Database: Get a Little Closer
Segmentation is a way of grouping people or organizations with similar demographic profiles, attitudes, purchasing patterns, buying behaviors or other attributes to help understand customers more thoroughly and thus market to them more effectively.
The problem is many businesses use segmentation to only recognize that every customer has some unique characteristics, providing a somewhat superficial view for each.
For this reason, traditional segmentation often can be a “blunt instrument,” leading to one-to-some marketing. It’s too simplistic and lumps together customer groups that have distinct preferences and behaviors—and it can perpetuate “accepted wisdom” about customers and the market that isn’t necessarily accurate.
However, marketers that add predictive analytics to the segmentation process can generate the insight needed to more effectively and efficiently acquire, grow and retain the right customers. The result is a better understanding of what products and services customers are likely to want next.
Predictive Analytics’ Role
The approach taken by many marketers using predictive analytics can be thought of as auto-segmentation. Predictive analytics technology can discover automatically which groupings exist in customer data and find relevant patterns that are likely to be much more subtle, extracting much greater predictive insight than traditional segmentation.
This ensures segmentation is objective, insight is obtained into what customers want and how they behave, and marketing decisions made are evidence-based and result in more profitable outcomes from one-to-one customer interactions.
Predictive analytics technology incorporates data collection, statistics, modeling and deployment capabilities, and drives the entire segmentation process, from gathering customer information at every interaction to analyzing the data and providing specific, real-time recommendations on the best action to take at a particular time, with a particular customer. The result is more effective customer relationship management strategies, including advertising and marketing campaigns; upsell and cross-sell initiatives; and long-term customer loyalty, retention and rewards programs.
Navy Federal Credit Union, one of the world’s largest credit unions and one of the 50 largest financial institutions in the United States, uses predictive analytics to study its 3.2 million members’ buying habits. The credit union created more sophisticated segmentation functionality, allowing it to determine which members best matched specific product and service offerings, thus eliminating the guesswork many direct mail solutions impose.
While it was important to know what products and services members were interested in, as well as how they used them, predictive analytics allowed Navy Federal to dig deeper into its segments to better understand members’ responses and predict how they might respond in the future.
Alan Payne, manager of the member research and development team at Navy Federal, says, “With greater insight of our membership, through segmentation, we were able to gain much deeper understanding of the value points of our membership and pinpoint where members came from … so we could quickly identify what those members’ immediate needs would most likely be in that geographic space at that particular time and enable sales and marketing to respond with appropriate offers.”
Segmentation Road Map
To get the most out of customer segmentation analysis, organizations should create road maps incorporating the following steps:
1. Determine the Overall Business Objective. Get everyone on the same path and in agreement with what you want to accomplish, such as improving the yield on lead-generation efforts, identifying cross-sell opportunities or identifying customers most likely to go to a competitor.
2. Capture All Potential Customer Data. Segmentation begins with gathering customer data from a wide variety of resources, including data warehouses, point-of-sale systems and loyalty programs.
A database of static customer information is valuable, but until key active knowledge gained from feedback is applied—like preferences or motivations—there’s an incomplete picture of the customer.
Capturing feedback from any touchpoint—in any language—provides a clearer understanding of customers’ needs, preferences and attitudes, and improves the segmentation process.
3. Perform Recency, Frequency and Monetary (RFM) Analysis. To obtain the most accurate picture of customer lifetime value, organizations first should perform RFM analysis to classify customers according to: those who have spent the most—the most often and most recently; those who have spent the most—the most monetarily, but may not have purchased in a long time; those who spend the most in the fewest number of transactions; and those who spend the least, or rarely, and have not purchased in a long time.
4. Outline the Segmentation Process. Once an organization has identified customers based on purchasing patterns, it then can begin segmentation analysis to get to the core of the audience it wants to target.
The key to a successful segmentation program is to first define the many ways the results can be used. A simple approach might take the following path:
- Create customer segments to enable differential marketing programs.
- Use past purchase data and demographics to construct customer subgroups.
- Isolate key performance factors linked to long-term customer value as major data drivers for the segmentation.
- Use cluster analysis to form homogenous groups of differently valued customers.
- Use techniques such as rule induction to automatically extract the profile of each cluster.
- Align the marketing spending priorities against each subgroup.
- Link product line or category affinity to each subgroup.
- Develop marketing plans incorporating value-based budgeting and category affinity to make programs more relevant and efficient.
5. Auto-Segmentation. With a customer base more clearly defined through effective segmentation, organizations then can add predictive modeling functionality within each segment to produce greater insight that’s required to more effectively and efficiently acquire, grow and retain the right customers, and also identify fraud and minimize risk.
The modeling functionality in predictive analytics technology helps organizations accurately determine which customers best match specific offers or campaigns. By eliminating the guesswork when targeting customer groups, organizations quickly increase ROI through more efficient use of resources and reduced spending.
6. Deploy and Share Results Throughout the Business. The final step is to create an environment in which an organization can manage and automate its analytical processes and easily deploy the results across the enterprise—thus improving productivity and collaboration and increasing ROI.
This includes the ability to automate the database scoring process, publish and distribute output and reports, and integrate the analytical process into other business applications. For example, when a customer calls a call center, that agent should be able to pull up information on that specific customer and know what type of offer should be made at that particular time.
Hit the Target
With predictive analytics technology, organizations can move away from mass campaigns toward a true, one-to-one conversation with customers. Insight gained from even the most elementary analysis of customer characteristics can have profound implications on the business and result in marketing nirvana.
David Vergara is director of product marketing at SPSS Inc., a Chicago-based predictive analytics software firm. He can be reached at (312) 651-3000.