Oct | 13
Customer Segmentation: Know the Limits
When companies are fed with more and more data about their customers, the natural path to take is to target or retarget on a more personalized level - divide and conquer, or rather divide and try to conquer more than what was possible before. Often the first step is partitioning your customer base into different groups, or segments, that have some common attributes and characteristics in order to provide more customized offers - a process known as segmentation. And then you test, test, test on the profitability of those segments.
So it was interesting to see that the largest restaurant chain in the world (in terms of sales) began to detarget, as this article explains. Not necessarily taking away the option, McDonald's made the move rather to stop marketing sodas as a drink choice with its Happy Meals in an effort to push its kids' health initiative. A bold move with its family segment? Certainly, especially knowing that the Happy Meal is a long-proven worthwhile item. A smart move? We'll have to wait and see. It's enough to make me wonder: where does segmented targeting have its limits?
The first thing to keep in mind is that you must be able to distinguish whether enough disparity exists in your market to justify segmentation. A homogeneous customer base may not warrant many segments. You could waste your time and money if you do not take measures at the onset of your campaign to be sure of actionability and profitability - this is where testing is critical.
Even when you have your cohorts finalized, recognize what you can and cannot do with them. Segmentation works at a high level - your customization has to do the same. Say that you have demographic segments dividing at the age, gender, and income levels. You may have a celebrity and a successful businessman fall into the same socioeconomic segment, but do they behave the same? Think the same? Likely not. Everyone within a given cohort is on an even plane; both the celebrity and the businessman will get the same offers. As such, your strategy has to remain broad enough at the same time.
But let's target further, you say! What can you do? Conduct different types of segmentation (not forgetting still to test, test, test). Cross your demographic segments with behavioral or even attitudinal segments, and you have a more individualized scheme - a modus operandi Anthem implements more and more with its clients. Furthermore, advanced techniques such as predictive modeling can provide truly customer-level response rates if you have the tools and resources to carry it out.
Take a step back and let it sink in. It's a careful balancing act to recognize where targeting comes into play and at what level. For McDonald's, there may even be enough wiggle room to detarget on a loyal customer segment with high lifetime value. To get to that level in the first place, it's important to know the limits of your segments and strategize accordingly.
Posted by Divesh Brahmbhatt on 10/11