11 March 2019

The biggest mistake you can make when using Predictive Analytics to hire your new colleague.

After reading this article you will know:

– The definition of predictive hiring in recruitment;
– The biggest mistake you can make when using predictive analytics to hire new colleagues;
– How to avoid this.


Last week I visited a company that asked for our advice to hire new colleagues by making use predictive data. In other words: they no longer wanted to base their hiring decisions on human judgements only. I fully agree on that. However, the way they tried to implement Predictive Analytics now couldn’t be worse.

And although they way of working was so bad, many companies are doing it exactly the same. The result: you create uniformity (in the Netherlands we call this ‘eenheidsworsten’ — a group of very, very similar people). If that’s what’s happening to your company as well, you’ve got a problem. A really painful problem that will cost you a shitload of money. Trust me.

Change. It’s happening. Hallelujah.

The recruitment industry is changing. Innovating. Finally. It took us quite a while, but it looks like it’s really happening now. Cool. One of these cool innovations is Predictive Analytics (PA) — analyzing historical data in order to predict future outcomes. Sounds cool, right?

When translating this into your hiring process, PA can help you hiring the right candidate for the job by predicting his/her job success. If you’re using it right at least.

Predictive Analytics : This is how it can destroy your hiring process.

So PA has a huge potential for your company, but only if you’re using it the right way. The key to PA is historical data. This data set will be analyzed and used to predict a future outcome.

HR leaders will use PA to evaluate candidates and their potential. Within my company Equalture — a pre-selection technology for recruiters- we use PA for our matching algorithm. Based on this algorithm we can predict a candidate’s job success and therefore help you hiring the best candidate for the job.

The part where you can really mess this up is the data part.

Internal vs. external data.

When companies start using data analysis to analyze candidates, 90% of those companies will start by analyzing its own team. And that’s fine, it’s definitely a good start. However, focusing on internal data shouldn’t be your only source of data. Why? Well, simply because it stimulated uniformity.

When purely focusing on the colleagues you already have in your team, it becomes very hard to bring people with other skillsets and personality traits into your team. The result: you will keep hiring exact same copies of your existing team.

Hi uniformity, bye bye diversity.

This uniformity within your company introduces a slow death of diversity in your team. The consequences:

– Your company’s variety in skills, personality and experience will decrease.
– Which counteracts innovation and criticism.
– And therefore badly influences your company’s performances.

The results: you can lose a shitload of money — even up to 42% of your company’s revenue.

So although it seems so promising to innovate your hiring process, it can backfire quite painful when doing it the wrong way. However, the good news is that, although this is quite a tough problem, it’s not that hard to fix it.

How to avoid this uniformity in your company

To avoid this copy paste effect, you need to bring another dimension to your dataset. Within Equalture we use the following types of data:

– Type 1. Historical data from company X (client outcomes);
– Type 2. Historical data from the industry in which company X operates (both databases and client outcomes);
– Type 3. Historical data from comparable jobs (both databases and client outcomes).

So by using these three types of data we focus on the client, its industry and the expertise fields of the job openings they have. Bye bye uniformity. Hello diversity!

How to get access to these types of data.

Type 1 is obviously the easiest type of data to get access to. You now just need to gather and structure it the right way, but I trust your expertise on that.

Type 2 en 3 might be a bit more difficult, because here you need to find relevant databases ánd you haven’t got access to data from similar clients. Within Equalture we make sure you don’t need to worry about any of these categories; it’s already in the self-learning matching algorithm.

To end this blog with the story I told you in the introduction: the cool thing is that the company I was talking about just called me this morning to become a new client. They now can successfully use data to optimize their hiring decisions. At least their teams won’t become a group of ‘eenheidsworsten’.

Curious to learn more about how technologies such as Equalture might help you to introduce Predictive Analytics into hiring process? Just send us a message, no strings attached.

Happy hiring! ?

Cheers, Charlotte