The Right Way to Use Recruitment Technology

June 24. 2019. 5 mins read

It seems like just about every tech publication cannot help themselves and feels the need to start taking a political spin. Here at Nanalyze, we avoid politics like the plague. Why? Because the United States – where the majority of our readers come from – is polarized to an extent that hasn’t been seen since the last Civil War. For every fifty people out there who believe that “political talking point A” is the way forward, there are fifty people who disagree vehemently and believe that “political talking point B” is the way forward. Still, most tech media outlets out there just plow forward blindly, assuming that we all hold the same political beliefs. A good example of this is “diversity.” Not everyone defines diversity the same way.

When we talk about topics like “diversity in hiring,” we’re very careful to avoid identity politics. Equality of opportunity is a core value that every single firm out there should espouse. Everyone gets an equal shot at success – and in today’s increasingly competitive global environment – that means working your ass off to get there. Equality of opportunity should be guaranteed, not equality of outcome. When we look at the plethora of startups out there that look at “changing the way we hire,” there are good ones (equality of opportunity), and there are bad ones (equality of outcome). This is a topic we touched on in our article on Catalyte Uses AI to Hire Better Software Engineers.

In order to measure the success of any recruitment technology solution, the answer is not looking at a group photo of the company and evaluating the various shades of skin color and genders. That method makes about as much sense as balling up half the world into a single category and then calling it “Asians.” What we need to do instead is to define an objective metric that can be used to determine if our hiring technology is having a positive impact on a firm. One metric that every firm wants to improve is something called “retention.”

In order to understand how we might use technology to make better hiring decisions, we turned to a firm we’ve talked about before – PredictiveHire – which uses artificial intelligence to improve the hiring process. They put together a few interesting examples which should serve as an example of how technology ought to be applied to the hiring process.

How PredictiveHire Works

The notion behind PredictiveHire’s technology is that top performers have a unique “fingerprint” that differs based on each organization. In other words, there is no one-size-fits-all approach to hiring. Combinations of certain personality traits have a higher impact on desired outcomes, whether that’s turnover, sales performance, or other business KPIs. PredictiveHire looks for data patterns in high performers using an increasingly sophisticated series of inputs – more than 60 at the moment – which increase their accuracy in predicting top performers.

In one example, PredictiveHire built a conversational chatbot called Phai, whose full name is PredictiveHire Artificial Intelligence, and it would seek out the lead indicators of job success by asking potential candidates a series of questions.

PredictiveHire Artificial Intelligence
PredictiveHire Artificial Intelligence – Source: PredictiveHire

The candidate answers 100 different questions that are then used to predict the likelihood that they will mesh well with the particular organization they’re applying at. While the company used to employ multiple choice questions, they’ve now moved on to comprehensive text analytics which can analyze natural conversations and look for patterns. The most important part of this whole process is identifying objective metrics that can be used to determine success. One objective metric that most companies would like to improve is retention.

Reducing 6-month Churn

The first case study we looked at was how PredictiveHire’s product was applied to a National Provider of aged care services (in other words, a government-operated old folks home) where there was a problem with “6-month churn,” which essentially means that the number of new-hires leaving before having worked more than 6 months was too high. Why is this a problem? Well, it’s because most companies employ an entire department which is tasked with finding candidates, screening them, having highly-paid managers interview them, on-boarding said candidates, and then training them.

Right away, we have a tangible and objective measure of performance that has nothing to do with whether or not you’re the type of person who would enjoy a steamy night at the Abbey in West Hollywood. We can call this metric “6-month churn,” and it’s simply the percentage of people who are hired that are still employed after 6 months. Using PredictiveHire, the client was able to accurately predict with an 82% accuracy if a new hire was likely to be “high-risk.” If every new hire in the next 6 months was hired based on the predictions made by the model, turnover would decrease by 50% in the next 12 months. It’s easy enough to assign dollar amounts to these results and it becomes an easy sell at the C-level. Another use case we looked at helped increase the likelihood that someone would be a high-performing salesperson.

Increasing Sales Numbers

The next case study involved a global sales organization with 800 sales people where – as is the case for most sales-focused companies – growth comes from having a productive high-performing sales team in the field. In this case, the sales organization would typically hire people with little or no relevant prior experience, meaning that they needed to invest a significant amount of time and resources into the newly-hired raw talent. PredictiveHire administered their test to 823 salespeople within the firm and used those results to improve the hiring process. Using their predictive analytic model, PredictiveHire was able to predict with 88% accuracy whether an applicant who completes the same assessment would be in the top quartile of sales performers. Furthermore, the incremental sales revenue from hiring all new hires using their predictive model would be 45k per hire, per year. Even more interesting is that if every employee in the lower performing group was replaced by a new hire in the higher performing group, the sales organizations revenues would increase by 41%. Again, this is an easy sell at the C-level.


A while back, we opened up the Fortune Term sheet – which we’re big fans of by the way – and were greeted with this statement:

Startups like Pymetrics, a developer of neuroscience-based assessments for staffing services, could see a big boost as the political headwinds (especially as political superstars like Alexandria Ocasio-Cortez take power) start demanding accountability.

There you go again Fortune, diving headfirst into politics instead of focusing on how we can use technology to solve real problems. At least half of America disagrees with the political stake you’ve just planted in the sand with the above statement, but yet you seem to assume that everyone sits on the same side of the political spectrum you lean towards.

Regardless of where you stand politically, attempting to use recruitment technology for promoting equality of outcome is a dangerous approach that does a disservice to every individual who wants to be evaluated on their own merits instead of being patronized based on some attribute that is irrelevant to their ability to perform. Equality of outcome is an approach that is driven purely by politics, not by a focus on making sure that we hire the most high-performing individual for any job. Using solid metrics like “retention” or “sales numbers” is how competent hiring managers will expect to measure the success of any recruitment technology, and PredictiveHire’s AI-powered solution should be commended for taking a scientific and quantifiable approach as demonstrated in the above examples.


Leave a Reply

Your email address will not be published.