employees, they use to data to identify the best candidates. Which is not the same thing.
The term “quality of hire” refers to the value employees provide to a company after they are hired. It is a function of selecting candidates using data that predicts post-hire outcomes such as job performance and retention. In other words, ensuring selection decisions are based on job relevant attributes. One might assume that most hiring technology solutions are designed to improve quality of hire. But this assumption would be wrong. What most hiring technology systems actually measure is “quality of candidate”, which is the likelihood that an applicant will pass a company’s selection process. The majority of hiring technology solutions use data to predict people’s potential to be hired into the job. They do not use data to predict whether these people are successful after they are hired.
There is a big difference between being a good candidate and being a good employee. A good candidate is someone who has the skills, social connections and attributes to get hired. A good employee is someone who has the skills, social acumen, and attributes to be successful in the job. These things often overlap, but they are far from the same thing. Hiring managers might be impressed by candidates who
have the “right look”, went to the “right schools”, or worked for famous companies but that does not mean these candidates will be good employees. A socially awkward but technically skilled applicant might come across as a poor candidate in a job interview but be an exceptional employee when placed in the job.
The mismatch between what hiring systems measure about candidates and what actually matters for job success lies at the heart of peoples dislike for candidate selection systems. The best action we can take to build credibility in use of hiring technology is to stop building solutions that only evaluate quality of candidates and build solutions that predict quality of hire. This requires creating solutions that empirically link pre-hire candidate data with post-hire job outcomes. This change is well within the capabilities of modern technology[1]. But companies must shift from just focusing on maximizing hiring efficiency to increasing hiring effectiveness. The change will take effort but it will have tremendous benefits for candidates and companies alike. Nothing good comes from making hiring decisions that place people into jobs they are unable or unwilling to perform. Great things happen when people are hired in jobs where they can realize their full potential.
