PredictiveHire tackles the problem of time and resources by turning the process of identifying high performers on its’ head. Essentially, rather than inferring or ‘informed-guessing’ a person’s likely performance from the usual recruitment methods of interview and background and skill review and perhaps Psychological Assessment, the actual performance data of real employees in a role is mined and used to identify those criteria that actually differentiate or identify higher performance. Put another way, actual performance is used to predict the likely performance of a candidate, rather than the other way around. This actual performance data comes from measures like sales volumes, response times to phone calls, Net Promoter Scores or other business information about actual employees. When applied, PredictiveHire’s data science tells you whether a person applying for a job matches the specific characteristics that are actually measured in high performers.
One reason this approach is so powerful is that it is standardised, as the advanced statistical analytics that underlie the methodology ‘learn’ what predicts the best people independently from a human’s judgment. No bias is possible because the identification of a higher performer does not draw on any information that could be subject to bias; there is no ability or criteria available to input subjective judgment into the model. The questions that a candidate is asked, and that ‘feed’ this model when they go through a selection process are suitably diverse, psychologically robust and designed with the same rigour as contained in standardised Psychological assessments. So, for the participant answering the questions, the process feels familiar, work-related and brief. In essence, PredictiveHire draws on the best Psychological science around what you would ask a potential employee to find out if they were going to be highly effective, and then compares that person’s pattern of answers with those who have proven to be effective as employees. These questions are hugely diverse, but only the most relevant end up being used for each role.
At the core of the PredictiveHire model of assessment is a key fact; from role to role and organisation to organisation, the one set of questions will deliver very different predictive models. The reason for this is that between the question set, and the specific candidate is the ‘machinery’, ‘engine’ or ‘model’ of data that differentiates actual performance. For example, in one role, there might be say 20 questions of the 100 or so questions asked that, when answered a particular way, indicate a strong candidate. In a different role, or organisation, perhaps only a few of these questions and several other, different questions would predict a high performer. It is the statistical model that works this out, not the person making the selection decision. Using this information, along with other candidate information available, the recruiter or hiring manger is then in a position to make a better decision. But this can be done over and over again as more candidates are assessed, further improving the accuracy of the selection ‘engine’ over time.
Very quickly, the model that a client organisation is using to select high performers becomes completely customised for their own, continued use. No other organisation can possibly have access to this kind of bespoke Intellectual Property, and such competitive advantage is unprecedented in Recruitment technology to date. The methods used by PredictiveHire are innovative and a departure from most thinking about how Psychological Assessment is used in business, however the ‘performance feedback loop’ that is used in this methodology is what sets it apart. The days of estimating and best-guessing a candidate’s performance, for the majority of roles are limited, as are selection systems that ignore the measurement of real people-performance in their design.