What I love most about analytics is that it teaches you how to ask the right questions.
This morning I attended a breakfast meeting put on by the Australian Institute of Management. While waiting for the doors to open, I had a lovely chat with a fellow attendee who works in an ethics role within a public sector organisation – I’ll call her Sonia. When I told her I worked in analytics, she asked if she could pick my brain for a few minutes.
Sonia’s organisation conducts ethics training as both an online option and through traditional classroom formats. She was curious if I would have any insights into how to determine which method of ethics instruction had a greater impact and how to measure that. Was it just a case of how many people took each course and the impact of ethical breeches?
For many of us, we think in terms of counting and quantifying when we talk about analytics. But in reality, analytics starts with asking the right questions. What is the problem we are trying to solve? Where are the correlations and causation which, with insight, we can attempt to shift?
In Sonia’s situation, it’s best to start with a focus on the outcomes rather than inputs to ethic training. So instead of looking at people who have been trained, it might be better to look at the incidents of ethical issues. If Sonia were to focus on the smaller employee population of those who had made poor choices or were subject to ethical scrutiny, she might begin to see patterns which may not be obvious when you start analysing the entire employee population and look at training alone.
In a popular case cited by human capital analytic thought leader Josh Bersin of Bersin by Deloitte, they were looking at incidence of fraud within a bank. There was no correlation between training and fraud. The strongest correlation was linked to the distance between the branch and the bank’s head office. Out of sight, out of mind. The further away an employee felt from scrutiny of the head office, the more likely the incidence of fraud.
For Sonia, I suggested she look at various employment data: length of service, time in their current position, age, remuneration rate, position, time since ethics training. Those and many other criteria could hold some of the keys to find correlations to ethical breeches. If deep analysis is too much, even just reviewing the training methods received by her smaller, targeted group may provide more specific results. Ultimately, should Sonia’s organisation choose to move forward with only one method of ethics training, she should continue to monitor the incidences of ethical breeches to see if there is a change in pattern over time.
It all starts with a question. What’s yours?