Slicing and Dicing

…enough to leave a thousand cuts

The phrase "slice and dice the data" usually gives me chills because it implies that effective people analytics is essentially looking at a series of "cuts" of the data. "Can we look at this by _____?", where the blank is filled with "location", "job level", "department", etc., often seems like an exercise in satisfying someone's curiosity, rather than increasing the impact of the analysis.

Doing these extra cuts on top of an already-solid analysis gives us a powerful tool, though - the ability to focus our recommendations on a smaller set of people. If our analysis finds that X is a major driver of "bad thing Y", that's useful! But it's even more useful if we can tell our leaders that X especially drives Y in, say, our early-career sales team. That gives them a much clearer path to action.

I'm not recommending that we entertain every follow-up question on our analyses. There were a few people who commented on LinkedIn that it’s important to have a hypothesis in mind before going off to do random cuts. That helps with prioritization, and is worth doing: when we've already done the difficult data cleaning and set up the analysis framework, it might be worth building in 10% more time into the project plan to make the analysis 50% more useful to stakeholders.

Alexis Fink made a similar point recently:

"you won’t always have time to add a bunch of extra analyses or insights or automation or whatever would elevate the question - but... often once the data are cleaned, the time required to do the incremental analysis isn’t that much" (link).