One of the many difficult experiences in an analyst’s life is when they present turnover data.
It starts easily enough as the analyst presents general trends. Then it gets ugly when someone asks which manager has the worst turnover. When the analyst pulls out that information, and if that manager is in the room, then that manager is likely to angrily go through the turnover, individual by individual, showing why each particular case was unavoidable and hence why the analyst’s data are essentially garbage.
The analyst, sitting there with abstract data, has little to say in light of the specific case-by-case knowledge of the manager.
This leads to a difficult question: If people really are unique individuals in unique circumstances, and if managers know each specific case, then does the sort of statistical data an analyst produces have any real value?
In one mid-sized organization, I know the CEO skips past the overall averages and goes over the turnover results individual by individual — that’s how he figures out if something needs to be done.
There are several answers to how an analyst should respond to this situation. One is that the manager may be right, and that there is nothing wrong with how they are handling turnover. If the analyst has presented it in a way that implies the data proves this individual is a bad manager, then you have trouble. On the other hand, if the analyst has presented this as, “Here are some areas, we should talk through to see if any are problems,” then everything is fine.
The second answer is that a more senior manager may look at the case-by-case explanations and question them. If one person leaves for “unavoidable family reasons,” that’s understandable; if the manager’s whole team leaves for family reasons then maybe that’s not a coincidence. In this case, the analyst genuinely has revealed a problem that doesn’t show up when you look at departures one at a time.
More broadly, we need to accept the strengths and weaknesses of the statistical data that analysts work with. It can be misleading, but it can also provide insights. Typically, a senior manager cannot look at each individual case, so they need analytics to show them where to focus their attention. It’s essential to look at the data, We can avoid trouble as long as we’re humble about what the data seems to show.
Here are some specific takeaways for analytics professionals:
- Be prepared for a hostile reception. Analysts should not walk into a meeting thinking they are just presenting the facts and don’t need to be aware of the politics. If the data makes someone look bad, then you will be up to your neck in politics.
- Never overplay the interpretation of data. A manager might have the highest turnover, but that cannot be interpreted to mean they have the “worst” turnover. Be sure to position the data as an interesting indicator that needs to be understood in the light of the context.
- Never overplay the brilliance of the data. More often than not, analytics won’t tell managers anything they don’t already know. If you try to imply the analytics team has brilliant insights that managers don’t have, then you are setting yourself up to be knocked down. The point is that data can confirm a manager’s suspicions, put some specific quantities around issues, and occasionally draw attention to a problem that managers didn’t see. That’s good stuff and will be appreciated as long as analysts don’t try to overplay their hand.
Data looks simple and objective, it’s not; be prepared to deal with the political fallout of the complexities.