Getting started with segmentation
Posted by marketingfrontier on November 17, 2006
Share of Wallet is a critical metric, but like most single metrics, it has limited value. True, Share of Wallet will prioritize targets for marketing activities based on size of opportunities. However, the fallacy comes from assuming that all the customers with the same size wallet opportunity are the same.
The differences between customer groups, usually referred to as segments, provide critical insight for marketing programs, product assortment, business processes, etc. that can change the fundamental nature of the business’s relationship with customers. In addition, segmentation can provide the foundation for metrics that actually measure changes in the dynamics of the customer base.
Segmentation can be a daunting task, however. How do you get started?
As with most marketing analytical activities, the key is not to overly complicate the activity at the outset. As the team becomes more experienced, you will find that segmentation becomes more precise and segments smaller and more carefully defined. But to start, you do not need a PhD nor a series of high-powered multi-processor computers. A simple laptop and a decent query language will be enough to get started. As a marketer, you might only have one of the two! 🙂 But I am sure you can find someone in the organization who can help you run queries against the database(s). Knowing the technical languages is not the hard part — asking the right questions is.
So where to begin. We always start with spending, in gross profit if you can, net revenue if you cannot get to the margin number. Take those customers and sum all their spending for the past 12 months and then divide those customers into deciles (by 10% ranges). Then identify frequency patterns, average order value, average net sales, average gross profit, preferred channel, etc. for each one of those deciles. Examine the patterns — how different are the top decile customers from the average or from lower deciles? What seem to be the biggest differences and what does that tell you about the customer base.
I then examine product purchase patterns within the deciles, since product patterns often describe differences in how customers are using your products. Examine concentrations of product purchasing among the higher decile segments and contrast those patterns with lower decile behavior. Often significant differences exist between top and lower deciles in product preference.
But you are only getting started at this point. Now here comes the fun part…