Every brand eventually asks the same question: how do we increase the lifetime value of our customers?
It's the right question. Repeat purchase rate, purchase frequency and customer lifetime value are the metrics that compound over time to create meaningful business growth. A customer who buys twice is worth more than a customer who buys once. A customer who buys three times is worth even more.
The market is full of products that promise to increase customer frequency and loyalty. Gift cards, discounts, rewards programs, sign-up incentives and countless variations of each are all positioned as solutions to the same problem. Most report impressive results. Many report increases in repeat purchases. Yet surprisingly few explain how they determine whether those outcomes were actually caused by their product.
Loyalty has a denominator problem
When a vendor reports that 40% of customers who received a gift card came back within 60 days, the first question should be: what was the repeat rate before the program existed?
In high-frequency categories, a large share of customers return within that window regardless of any intervention. Reporting that they returned after receiving a reward is not a loyalty outcome. It is a description of existing behavior with a reward placed in front of it. The gift card may have had nothing to do with it.
Then there is the attribution window. A customer who receives a gift card and buys again three weeks later may have been driven by the card, a promotional email, a retargeting ad or their own purchase cycle. Isolating the variable requires controlled testing, meaningful sample sizes and enough time for behavior to develop.
Most loyalty claims skip both questions. The number sounds compelling in a slide, but without the denominator, the control and the window, it says very little about whether anything incremental happened.
The incentive and the measurement are both part of the product
This is where most reward programs break down. They can tell you a customer came back. They cannot tell you whether the incentive is why.
Every incentive carries a cost. If a program cannot separate customers who would have purchased anyway from those whose behavior genuinely changed because of the reward, there is no way to know whether the program created value or simply took credit for behavior that was already going to happen. A repeat purchase rate without an incrementality measurement is just a number with the most important context removed.
Building that measurement requires integrating directly with transaction data so you observe real purchases, not proxy metrics like engagement, redemption rates or email opens. It means defining the behavior you want to influence before launch and measuring against a control group so you can see the gap between what happened and what would have happened without the incentive.
When those elements are in place, the numbers look different. A 6% increase in repeat purchases among a targeted cohort, where you can prove those purchases would not have occurred otherwise, is a foundation you can build on. A reported 40% return rate with no control and no baseline is not.
Loyalty and incentives require precision
Part of the confusion is that the word "reward" gets applied to very different things solving very different problems.
A gift card applied at the point of purchase helps someone complete a transaction they were already considering. That has real value, but it is a conversion mechanic. A reward delivered after a purchase gives a customer a funded reason to come back. Whether they actually do depends on how relevant the incentive is, how easy it is to redeem and how naturally the customer returns to that category. Measuring the real impact still requires a control group, because without one you are watching customers do what some of them were going to do anyway.
The hardest version of this problem is a targeted incentive tied to a specific customer behavior (buy this brand, hit this basket size, return within this timeframe). That is the most direct path to a meaningful change in customer behavior, but it requires customer data, a cost model the business has signed off on and attribution that closes the loop at the point of transaction.
These are different products solving different problems. Treating them interchangeably is how brands end up measuring the wrong outcome and calling it loyalty.
Where this leaves us
Increasing customer lifetime value is a worthy goal. It deserves scrutiny proportional to the claims being made.
When a vendor reports that a large percentage of customers come back within a set window, the right response is to ask what the baseline was before the program existed, and whether incrementality was measured. Those two questions separate programs that can prove their impact from those that cannot.
