How do I enter a business field?

What statistics should one be familiar with for business intelligence and web analytics?

• For someone from, say, the academic field who is looking to enter the business intelligence and analytics field, what statistics should he or she be familiar with? P-values, T-tests, and segmenting groups comes to mind immediately, but what else?

• Answer:

  Multivariate regression, mostly linear. Understanding linear models is important. Understanding which models to use when your predictor variables are continuous vs discrete is also important. Bayesian methods will be helpful for developing supervised learning algorithms. I use logit regression a lot because the type of questions I have to answer generally lend themselves to that model (ie if we change A but leave B and C the same, what happens to Y?) In web analytics you probably won't dig too deeply into advanced statistics; like another poster said, being able to track simple univariate statistics will probably be enough. But it's important to note that I see people mis-using even these all the time (like taking averages of averages, or reporting averages when the dataset exhibits significant skew or kurtosis, or using OLS regression when the dataset is significantly heteroskedastic). Avoid the urge to over-regress, under-sample, and immediately spit out averages and you should be ok.