Questions that appear simple to the person asking them can be messy to an analytics pro who has to answer them. One nasty form of messiness is changes to the organization structure. That is, any question that involves a longitudinal look at a department will be difficult if that department has been affected by structural changes.
Let’s imagine the people analytics team is asked a dead simple question such as, “Compare turnover in HR vs. Marketing over the past six years.” Surely, analytics pros can get you a straightforward answer in a matter of minutes.
But now here’s where the problem of reorganizations fits in. Let’s imagine that last year, Learning & Development, which used to be a separate department and has very high turnover, was merged into HR. Do you leave in the Learning & Development data for the one year it was merged with HR even though it skews the results? Do you exclude it? Do you go back and add it in for all six years? Perhaps to add to the fun you find that the Communications department split off from Marketing three years ago. Again, you now have a similar dilemma.
Maybe you should go back to the leader (presumably after more than the few minutes they thought this would take you) and say, “Here are three ways of looking at HR turnover, and three ways of looking at Marketing turnover, and hence nine different answers when we compare them.”
Somehow that doesn’t feel satisfactory.
How to Approach the Reorganization Problem
The most important takeaway is that the notion that the problem was simple is an illusion. There really are nine different answers to the question of HR vs. Marketing turnover. Since leaders typically don’t want nine answers to one question, the analytics team will need to make some choices.
The choices on how do to the analysis revolve around two things:
Article Continues Below
Phenom AI Day | Dec. 9 | 11am ET
- What can we practically do given the time and resources available?
- What way of handling the data best informs the decision the manager who asked for the analysis is trying to make?
With respect to practicality, sometimes the complexity of the reorganization makes analysis very tough. In that case, one simply has to be realistic about what can be accomplished. You then do what is practical and make sure the manager you are presenting to understands which choices you made and why.
With respect to the decision that needs to be made, analysis is always easier and more relevant if you know what the manager intends to do with the answer. Perhaps the manager is really interested in whether the incentive scheme used in marketing, which differs from that in HR, affected turnover. If you uncover the decision they are trying to make, rather than just giving them the data they ask for, then you are more likely to be able to make the right choices in your analysis.
A Tip and a Conclusion
One tip is that it’s often easier to do analysis by cost center (which is how finance organizes data) than doing analysis by the department as it appears on the organization chart (which is how HR organizes data). The cost centers are typically more stable, so there are fewer changes to deal with.
And one conclusion is that people analytics is almost always harder than it appears at first glance. The analysis is going to be more complex and take more time than you thought. Furthermore, the answer is going to need a lot of caveats and explanations if it is going to be interpreted properly. That’s just the reality of people analytics, and it’s something we need to accept.