Email : Segment Experiments
Stop thinking ‘email list,’ start thinking ‘customer community’January 2012 By Tim Watson
The Email Statistics Report, 2009-2013 (published by The Radicati Group in 2009), estimated that there will be 1.9 billion email users receiving 507 billion messages by 2013. Yet, the idea that you can simply build a list and send everything to everyone on it has continued. The planet may have 7 billion residents. But, as platforms like Facebook have shown, community is key.
Where marketers once thought, "Customers subscribe to lists so we can talk at them," marketers should now think, "We are building a community around our brand, and customers are joining that community by giving us permission to communicate with them."
The "send to a list" paradigm needs to be replaced by carefully crafted messages to highly targeted segments of individuals within a community.
The job of the marketer is not to define segments, but to understand how individuals in communities segment themselves. Rather than enforce artificial divisions based on what the marketer thinks, the better approach is to understand segments that occur naturally. Customer intelligence is the key to this understanding.
Take, for example, a car dealership. Buyers can be segmented according to budget, income or gender. Or by vehicle specifications—size, category, safety features, equipment specifications, environmental credentials and aesthetic appeal. With customer intelligence, the car dealer can better understand what is and what isn't important. The dealer may even find that gender isn't a strong determining factor, but beverage holders are.
One excellent way to identify segments is through customer purchase behavior. Analysis of prior purchases can show whether the customer is driven by brand, article category, gender, price and so on. The marketer needs to create hypotheses about the different combinations of factors that might define each segment and, by analyzing actual data, verify or reject those hypotheses. If verified, then a new segmentation criterion is defined.
Imagine a pizza delivery service is promoting a new jalapeño pizza. A marketer might assume that males favor spicy pizza, and plan to distribute an early evening message to the men in its database. To truly maximize the tools to maximize ROI, a better idea would be to run an analysis of single purchases looking for the combination of "male + spicy" and compare that against "females + spicy" or "males + not spicy." If the results demonstrate that men order proportionately more spicy pizza than the rest of the population, then the hypothesis is verified and can be used for segmentation.