Geek Time: Stats 101 for Cannabis Managers

"All models are wrong, but some are useful" - George. E.P. Box, British Statistician

It’s good to be a data-driven leader but do you really understand the insights and models that are presented to you?   If not, you may end up spinning your analytical wheels, hiring the wrong people, or making the wrong decisions.

Yesterday, I posted an article on building a data-driven enterprise.  I received several heartfelt DMs asking me what are the key statistical principles that underpin data analytics: Stats 101, if you will.

I get it.  Stats and math can be daunting especially for those that never studied it in school or are far beyond their engineering degrees or MBAs (like me)

For your weekend reading pleasure, here are 4 simplified concepts you need to know:

Mean, Median and Mode
These are the simplest but often most misunderstood statistical pillars. I frequently witness their misuse in estimating basket size, retail SKU distribution & penetration and analyzing customer satisfaction scores.  The mean is the mathematical average of a number of data points.  The median is the midpoint of a set of numbers (i.e. the 50th percentile), and is often different from the mean.  The mode is the most common (and important) occurring value(s) in a data set.

A/B Testing
This is one of the most common analytical exercises. Essentially, A/B testing compares two versions of something to see which is best. For example, you could test which two web site messages or product names resonate better with a consumer.  Or, which version of a product or promotional offer has the highest potential?  When undertaking A/B testing, you should make sure you choose the right metrics, clearly define the choices, and run the experiment long enough.

Regression Analysis
This tool enables you to find the relationship between variables. For example, you may want to know what the relationship is between terpene levels and volume. Regression analysis can help you determine WHETHER and HOW terpene levels impact volume at a given price point.  But, be mindful of a common RA landmine:  confusing correlation (two things happen together, just independently) with causation (one thing causes the other).

Statistical Significance
Once you have employed the tools, you need to determine what your results mean - if anything.  Basically, you are examining whether the experiment’s results are due to chance or from the factors you were measuring.  This is an important consideration when facing key decisions such whether to raise prices, change the product or run a new marketing campaign.

Geek, out.

#data #dataanalytics #statistics #decisionmaking #strategy