Loyalty programs are becoming an essential part of the marketing mix for large retailers. According to data from COLLOQUY, there were 2.65 billion U.S. loyalty program memberships in 2012, with the average U.S. household now participating in 21.9 programs.
Their popularity is growing because they work. A study by loyalty platform vendor FiveStars found that over their lifetime, loyal customers spend 10x as much as average customers. Further, loyalty programs can increase the frequency of customer visits by 12 to 44 percent, depending on the type of business.
Thinking of building and launching a loyalty program of your own? We recommend this three-step approach.
Phase 1: Laying the foundation. During this phase, you must gain an understanding of the current state. In terms of where the company stands, how many repeat customers there are, are there any distinct segments of customers behaving differently, etc. Post this evaluation, determine what infrastructure needs to be put in place in order to support the loyalty program. This generally consists not just of technology, but also people and processes.
When a Fortune 500 pharmacy chain approached Mu Sigma asking for our help in developing a data-driven loyalty program, we jumped at the chance to help make an impact on their business. The chain had no loyalty program in place, and its long-term objectives were to increase customer engagement and increase visit frequency. The company wanted its program to be data-driven, rather than presenting generic offers to members. The problem was that there was no analytical framework in place.
We started by defining the key metrics for success, including member enrollment, usage, lift analysis, basket analysis and redemption rates, and implemented the necessary data engineering foundation to track these metrics, including SAS and Mu Sigma analysis tools. Then we populated the database with previously collected demographic and behavioral data on customers' past purchases (the company had collected the information, but never used it before). We used this data to identify three test markets for a pilot, determined measurable goals for the pilot, and developed three variants of the program to test.
Phase II: Test and learn. During this phase, we work to identify which loyalty approaches and promotions are the most effective.
For the pharmacy chain, we executed quick iterations of test-and-learn cycles, in order to gain insights on a daily basis. We tested various aspects of the loyalty program such as:
- Pure cash discount ($5 reward) versus coupon-linked discount ($5 off oral care)
- The right mix of base points versus bonus points
- Effectiveness of redemption options
- Response rates for various campaign designs
- Offers on in-house versus market brands
This phase involves a lot of collaboration among marketing, merchandising, finance and IT, and was critical for learning which approaches were most effective for driving satisfaction and lift while reducing churn.
Testing is the key at this stage. You need to know which approaches are most effective so you can repeat and scale them.
Phase III: Prepare for scale. Here we focus on how to take what we've learned and scale it up for a national launch.
In the pharmacy chain's case, we operationalized the various performance metrics, so there was continuous tracking of the program as it was scaled. We also created various customer segments using characteristics such as potential, channel preference, promotional response propensity, and product preferences in order to more precisely target promotions.
The end result: The chain has seen significant sales life of 1.5 percent, with consistent 10 to15 percent response rates across all test groups. Renewal rates, sales lift, and response rates have consistently remained above industry average.
For this retailer, the entire process-from the beginning of the pilot through the readiness for national launch-took 18 months. Can it be done more quickly? Certainly. We've seen retailers build and launch a program in as little as six months. The timeline is somewhat dependent on where you start; if you already have some infrastructure in place and have a culture that's analytically inclined, the process will move along more quickly.
The key consideration for making this happen will include:
- Long-term view by leadership-an understanding that the benefits of the setup will be long term.
- Focus on understanding behavior and not just quick wins.
- Focus on understanding how consumption of insights can be better enabled.