Anyone who has tried to recruit and hire talent knows how difficult it is to estimate a candidate’s fit. The fundamental challenge in recruitment is to find the talent with the competencies to fit the job and the personality to fit the firm’s culture. This is difficult because candidates give their best impressions during the hiring process, and thus it becomes your job to separate fact from fiction. The traditional approach to improving hiring accuracy has been to use technically rigorous hiring practices that include a systematic analysis of the job, structured interviews with job-related questions, and psychometric assessments and simulations. These approaches are effective and there is an abundance of scientific evidence suggesting that hiring decisions will be more accurate when these best practices are used.
Yet in the modern world these practices seem out of date. Recruitment and hiring have been “digitised”, as firms now routinely review a candidate’s online information as part of the hiring process. Some of these practices appear sensible, such as searching LinkedIn profiles for potential candidates to source. Other practices seem more risky, such as reviewing personal data or social networking sites (such as Facebook) as a means to learn about a candidate’s personality.
Digital hiring: fact vs fiction
It is understandable why you would want to review online information. The pressures for speed, low cost, and candidate supply (versus demand) make digital hiring very appealing. The information available in digital platforms is believed to be more revealing about a candidate’s “true self” than the best impression one usually sees in an interview. The fact that the data exists in digital platforms enables you to use big data analytics to source and select candidates. These approaches range from simple keyword searches to sophisticated models based on neural networking. This digital data provides the impression that you can make better, faster, and cheaper hiring decisions.
It is a nice simple story, and the reason is because it is a story based on fiction rather than fact. There is surprisingly little objective or scientific data that speaks to whether digital hiring provides any tangible benefits in terms of candidate quality in a manner that saves money and time, and is legally defensible. There is little objective evidence to suggest digital hiring (such as reviewing social networking sites) is more effective and efficient than traditional approaches for sourcing or hiring quality candidates. Most of the evidence one finds about the “benefits” of digital hiring are based on anecdotes, vendor reports, and surveys of whether recruiters believe the approaches are more effective. Unfortunately, anecdote and belief are no way to run a high-performing organization (can you imagine an executive team meeting about operational or financial choices that would be based solely on belief rather than data?).
There are other risks associated with digital hiring. Countries differ in the extent to which personal information can legally be used as a basis for hiring. Many countries require that hiring decisions must be based on job-related information. In these countries it becomes risky to even review LinkedIn information because that information often includes pictures of the individuals—and thus creates the potential for discrimination in the hiring process.
Even when there are attempts to employ analytics to remove discrimination from the hiring process, bias can still occur. Algorithms and statistical models are blind to the nature of the data, and without careful guidance from experts knowledgeable about hiring and legal issues, it is easy for statistical models to be highly predictive but also highly discriminatory in a manner that is not legally defensible. A statistical model is simply a machine that will try to maximize the prediction of some outcome given a set of inputs; it does not care whether those inputs are legally or ethically appropriate.
Regardless of a country’s employment laws, the way to maximize performance and fit through recruitment is to focus hiring on job-relevant competencies. Traditional best-practices in recruiting and hiring are based on isolating job-relevant competencies so that applicants can be compared to each other and judged against important requirements for a job. Newer approaches that employ digital hiring have a certain “sexiness” that is appealing to recruiters who are challenged with a very difficult job. However, you must ask yourself whether it makes sense to use digital hiring—practices of unknown effectiveness and high legal risk—versus traditional practices with known benefits and legal support.