Most retailers launch a points program, watch enrollment numbers grow, and conclude the program is working. They rarely measure whether the program changed customer behaviour or simply rewarded behaviour that was already happening. The fundamental question a loyalty program must answer is: does membership increase purchase frequency, basket size, or both, beyond what the customer would have done anyway? If the answer is no, the program is a cost centre, not a revenue driver.
Why Most Loyalty Programs Fail to Move the Needle
Loyalty programs are one of the most universally adopted tools in retail, and one of the most frequently misunderstood. Most retailers launch a points program, watch enrollment numbers grow, and conclude the program is working. They rarely measure whether the program changed customer behaviour or simply rewarded behaviour that was already happening.
The fundamental question a loyalty program must answer is: does membership increase purchase frequency, basket size, or both, beyond what the customer would have done anyway? If the answer is no, the program is a cost centre, not a revenue driver. You are giving discounts to customers who are not influenced by them.
Research consistently shows that the top 20 percent of customers by spend generate 60 to 80 percent of revenue for most retailers. The goal of a loyalty program is not to reward these customers for behaviour they have already demonstrated. The goal is to move customers from the second and third quintile into the first quintile behaviour pattern, and to slow the drift of at-risk customers before they lapse.
That requires a program designed around behaviour change, not benefit delivery. Points per dollar spent rewards existing behaviour. Tier upgrades triggered by frequency changes reward and reinforce new behaviour. The design difference is significant.
Points, Tiers, and Cashback: Choosing Your Reward Structure
Most loyalty programs use one of three reward structures, or a combination: points-for-purchase, spend-based tiers, and direct cashback. Each has different psychological and financial properties.
Points-for-purchase programs award points for every dollar spent, redeemable for discounts or free products. The appeal is simplicity and ongoing engagement. The risk is that points become expected, and customers feel penalised when they cannot redeem them. Points programs work well when redemption is easy and the value proposition is clear. They fail when points accumulate without a clear redemption path, or when the points-to-discount conversion is so low that customers stop tracking their balance.
Spend-based tier programs (Bronze, Silver, Gold, Platinum) create aspiration and status. When a customer is 200 points from a tier upgrade, they are far more likely to make a purchase than when they are 1,200 points away. The tier gap creates urgency. Tier programs also allow retailers to concentrate benefits on the highest-value customers without giving the same benefits to occasional buyers. The risk is program complexity: too many tiers or complicated qualification rules reduce participation.
Cashback programs (earn 5 percent back on every purchase, credited monthly) are the simplest structure and often the most financially transparent for both the retailer and the customer. Cashback is easy to communicate and easy to measure. The limitation is that cashback provides no aspirational element and does not create tier-based urgency.
Hybrid programs that combine a points earn rate with tier-based benefits tend to outperform single-mechanic programs. Points create ongoing engagement; tiers create aspiration; bonus point events (double points on Tuesdays, bonus points for trying a new category) create urgency and frequency.
The Data You Need to Run a Loyalty Program That Works
A loyalty program without purchase history data is not a loyalty program. It is a discount card. The difference is whether you can identify individual customer behaviour patterns and respond to them.
The minimum data requirements for an effective loyalty program are: purchase history at the transaction and SKU level, purchase frequency and recency, average transaction value, and category preferences. With these four data points, you can segment your customer base, identify at-risk customers before they lapse, and build targeted interventions that are relevant enough to change behaviour.
Customer lifetime value (CLV) modelling takes this further. CLV estimates the total revenue a customer will generate over their relationship with your business, based on their historical frequency, spend per visit, and retention rate. Customers with high CLV and high churn risk are the highest-priority segment for loyalty interventions. Identifying them requires combining recency, frequency, and monetary value data (RFM analysis) with a model of churn probability.
The practical challenge for most retailers is that this data is spread across their POS system, their loyalty platform, and their email marketing tool. When these systems are separate, the data analysis that drives effective loyalty management requires regular manual exports and reconciliation. When they are unified, the signals are always current and the interventions can be automated.
- Transaction history: date, SKU, quantity, price, and channel for every purchase
- Recency: days since last purchase
- Frequency: number of purchases in last 90 days, 180 days, 12 months
- Monetary value: average transaction value and total spend
- Category preferences: which product categories the customer buys from
- Churn risk score: probability the customer will not return based on recency and frequency trends
Automations That Drive Loyalty Program ROI
The highest-ROI loyalty program automations are triggered by specific customer behaviours, not by calendar dates. Calendar-triggered campaigns (monthly newsletters, seasonal promotions) go to all customers regardless of their current behaviour. Behaviour-triggered campaigns go to customers at the exact moment their behaviour signals an opportunity or a risk.
Win-back campaigns trigger when a previously active customer has not purchased in 60 to 90 days, depending on your category's typical purchase cycle. A single win-back offer with a meaningful incentive, sent at the right time, can recover 15 to 25 percent of lapsed customers. The key is the timing: too early and the customer does not feel the gap, too late and they have already formed a habit with a competitor.
Tier upgrade nudges trigger when a customer is within a defined threshold of the next loyalty tier. A message showing the customer exactly how close they are to their next tier, with a specific purchase suggestion that would close the gap, drives incremental visits. This is one of the highest-converting loyalty automations in the playbook.
Birthday and anniversary campaigns are industry standard but still effective when the offer is genuine rather than symbolic. A meaningful discount on a birthday visit, especially one that arrives a day or two before the birthday to trigger planning, outperforms a post-birthday acknowledgment that arrives after the spending moment has passed.
Category trial campaigns target customers who buy frequently in one category but have never purchased in an adjacent category. A customer who buys running shoes regularly but has never bought running apparel from you is a high-probability conversion target. A targeted offer in the untried category, tied to their purchase history, is far more relevant than a general promotional email.
Measuring Whether Your Loyalty Program Is Working
The right metric for a loyalty program is not enrollment or points issued. It is the difference in purchase behaviour between loyalty members and non-members, controlling for the selection effect that higher-frequency customers are more likely to join in the first place.
A clean measurement framework compares purchase frequency, average transaction value, and 12-month retention rate between members and non-members in matched customer cohorts. If members are not meaningfully different from non-members on these metrics after controlling for pre-enrollment behaviour, the program is not changing behaviour.
Redemption rate is also important. Programs with very low redemption rates (under 20 percent of issued points ever redeemed) suggest the program is not engaging. Either the earn rate is too low, the redemption value is not compelling, or the redemption process is too complex.
Finally, measure incremental revenue contribution: for every dollar you spend on loyalty rewards, discounts, and program operations, how much additional gross profit does the program generate? This requires comparing the margin contribution of loyalty-driven purchases against the cost of the program. Programs that generate less than $2 in additional margin for every $1 in program cost are not economically justified.
A loyalty program that rewards existing behaviour without changing it is a cost, not an investment. The programs that generate real revenue lift share a common design: they are built around customer data that reveals behaviour patterns, structured around tier mechanics that create aspiration and urgency, and automated to reach customers at exactly the right moment with exactly the right incentive. The infrastructure requirements are a POS system that captures transaction-level data, a customer profile system that unifies purchase history across channels and locations, and an automation layer that acts on behaviour signals in real time. With those three components, a loyalty program can consistently move customers up the frequency and spend curve rather than just rewarding the ones who are already there.
Loyalty Built Into Your POS, Not Bolted On
Momentum's customer and loyalty module gives you unified customer profiles across all locations, automated tier assignment based on spend thresholds, CLV and churn risk scoring, and behaviour-triggered automations built from your live POS transaction data. No separate loyalty platform required. See how it works.
