Hey! Just a heads up I've moved to pcward.github.io!

2. Remember K.I.S.S.

When attacking a data-driven business question, keep it simple, stupid!

At least do this until you have a good understanding of the problem, and your most important stakeholders are on board and do too.

Doing this accomplishes a few critical things:

One, you save yourself a lot of unnecessary work if your analysis isn’t what your stakeholders want.

Two, keeping your analysis straightforward builds trust in your recommendations and findings. People won’t get the impression you’re trying to pull wool over their eyes.

And three, there’s usually just not a really good reason for  sophisticated statistics in most real-world analytics. At least not where the field stands today. Besides, why start an (unnecessary) arms race when you don’t have to? 

Clarity trumps sophistication, every time.