725 Skippack Pike Suite 320 Blue Bell, PA 19422

Testing and Employment – How a Good Idea Goes Bad

Let me start this White Paper with an admission. I have never been a huge fan of personality profiles and testing in general. I think every person is unique and every hiring situation and employer is different. I think that it is impossible to develop a test that can accurately generalize about every potential candidate for every possible job. But, I also know that tests have become a common tool in the employment landscape and many of my clients use them They are a fact of life.

 

So, the purpose of this White Paper is to do two things. One, explain the best type of test to use as an employer. Two, to give a real-life example of how a well-intentioned testing policy can morph into a program that basically runs your company on its own. As we enter this discussion, please remember that I am talking about tests for professional level candidates. I know nothing about tests for retail, hourly or related positions. For all I know, they may be very valuable. Or maybe not.  Right now, we are only talking about personality profiles for professional level employment. We are not talking about skill tests. That is also a different animal.

 

First point – good tests are expensive. If your tests are basic 5 minute profiles that pigeonhole people into a box or graph, you are wasting money by saving money. Those tests, at best, tell you nothing and, at worst, are easy to deceive. If you want to test people, do it right and find a test that can give you a report specific to that individual. If you want to hire a metric on a graph, wait till Google comes up with a robot to do the position. Of course, by then your job may be done by a robot, too.

 

Good tests should also be used at the end of the project, not at the beginning. This is true for two reasons:

 

  • They are meant to give perspective on your impressions from an interview. They are not meant to give you an impression of a candidate who you have never met. Tests fit into the employment process at a specific place and time. If, after the first interview, you and the candidate have mutual interest, the test is appropriate. If the test validates your interview impressions, good. If not, explore the discrepancy in the second interview.
  • I once met with a client who wanted to test people prior to an interview. I said that was wrong for two reasons. One, you are asking someone to invest an hour or more of their time before they have even met you. It is rude. It implies to a recruited candidate that you see the employment process, and by extension, your employees, as digits and not people. Many good people will balk at that and it will shrink your candidate pool. Two, it’s too expensive. You will be testing everyone and not just finalists. You are better off investing in better tests for fewer people than cheaper tests for everyone.

 

So, if you want to use tests and personality profiles, use them wisely. Invest in good ones and make them a useful part of your process but not the dominating or entry point of your process. They should be food for thought and not the actual decision making entity. That last point is critical. Tests can take over your process, as we will see in the second part of this White Paper which is a description of a good idea gone bad – The Frankenstein Test Monster.

 

We once worked with a client that tested everyone. Every person got the same test – from non-degreed supervisors to VP level candidates. It was a half hour online test that turned the person into a point on a graph. As a rule, the test just confirmed the interview impressions but, occasionally the test came back with a surprise. Here is a typical scenario.

 

The manager and HR person would love the candidate after the interview. Excitement would rain about this great person they interviewed and, after the test (a mere formality), an offer would be forthcoming. Then came trouble. The test would show something unexpected. Maybe an aloof attitude, an arrogant attitude or a tendency to micromanage. A hiccup. There would be talk about ignoring the test or re-interviewing to explore in greater depth but, in the end, the test anomaly always ended the process to everyone’s disappointment. This happened once a year, maybe in 10% of their employment situations.

 

I once asked one of their VPs why a potentially off-kilter test result ended the process so abruptly and he had an interesting response. The CEO was the type of guy who loved to apportion blame. If any new hire did not work out or anyone ever quit, the CEO would immediately look at the test results. If there was a hiccup in the test and the person was hired anyway, blame would fall on the Manager and HR for hiring someone who the test said was unqualified. I asked the obvious next question – did anyone ever not work out or quit who had good test results? The VP said, of course, but a normal test protected the manager from blame.

 

This is a classical example of test mission creep. In reality, the test had become the deciding factor and managerial judgment had been neutered. Even assuming the accuracy of the test, which is a dubious assumption, the dominance of the test basically removed managerial input from the equation. To make it worse, it also removed candidate talent from the equation. It created a culture of blame and fear, which spread into other parts of the culture. Over the years we worked with them, the company went from entrepreneurial to so risk averse that nothing got done.

 

So, think of your testing process as the canary in your coal mine. Your employment process is the first exposure people have to your company and affects all employees, both new hires and current managers. If you want to make testing an integral part of the process choose a good test, use it correctly and constantly reevaluate the process to be sure that you are not, inadvertently creating your own Frankenstein Monster.

 

As always, thanks for getting this far and don’t forget Right Recruiting for all your recruitment needs.

Leave a comment

You must be logged in to post a comment.