Recently, a major retailer asked Lenati to assess its loyalty rewards program, which had first launched in the late 1990’s. By one metric—retention rate—the retailer’s program was still successful. But the program was failing to drive a large share of wallet, and execs wanted to know why.

Because the program had existed for so long, reports didn’t incorporate new data sets captured by more modern loyalty programs (for example, customer profiles or browsing behavior). However, even basic transactional data dating back to the 90’s revealed major opportunities to increase share of wallet and improve customer lifetime value. The answer to the problem lied within the data the retailer already had.

Even with large, sophisticated brands, we see a failure to feed data from existing rewards programs back into the design of the program itself. Even a few simple data analyses can multiply the effects of successful incentives, and reveal otherwise unnoticed places to respond to customer needs.

Answering the following questions will reveal several quick ways to use data for a rewards program refresh—and the right analysis to help you get there.

  • Are you driving the right behaviors? Well-designed rewards programs can incentivize behaviors like frequency and shopping across channels, but is your program targeting the behaviors that can best influence profit? Which behaviors correlate to your top spenders? Where are your biggest moments of attrition?

In the retail example above, a “survival curve” of members showed many first purchasers never returned to make a second purchase—an expected “one and done” problem. More interestingly, still fewer customers made a third purchase, which was new information. But after the third purchase, customers would remain loyal shoppers. For this retailer, refreshing the program with incentives to visit the store again after the second purchase has the potential to further reduce churn.

  • Are you rewarding the right people? Tiered loyalty programs often reward an extremely small group of their very best customers while ignoring very good For example, another company’s tier structure was rewarding the 3% of customers who represent 10% of their revenue. By analyzing the customer spend distribution over possible alternate tier structures, we discovered they could shift their approach to instead reward the 30% of customers who represent 55% of their revenue, making a major profit impact for a relatively small added cost.
  • What’s the customer experience like? Even before you survey rewards program members for their opinions, transactional data from loyalty programs can reveal gaps in the program experience. For example, how many customers stay within your loyalty program tiers? Analyzing transaction data at the customer level revealed churn rates out of the top tier: Almost two- thirds of members who earned top-tier status failed to achieve the same status the following year—and nearly 10% of the top tier members (and biggest spenders) never returned to shop at all. Program design can address this issue by lowering the tier spend threshold, or by specifying different requirements to maintain a tier status than those needed to earn it.
  • What’s the financial impact of changing the program? Refreshing or redesigning an existing rewards program offers a more predictable ROI than designing a program from scratch. Existing member transaction data can help you predict how real customers will react to benefits more effectively than just making assumptions with personas. And you can adjust program response rates and costs by customer segment, since customer groups will take advantage of benefits in different ways and at different frequencies. My company has developed an ROI Simulator to test the financial impacts of various program adjustments and quickly compare several program options side-by-side, revealing the optimal combination of benefits for real-world customers.

Finally, are you getting the right data? A program refresh is a great time to review the way you are collecting data. Which key business questions can’t you answer right now? What information is missing that you wish you had? Do you have a database that maps to other databases? Now is the time to build data capabilities in IT for the next round of improvements.

For related reading, be sure to check out Lenati’s New Methodology for Predicting ROI of a Loyalty Program: Using customer and data-driven simulation to more accurately project how a loyalty program will perform in the future.

Originally published on