To a lot of marketers, statistics are just about dull, one-dimensional numbers. If they see any nuance to them, it’s in varying shades of gray along a black-white scale. So many marketers struggle to make effective use of stats in their efforts to move past the limited and limiting world of statistical gray-scale. The ones who “get it” of course, are thinking not about numbers but about the humans behind the numbers. And no, I am not talking about the overreach of behavioral targeting-I’m talking about persona development, one of the most difficult, yet (ROI) rewarding strategies a marketer can master.
Personas not only add vivid color to statistics, they breathe life into the pool of numbers marketers swim in (or drown in) every day. A useful definition is found on Wikipedia: “In marketing and user-centered design, personas are fictional characters created to represent the different user types within a targeted demographic, attitude and/or behavior set that might use a site, brand or product in a similar way. Personas are a tool or method of market segmentation.”
Tip #1 Start with the Math
The math behind persona development is pretty straightforward, though the lingo can be a bit daunting at first. Basically, personas are generally developed from large data sets of customer purchase activity, demographic attributes, primary research, etc. The most common statistical techniques are K-Means Segmentation and Self-Organizing Maps. Both of these methods are based on the concept of distance. You can view a complex data set as something called an “n-space hyper-dimensional model,” where dimensions are defined for things like demographic attributes (age, income), product attributes (class, price), velocity attributes (time between purchases, time since last purchase), and survey attributes (responses to questions). As you might guess, it’s pretty easy to get hundreds or even thousands of dimensions in a clustering data set.
Tip #2: Utilize Domain Expertise
The notion of “domain expertise”—meaning the innate knowledge that people who work in the business have about that business—is very important in persona development. Here’s why: Because clustering is based on the concept of distance, the very definition and scale of the dimensions are alterable and subjective, which can lead to confusion. Suppose you have four data points on a two-dimensional axis. The relative scale of each axis will dictate how the data points cluster. Two data points will be closer to one another with one relative scale, and farther apart with a different relative scale. This is where the fun starts, and why domain expertise is so necessary to successful persona development.
Tip #3: Understand Your Scale
Scale is determined by many things. For example, categorical data like income and age obviously have differing categorical scales. And product data can be grouped into different product classes. Suppose two people are looking at a dress on a model at a boutique fashion show. The first person might think, “That dress is for the woman who wants to be classy and confident at work,” and the second person might think, “That woman is going into a nightclub with her toothbrush in her wristlet!” Meanwhile, the merchandiser who sourced that product actually bought it for a married woman for her third anniversary.
Tip #4 Supplement and Correct
When it comes to persona development, once you let the math do the work, you need to supplement and correct it with multiple iterations through domain expertise, a.k.a. subject-matter expertise (SME). That is why any useful persona-development effort should be expected to take several months after the initial data set has been gathered. And that exercise—working with industry experts to decipher data in ways that only they can—is always some of the most interesting and rewarding work that I do at Pluris.
Why is all this so important? Simply put, successful persona development infuses organizations with the spirit and energy to truly connect with those personas—those actual consumers—and make them happy customers! To extrapolate from the boutique-fashion example above, if a merchandising group realizes from its data that it “owns” the archetypal “Samantha” consumer—think Sex and the City—then they are sourcing things that Samantha would like, and thinking of her needs in a different way than the merchandising group that owns Carrie. I’ve seen organizations create real-life stories, contests and events around their personas, to the point of fictionalizing the wedding of their persona or the birth of the persona’s first child. (You may not want to go that far!)
Tip #5: Find What Data You Can Do Without
A few other things you should know about persona development – it can provide insight into which data attributes are most important, and which you can do without. For example, with most specialty niche retailers, I’ve found that personas are generally demographically indistinct. That is, the same demographic is attracted to a brand, but people vary markedly by behavioral attributes like first-purchased product, overall market-basket characteristics and channel behavior. Of course, the more general-interest the retailer, the more demographics will come into play in coloring personas.
Persona development is essential when designing loyalty programs. I don’t know how any self-respecting loyalty-program designer could ever do his or her job without the color of personas, as each consumer can and will be driven by different needs and desires, and thus a one-size-fits-all loyalty program will typically fail.
Tip #6 Integrate Persona Information Across Disciplines
Personas can and should be used in conjunction with other marketing and analytic efforts. For example, when sending emails without varied body-copy content (though you should!), the subject line can be flavored to each persona, increasing open rates. And methods like Next Logical Product can be used in conjunction with personas to a) compete with them, b) validate them, or c) supercharge them.
With all of these advantages, I can’t think of a retail marketer, a financial-services marketer, a telecommunications marketer or, in fact, any marketer who couldn’t benefit from persona development. Because buried in the data marketer’s collection is a lot of color—amazing intel about living, breathing customers just waiting to shine through.