A Segmentation Primer
What is segmentation, anyway, insofar as Web analysis is concerned?
What is segmentation, anyway, insofar as Web analysis is concerned?
These days, “segmentation” is a hot Web analytics topic. But what do people mean when they talk about segmentation? It’s a word used more often than is understood, at least in the marketing sense.
One of the largest, most successful Web analytics systems vendors has a report section called “Segmentation.” Yet it contains reports on the site’s most popular pages and sections. I’m not sure what that has to do with segmentation. Other vendors talk about segmentation as well, but they mean different things. Sometimes they mean the ability to filter along different dimensions or to analyze the data by combining different variables.
So segmentation could mean reporting particular data, filtering data, or analyzing data. All these things are good and potentially useful, but are they segmentation?
I dug out some of my marketing textbooks to see if there was a consensus definition of segmentation. What the books talk about is segmentation as a means of identifying different groups of people to develop different strategies for each group. Segmentation, then, is a purpose rather than an outcome. It’s the difference between classification, which is what a lot of analysis tools do, and segmentation, which is what marketers and marketing analysts do.
The point of segmentation is you do something as a result of having it. For example:
In one book I looked at, “Marketing Decision Making,” written about 20 years ago, the authors describe three conditions of good segmentation:
That all sounds pretty theoretical (it was in a textbook). What does it mean in practice?
A good segmentation must be robust, useful, and actionable. There are many ways you might segment your customer base, from simple classification approaches to complex statistical techniques, but they must pass the sense check of being robust, useful, and actionable.
Let’s say you classify your customer base into males and females. The segmentation is only robust and useful if men and women exhibit differences that are potentially useful to you and only actionable if you can realistically target them in different ways.
Alternately, you might develop a segmentation based on attitudinal variables. Many years ago, I was involved in a project where we segmented visitors to our European sites according to their attitudes about online shopping and their motivations for visiting. Though results were interesting and highlighted some differences in different sites’ visitor profiles, we had to question how useful the information was to us. How could we act on the insight? We couldn’t identify and classify people arriving on the site by their attitudes. Nor could we easily use the information in our retention marketing activities, as we didn’t have people’s attitudes stored in our customer database.
Satisfying those three conditions of homogeneity, parsimony, and accessibility in a good segmentation is a balancing act. In our own work, we tend to use behavioral segmentation approaches, as they make acting on the outcomes easier. We might use statistical methods such as cluster analysis to segment customers into groups that are distinct in a meaningful way, such as their browsing or purchasing behavior.
We are also mindful of the client’s ability to act on the results. There’s no point in developing a sophisticated methodology that identifies some really meaningful segments if there are neither the skills nor the tools available to realize the opportunity. If your email tool isn’t easily integrated into your customer database, for instance, it’s difficult to execute improved target marketing initiatives. Start with something simple and develop the capabilities to act in line with the development of the insight itself.
Next time: a look back at 2005, and a sense of where we may be heading in 2006.