Your loyalty members spent EGP 48 million last year. That number sounds impressive, but it does not tell you what the program produced. Many of those customers may have bought from you anyway. A credible ROI calculation begins by estimating the behaviour the program actually changed.
Begin with the counterfactual
The central question is not how much members spent. It is how much they would have spent without the program. That missing outcome is the counterfactual. You cannot observe it for the same customer at the same moment, so you need a credible comparison: a random holdout group, a phased rollout, matched stores, or a carefully constructed pre-program baseline.
This distinction matters because loyal customers are often more likely to enrol. Research using supermarket panel data found that a simple comparison between members and non-members overstated the program effect substantially once customer self-selection was considered. Treating all member revenue as a program result gives the program credit for loyalty the brand already had.
Choose one primary behaviour before choosing the measurement method. It might be an additional visit within 60 days, a larger share of category spend, reactivation after inactivity, or a second purchase. A focused business question produces a much cleaner test than a dashboard containing twenty metrics and no decision.