How can we align segmentation with the business view of customers?
We know that segmentation, targeting and personalization are the right thing to do - well most of the time anyway. And we know that this logic can ultimately lead us to potentially segments of one. And we know all too well that we can't manage the business at that level. So, where do we start along the path of segmentation?
In marketing 101 we should have learned that a segment is simply a group of consumers with common needs who can be reached efficiently via marketing tactics. And if we make the assumption that behavior is a reasonable substitute for needs then we can leverage purchase history (or digital breadcrumbs) as the basis for segmentation.
Now comes the challenge of sorting that data into useful buckets in order to fuel the targeting of campaigns as well as understanding their impact. A good place to start is to identify the ways the business currently views customers and use these lenses as the basis for straight-forward segmentation.
Here are some typical ways a business could look at a consumer:
• Value and transaction history
• Product and category usage
• Interaction and engagement
Each of these lenses reflect two separate dimensions resulting in a classic 2*2 matrix that is easy to explain. For example, the Value - History pair results in a variant of RFM segmentation (recency, frequency, monetary) and it looks like the following:
In this example, the break points are typically the average/median
values for each metric, i.e. above and below the average spend plotted
against above and below the avg. number of orders. We typically trim
the outliers so as not to skew the results too much.
The "Best" customers spend more and shop more frequently whereas the "Uncertain" are literally that - too few transactions and too little revenue to know where they are heading. While sounding a bit like "duh" it is amazing that few companies actually design their campaign tactics around the simple economic value of their customers.
For the second lens of Product and Category usage, the variables might be the number of categories shopped vs. the number of individual products bought per visit.
• Multi-category: high categories, high products
• Cherry Picker: high categories, low products
• Intensive Buy: low categories, high products
• Top Up Staples: low categories, low products
And the list can go on.
A set of simple segmentation schemes that cover the basic business points of view can result in a large number of potential segments. A recent project using a similar approach using four different stakeholder views resulted in 256 possible ways that a consumer could be classified. In this case, that was plenty of variations to drive the initial communication blueprint and creative needs for starting a 1:1 program.
A "Frequent - Cherry Picker" would likely get different offers, incentives than a "Spender - Top Up Staples".