Using Cross Tab Analysis to Find Relationships in Marketing Efforts
By Kimberly Grant
Sure. Relationships are hard. They require patience, analysis, checking in with others, and determining their value. Now, this theory applies to human relationships and marketing. When trying to figure out and understand the way humans think, marketers have tried many different ways of analysis. For instance, taking a survey these days isn’t just saying “yes” and “no” a bunch of times, like the ones developed via Qualtrics. There’s a lot more work involved; like human relationships.
Cross-Examining
As an example, if a marketer wanted to determine if there was a relationship between favorite colors and soft drinks, they need merely to run a survey, then a cross tab analysis. For the uninitiated cross tab analysis is not like that time you had two crushes on a pair of friends. No, no. A cross tabulation is an analysis of relationships that don’t involve a therapy session. This type of tabulation determines if a relationship exists between two variables, like consumer purchasing habits and favorite color.
Will They? Or Won’t They?
Getting back to our example, let’s say Grant Soft Drinks, LLC decided to make their own cola-type soda. Grant Soft Drinks would need to analyze the competition. When it comes to p values (relationship-wise) the company would want to know about product design. A research question for this design template could be: Is there a correlation between those who like the colors red and blue and their preference of Coca Cola vs. Pepsi. A hypothesis could be: People who like red are more likely to buy Coca Cola, whereas people who like blue are more likely to buy Pepsi.
After conducting a survey of soft drinkers and asking about their favorite colors and their favorite brand of cola, marketers would run a crosstab analysis to find out if there is a relationship between color and cola purchase. Analyzing an observed frequencies table of data, Grant Soft Drinks, LLC may see a correlation that supports the above hypothesis. However, they wouldn’t know for sure.
Making it Official
Surety would lie in running further analysis with a CHI square test of frequencies to find a statistical significance. If the p value in a CHI square test is lower than .05, there is now statistical significance in their hypothesis. Not to mention, the marketer would have bragging rights. Their hypothesis would prove correct that the color design of the soft drink can attract certain types of people. Further analysis would be needed to capture color associations in order to make a final decision on the color choice of the cola can.
When it comes to relationships and analysis, it’s not always “yes” and “no”. There’s more work, patience, and thought that goes into it. And, when the relationship status is questioned, there’s always statistics.
To further develop your relationship with K. A. Grant’s blog, you can subscribe to her newsletter below or click here for her previous post.
