Brand Analytics – Branding and Gender
Text Analytics Tips: Branding Analytics – 500 Major Brands and Gender- by Tom H. C. Anderson
(Continuation from yesterday’s Brand Analytics post)
Thank you everyone who contacted us for more information about OdinText yesterday. As a result, I’ve decided to dig into the same branding question a bit deeper taking a look at one or two additional variables today and tomorrow. Of course, if anyone is interested in seeing just how easy and powerful OdinText is feel free to request info or a demo.
Yesterday we looked at Brand Awareness by Age quite a bit. Today, I thought we’d look at gender. But before that, I’ve posted another visualization from the Brand vs. Age data below.
Though popular, we’ve found at OdinText that typical word clouds are almost completely useless. Yesterday, I showed a co-occurrence plot of the data which certainly is more meaningful than a word cloud, as unlike a word cloud, position is used to tell you something about the data (those terms mentioned most frequently together appear closer to each other).
In the chart above, we are plotting two variables from yesterday’s data, Average Age on the x-axis and frequency of mentions (or popularity) on the y-axis. Visualizations like this are a great way to very quickly explore and understand unstructured data. Even without getting into the detail, often just the overall shape of a chart will tell you something about your data. In the case above, we have a triangle type shape where the most popular brands, such as Samsung, Sony and Coca-Cola, tend to appear in the middle. Two outliers are Apple and Nike who are not only our most popular brands, but also skew a bit younger.
But let’s leave Age and take a look at Gender. Though a nominal variable because gender typically only has two values (male and female), it is also a dichotomous variable and thus lends itself nicely to more advanced visualization. Basically, any dichotomous variable including Yes or No (present vs. not present) can be very useful in OdinText patented text analytics process. What can we tell about brands when we look at gender?
Brands by Gender
There are certain stereotypes that seem validated. You’re far more likely to be a guy than a gal if, when you think of brands, the first things that pop into your mind are Software and Electronics such as Microsoft (15% vs. 6%) and Sony (17% vs. 11%), or if you think of an auto brands like Ford (11% vs. 7%) or a McDonalds (4% vs. 1%).
Perhaps as expected, women are far more likely to think of consumer packaged goods brands like Kraft (14% vs. 7%), Johnson & Johnson (6% vs. 2%), Kellogg’s (5% vs. 0.4%), General Mills (6% vs. 2%), Dove (4% vs. 0.2%), and P&G (4% vs. 1%).
Interestingly the list doesn’t stop there though, women tend to be able to mention several more brands than men including many less frequently mentioned brands such as Coach (3% vs. 0.3%), Gap (3% vs. 1%), Colgate (3% vs. 0.5%), Tide (2.6% vs. 0.4%), Victoria’s Secret (2.5% vs. 0.2%), Michael Kors (2.8% vs. 0.7%), Kleenex (2.5% vs. 0.5%), Tommy Hilfiger (2$ vs. 0.2%), Huggies (1.7% vs. 0%), Olay (1.7% vs. 0%), Hershey’s (1.7% vs. 0.2%), Mattel (1.5% vs. 0%) and Bath and Body Works (1.3% vs. 0%).
That’s it for today but come back tomorrow and we’ll look at one last data point related to this one unstructured branding question we’ve been looking at to see whether brands can have a political skew.
Of course in the meantime, please feel free to request more information on how you too can become a data scientist with OdinText. Text Analytics can be a great tool for brand analytics including answering brand positioning, brand loyalty and brand equity questions.
[NOTE: Tom H. C. Anderson is Founder of Next Generation Text Analytics software firm OdinText Inc.]