Originally shared by Avinash Kaushik
Analytics 101: Averages suck, distributions rock! goo.gl/fmhdCe
If you query how much alcohol American's drink on average, the answer is four. But that does not tell you much.
Turns our, 30% don't drink any alcohol (I'm relieved, I'm not weird!) and the top 10% (24 MILLION!) drink 74 drinks per week, or a little more than 10 per day.
Now, that is interesting. Lots of good questions arise (primarily, #omg !), lots of interesting and specific avenues can be taken for data analysis or programs to assist the top 10%, if they would like that.
It also brings about media/advertising/analytics challenges. If just 10% of the people are responsible for over 50% of the alcohol consumed, and 30% don't drink at all and the next 30% barely any, why is alcohol advertising so pervasive and wasteful in its targeting? Can't all those marketing dollars, those customer retention and loyalty dollars be spent more efficiently, or just go to reducing cost of the product?
If you take off the business lens and put on your human lens, you can also see that if our peers are having 10+ drinks a day they have a challenge. How should alcohol businesses react to that? Today they even more heavily court that audience for business profit.
The article on this graph is here: http://goo.gl/vcq9ct It ends with this astonishing fact: "If the top decile somehow could be induced to curb their consumption level to that of the next lower group (the ninth decile), then total ethanol sales would fall by 60 percent."
While business would suffer, some job layoffs will happen, think of the impact on the quality of lives.
#foodforthought or #drinkforthought :)
Comments
Post a Comment