Netguru released a loyalty program ROI measurement framework in 2026 that teaches physical product brands how to measure the true incremental value of their loyalty programs using holdout testing. According to the published framework, the method isolates program performance by comparing member behavior against a randomly selected control group that receives no loyalty benefits, then calculates the net financial gain after program costs.
The framework centers on a three-part formula: calculate the incremental revenue generated by loyalty members versus non-members, subtract the full cost of running the program (platform fees, rewards expense, labor), then divide by program cost to yield ROI as a percentage. Netguru's worked example shows a $500,000 loyalty program generating $1.2 million in incremental revenue for a 140% ROI. The critical step is the holdout group—a statistically significant sample of customers randomly excluded from the program who serve as the counterfactual. Without this control, brands mistake natural repeat purchase behavior for program-driven lift.
This works because most loyalty programs suffer from selection bias: high-value customers join loyalty programs because they already intended to buy again. The revenue looks impressive until you realize those customers would have returned anyway. The holdout test isolates true causality. Netguru's method requires splitting new customers at acquisition into two groups—loyalty-eligible and control—then tracking purchase frequency, average order value, and total spend over a defined period (typically 12 months). The difference in behavior, multiplied by the size of the member base, yields incremental revenue attributable solely to the program.
The framework also accounts for full program costs that many brands overlook: technology platform subscription, customer service overhead for point inquiries, discount and reward fulfillment, and internal labor to manage campaigns. Netguru's worked example includes a $200,000 annual platform cost, $250,000 in reward redemptions, and $50,000 in labor, totaling the $500,000 denominator. The discipline of this accounting prevents brands from celebrating gross revenue increases while running loyalty programs at a loss.
A small physical product brand can run this test with basic tools and modest scale. At acquisition, randomly assign 10% of new customers to a control group—tag them in your CRM but do not enroll them in the loyalty program, do not email them point balances, do not offer member-only discounts. Treat them as regular customers. After 12 months, pull purchase data for both groups: calculate average orders per customer, average order value, and total revenue per customer for loyalty members and for the control group. Multiply the per-customer revenue difference by your member count to get incremental revenue. Subtract your annual loyalty costs (Shopify app subscription, reward costs, time spent). If the result is positive and exceeds your cost by a meaningful margin, the program works. If the difference is small or negative, you are spending money to reward customers who would have bought anyway.
The method scales down cleanly. A brand with 1,000 new customers per year can hold out 100 as a control group—statistically sufficient for directional insight. A $30/month loyalty app, $2,000 in annual reward costs, and 10 hours/month at $50/hour (email management, customer support) totals roughly $8,360 in program cost. If loyalty members spend $180/year versus $120 for control, the $60 per-customer lift across 900 members yields $54,000 incremental revenue, minus $8,360 cost, for a 546% ROI. If the lift is only $10 per customer, the program generates $9,000 against $8,360 cost—a 7% ROI, barely worth the operational friction.
The broader pattern here is the shift from vanity metrics (total loyalty revenue, member count, engagement rate) to true incrementality measurement. Holdout testing is becoming table stakes in performance marketing, and loyalty programs are finally being held to the same standard. Brands that cannot prove incremental lift will find their loyalty budgets reallocated to channels with clearer attribution.