Clustering is a means of using algorithms to find and create customer segments by looking through hundreds of customer dimensions and millions of data points to find meaningful relationships with buying behavior. Clusters are mathematically derived and more effective than segmentation for creating groups of similar customers and describing the group’s behaviors.
However, some marketers still haven’t harnessed the power of big data and are still using traditional methods to segment their customers.
1. First Acquired Product
This segmentation is based on a customers’ first acquired product. When a customer buys a certain product, selling them similar or complementary products is easy. For example, if the first thing a customer bought at Amazon were textbooks, Amazon would try to market study guides and notebooks. Makes sense, right?
However, making recommendations based on first purchased product alone has the hidden side effect of ignoring customers’ other needs and wants. If Amazon marketed ]only notebooks to customers who bought a text book they’re missing out on the opportunity to cross-sell other products that the customer in that segment may want or need. A college student may also want to buy furniture, posters, electronics, etc. If Amazon doesn’t recognize that then they will lose many potential sales and there is also the risk of turning off the customer with repeated irrelevant recommendations.
Segmenting customers using demographics can be a great approach to segmenting the market. Demographic characteristics have to do with people’s vital statistics, and that information is easily available from third party sources, although most companies barely harness that information. You would use demographic information from your target population and use that to steer marketers in selecting appropriate advertising, and marketing channels.
You can market to different demographics that fall within statistically calculated customer segments; however, you cannot create customer personas using demographics. A person’s age, gender, education, income level, religion, citizenship, etc., don’t always correlate to buying behavior. Demographic segmentation may be useful to tailoring content, but only after the after customers have been segmented using more sound statistical clusters.
3.Recency, Frequency, Monetary (RFM)
Creating customer segments based on how recently, how frequently, and how much money they spent is one of the most used customer segmentation tool. Most businesses have that data readily available and use it to assign customer rankings. It’s a great tool for raising short-term lift for promotions but in the long term, reliance on RFM can impair predictive judgement.
Relying on RFM alone leads to the neglect of some customer segments and the overuse of others because it is more descriptive than predictive. RFM ignores other types of behaviors such as time spent on site, days between purchases, discount sensitivity, etc., that are better indicators of a customer’s future behavior. A customer’s LTV or likelihood to purchase may change & RFM has no way of predicting that change.