If your business has a loyalty program, you are probably sitting on a gold mine of customer data. Are you using those data to gain insight into your customers and improve your marketing effectiveness? A survey of retailers in the Netherlands shows that gaining customer knowledge through loyalty program data is crucial to realizing the loyalty enhancement potential of such programs. So if you have not been leveraging your program data, it is important that you start right away. In this article, I will describe six sample uses of loyalty program data.
Use #1: Customer Lifetime Value Analysis
The beginning of loyalty programs is often to make the best customers feel appreciated. But who are these best customers? Loyalty program data can help you answer that question. Based on each customer’s transaction frequency and amount, it is possible to calculate the expected lifetime value for the customer. Refer to this article for how to calculate customer lifetime value. Once you are able to assign a lifetime value to each customer, you can design offers and campaigns to ensure that your best customers’ needs are satisfied.
Use #2: Customer Attrition Risk
Related to customer lifetime value analysis, your loyalty program can also tell you if some of your customers are at risk of leaving you. This knowledge gives you precious lead time to proactively address the problem and retain customers.There are different ways of identifying such risk levels. One popular approach to predicting customer churn (the BG/NBD model) uses simply the number of transactions a customer has made, when the last transaction happened, and how long the customer was observed. This model can be implemented as an EXCEL spreadsheet and through the BTYD package in R.
Use #3: Customer Segmentation
Marketing 101 often teaches examples of market or customer segmentation based on demographic information. But in practice, such demographic segmentation is often ineffective, especially as the tastes and needs of today’s consumers have become increasingly complex. Loyalty program data allow you to segment customers by what they actually do instead of simply who they are. This is called behavioral segmentation. To do so, you will need to first create some summary metrics to describe each of your customers (e.g., frequency, timing, amount, etc.). These metrics are then entered into some clustering algorithm to identify distinct groups of customers within your customer base. K-means clustering and hierarchical clustering are both popular methods used for such purposes. K-means clustering is implemented in the kmeans() function, and hierarchical clustering in the hclust() function, both in the R stats package.
What type of customer segments you will end up having depends on the nature of your business and your customers. As an example, in one of my earlier posts, I described how to create a habit metric to help identify the habitual vs. non-habitual individuals among your customers. Combined with customer purchase levels, it is possible to segment customers based on habit and loyalty levels. My research shows that making such distinctions between loyalty and habit is very important to effective marketing.
Use #4: Customized Marketing Communications
Closely related to the first three use cases is the ability to customize your business’ interaction with customers based on loyalty program data. A straightforward example of this involves creating different offers and communication messages to different customer segments identified from use case #3 above. At a more sophisticated level, it is possible to create truly one-to-one marketing messages based on a single customer’s specific combination of behavioral patterns. For example, you can customize how you provide progress feedback to your loyalty program members based on their location on the reward path in order to generate maximum motivation.
Use #5: Basket Analysis
If your loyalty program captures not only when and how much members buy but also what they buy, you can use that data to analyze the shopping baskets of your customers. Who are your top customers for specific product lines? Do shopping basket changes indicate significant changes to your customers’ life? Are product A and product B frequently purchased together so that perhaps product B can be promoted to those who buy A? Answering questions like these can help you identify profitable cross-selling and up-selling opportunities. Data mining methods can be used for basket analysis. For example, the last question above can be answered by identifying association rules in the data.
Use #6: Testing and Experimentation
If you are on the more adventurous side, a loyalty program offers great opportunities for testing and experimentation. Marketing campaigns can be pretested with a group of loyalty program members to assess their behavioral responses before rolling out to everyone. Customized communications with program members can be leveraged to compare the effectiveness of different creatives, different offers, and different product options. The ROI of different social media messages can be measured by linking social media with loyalty program data, a great example of social CRM. The possibilities are endless. Frequent testing and experimentation will help identify new opportunities and keep your customer knowledge fresh.
Concluding Remarks
The use cases listed here are just some examples of how loyalty program data can be used to enhance your business. I hope they will inspire you to engage in more program analytics. One word of caution is to be careful when generalizing insights you’ve derived from your loyalty program data. Two reasons for this: (1) your loyalty program members may not behave the same way as your other customers that do not belong to the program; and (2) your loyalty program members may not make all their purchases through the program (e.g., making a purchase without the loyalty card).
The more of your customers belong to your program and the more complete your members’ purchases are captured under the program, the more generalizable your insights will be. You can help by encouraging program enrollment and by making point earning more automatic (e.g., without having to show a program card). Remember that insights are as good as the quality of your data. So better quality data will offer you better-quality knowledge about your business and your customers.