You come to a fork in the road. The sign indicates 100 travelers have taken the left fork and 14 have fallen to their death; it also shows that 50 travelers have taken the right fork and 8 have fallen to their death. Which road do you take?
Welcome to the world of HR analytics.
The answer to this "which path" puzzle is one you probably won’t learn in a statistics class, and it demonstrates something HR leaders need to know: the road to analytics success isn’t always paved with data scientists.
Now, back to the decision: which path? The natural reaction is to take the path with a lower percentage of deaths. But wait! You might (correctly) suspect that the difference is not statistically significant. Does this mean it doesn’t matter which way you go? No, it still matters, and the reasoning behind the answer will solve many problems in people analytics.
The typical HR professional should be learning methods to enhance decision clarity.
What Question Are We Trying to Answer?
If you were asking the question, “Is there strong proof that one path is safer than the other?” then the answer would be “no”. But that’s not the question is it? The real question is, “Which path should we chose, given that we have to pick one?” In this case, we can only look at the best available evidence: the percentage of deaths. We should take the path with fewer deaths per traveler even though the evidence we have is weak.
If other evidence about path safety comes along, or if one path is more costly than the other, then we may change our decision. In the absence of that additional evidence we still have to forge ahead. We don’t need a calculation of statistical significance to make our choice.
This simple story encapsulates four important elements of successful people analytics:
- We need to be clear about what question we are trying to answer.
- We need to gather the best available evidence—which, even if it not good, will be better than no evidence.
- We need to assess the quality of the evidence so we can make an informed judgment.
- Often, basic math is all we need to inform our judgment.
Of these points, it is the first one that matters most. We don’t do analytics as a textbook exercise, we do it to make a business decision. When we are clear about the decision, then the rest follows.
What to Do?
In my analytics workshops, HR professionals are often relieved that I’m not teaching statistics. There is a role for statistics, but that’s not what the average HR person needs to learn. The typical HR professional should be learning methods to enhance decision clarity—i.e., be trained in asking the right questions. That’s the single biggest driver of analytics success.
Secondly, they should be trained to use the basic math skills they already have. We can go a long way to better decision making with counts, percentages and estimates; people need to recognize the value of this basic math.
Finally, they need to understand when to call on extra skills sets for problems that can’t be answered with simple forms of evidence—and this is where we do want unleash the data scientists.
HR will find that most of the wins in analytics fall somewhere between the rudimentary world of HR reporting and the exciting world of advanced statistics. To get there the average HR professional just needs a little extra help in bringing rigor to decision making. If they have the confidence to choose the statistically insignificant fork in the road, and explain why they made that choice, then they are on the right path to analytics success.