I was shamed by their good will and mortified by their cooking. There seemed to be some correlation between devotion to God and a misguided zeal for marshmallows.
- David Sedaris, c.o.g.
Thank the Muses for correlation. Planners, strategists and statisticians alike all preach ad nauseam that correlation and causation are not even mathematical distant cousins. A, however, client once put me in my place with a pithy truth on the topic: 'Correlation may not prove causation, but it's something!' My client, of course, was absolutely correct: correlation does not prove causation, but correlation damn sure is something. And not only is it something but, truth be told, correlation is mostly what we look for when we pour through pivot tables of cross-tabulated behavioral, demographic and purchasing behavior.
Put another way, not only is correlation most definitely something, but that something might very well be a sound customer insight about a targetable segment.
Davis Sedaris' claimed correlation about religious faith and marshmallow culinary ingenuity is easy to test. All one need do is create a two question survey - possibly even one question - devise a random sampling strategy and stand outside churches, temples and mosques following regularly scheduled services. The sampling strategy might be weighted towards interviewing women, who tend to be the household member responsible for food shopping and meal preparation. To be thorough, a representative sampling of zip codes might be covered so that demographic information can be layered atop the survey data to explore whether or not this alleged marshmallow moxie also extended into geographic location, educational level, and household income.
A similar study could even more easily be conducted via telephone by adding a screening survey on the topic of religious faith and a sampling strategy to yield a representative sampling of the entire United States population.
If the Sedaris correlation bore fruit the correlation between faith, marshmallows, and demographics could easily lead to a sound marketing strategy that included targeted messaging, product development (marshmallows that melt at specific temperatures for specific recipes, marshmallows mixed with different spices), and product extensions (Beyond Smores and Marshmallow Pie - The Complete Marshmallow Cookbook).
Consumer insight? Yes!
Actionable? Most definitely!
Causation? Totally irrelevant!
Which brings us to the Harvard Business Review's Daily Stat of July 16:
- Blondes earn 7% more than brunettes
- Men blond women wed earn an average of 6% more than the husbands of women with other hair colors
Consumer insight? Almost.
To take these data points and turn them into a proper, actionable consumer insight about a targetable audience for, say, a women's beauty product we would first have to determine how propensed non-blond women are to change the color of their hair either to increase their salaries, or to marry wealthier men.
Testing here is not as straight forward as testing the Sedaris correlation. To be thorough, several difficult-to-identify groups would have to be recruited: employers of at least two women where one of those women is blond and one is not; men married to blonds; men not married to blonds; and, of course, non-blond women.
Though testing could be limited only to the non-blond women group to save time and money, this direction would ignore the secondary market of men married to non-blond women who want to encourage their wives to dye their hair blond to give the appearance of a 6% higher salary.
Complicating the research even more is the fact that answers to questions on such topics as men's taste in women, income, and the lengths to which women are willing to go to make themselves appealing to men are prone to inaccuracy. But all of these risks can be minimized through rigor and methodology. In short, the study is very doable.
If ever there was a time for the women's beauty industry to take up arms against the perpetual accusations of deliberately distorting womens' collective self-image to create an insatiable demand for their products, that time is now! Right there, in the HBR no less, are two statistical data points that speak directly to quantifiable and positive effects that beauty products can have on the lives of consumers, and those effects are far more quantifiable and positive than those of virtually any other product gather dust on CVS and 7-11 shelves.
So let the testing begin! And let it reveal truths about our nature, self-image, shared values, and other deeper mysteries of the human animal.
[NOTE: Special thanks to Julie Salles, who served as wardrobe, stylist and model in the photo above, and who is first and foremost a fine photographer in her own right.]
Posted via email from Plastic Spoon's Posterous
