Is Your Loyalty Program Really Worth It?
Dear Dr. pROfIt: Our company has a loyalty program: customers get five points for every $100 they spend. Once a customer acquires 30 points, he/she receives a free gift and a $20 gift certificate. Customers have to enroll in the program in order to receive the benefits. Our CEO wants me to measure the impact of the loyalty program on sales, and thinks I should simply compare the annual spend between loyalty customers and nonloyalty customers. Is this the right approach?
Dr. pROfIt: Though it certainly is an easy approach, I don’t think it's the right one. Loyalty program customers often self-select themselves into a loyalty program because they enjoy receiving benefits, so they're inherently different to begin with.
One approach I’ve used that appears to be somewhat less biased is to “equalize” customer segments. In other words, select all customers in your database who spent between $300 and $500 through August 2009. Then split this audience into loyalty and nonloyalty customers, and measure spend from September 2009 through August 2010.
By executing this simple level of segmentation, you eliminate much of the self-selecting bias of high-value customers enrolling in a loyalty program.
This segmentation typically reveals that the impact of a loyalty program on annual sales is far less than one might think. Businesses often find that the percent increase in sales between loyalty and nonloyalty customers is reasonably proportionate to the amount of benefit the customer receives.
In the program referenced above, customers receive a 3 percent cash benefit ($20 after spending $600) and a free gift. The company is likely to see a 5 percent or maybe 10 percent increase in sales using this segmentation strategy, and after it runs a profit-and-loss statement on the program, it's likely to find a program that's close to breakeven.
Loyalty programs are frequently evaluated on the basis of biased metrics, which lead companies to believe that the programs are more effective than they really are.