Why you should always start with a Question when applying Human Resources Analytics

This blog explores the first important step towards finding the real value in Human Resources data: starting with a question that is important to your business.

A pile of gold

Smaug the dragon is a fearsome creature that plays a major role in the film series The Hobbit. For centuries, Smaug slept atop a pile of gold of an unimaginable value, keeping all the golden beauties to himself. Now, what does this have to do with Human Resources?

The image of Smaug always pops into my mind when I think about HR and the huge amount of data it possesses. HR is a lot less fearsome compared to Smaug but there is an important similarity: HR is excellent in collecting data through their reporting process. However, HR does not do much with this data, even though it might be just as valuable as Smaug’s gold.

The risk of cherry picking

Most data-savvy people recognize the temptation of digging through data, conducting lots of data analyses and searching for those juicy, significant effects. There are even entire master programs that teaches people how to do this most effectively (I am looking at you, economists!). However, there is a problem when you cherry pick data. Any analysis without a proper hypothesis or question has a high risk of resulting in seemingly interesting facts and figures that, in fact, do not help the workforce at all.

I recently came across a quote from Albert Einstein that can be applied to this issue rather well:

“If I had an hour to solve a problem and my life depended on the solution, I would spend the first 55 minutes determining the proper question to ask, for once I know the proper question, I could solve the problem in less than five minutes.” – Albert Einstein

Albert Einstein

hit the nail right on the head. When you have enough Human Resources data, you will always find interesting relations. Still, most of these relations are either non-findings (also called false-positives) or they do not add real value to the business.

Look for real challenges

This point is emphasized in a recent publication on Human Resources analytics by Rasmussen and Ulrich – Human Resources Analytics (2015). According to Rasmussen and Ulrich, HR analytics (read: What is HR analytics?) too often starts with studying the data without looking at the real challenges the business faces. This approach greatly diminishes the value of HR analytics and potentially reduces its impact to a short-lived craze.

Whenever you want to analyze something in HR, and especially when you want to engage in predictive analytics, it is essential to start with a question or hypothesis. The best questions are of importance to both HR and the CEO. A few examples of questions are: Do I have the right employees? Who are my high potentials? What is the cost of employee turnover? How does the way people are rewarded impact their performance?

Only when you have defined what you want to measure, and (consequently) have a specific question, you should look at the data.

Now, problem-solving is not as easy as Einstein makes it sound. Contrary to his optimism, it often proves to be quite difficult to solve a problem within five minutes (especially if you take into account the quality of most Human Resources data). In HR analytics, aggregating, structuring and cleaning the data is often a difficult and very time-consuming. Once all the data is ready to be put into use, the actual analysis only takes a few seconds.

Finding the right answer to the wrong question is easy

In other words, the real value of HR analytics does not lie in the accumulation of massive amounts of data. The value lies in having the right data to answer specific questions that add value to the business.

To send this point home, I will end with a quote by Tukey (1962), a brilliant mathematician and statistician, who is also quoted in Rasmussen and Ulrich’s paper:

“Far better an approximate answer to the right question, which is often vague, than an exact answer to the wrong question, which can always be made precise.”

In other words: proper people analytics always starts with a specific question.

Sources:

Rasmussen & Ulrich, 2015. Learning from practice: how HR analytics avoids being a management fad.
Tukey, 1962. The Future of Data Analysis