A reported return gives you a reason to investigate. Before using the figure in your business case, understand whose results sit behind the number and how the return was calculated.
When I read a claim about loyalty ROI, my first question is what the calculation includes. Revenue from members, additional revenue caused by the program and profit after program costs answer different questions. A figure labelled ROI needs to explain which question the author has answered.
For a commercial team deciding where to invest, this distinction matters. A program serving frequent customers will record substantial sales. Establishing how much of those sales resulted from the program requires a separate analysis.
What the published figure tells us
Antavo’s Global Customer Loyalty Report 2026 overview states nine out of ten program owners measuring performance reported positive ROI. The reported average of 5.3 times applies to respondents measuring a positive return. This is a selected group, rather than an average across every program surveyed. [1]
I would keep this qualification beside the number whenever quoting the finding. Otherwise, a result from programs reporting positive returns becomes an expectation for programs with different customers, costs and measurement methods.
The research offers useful context. But the published average alone gives me insufficient grounds to forecast a return for another business.
Higher member spending needs an explanation
Suppose members spend €600 a year and non-members spend €400. The €200 difference is real within this illustrative example. Its cause is still open.
Frequent customers have more reason to join because they expect to earn rewards sooner. Some of the spending difference might therefore exist before registration. Comparing the two groups after joining does not separate this existing preference from the effect of the program.
I would start by examining purchase patterns before enrolment. Were future members already visiting more often? Did their spending change after joining? Did similar customers outside the program experience the same change?
These comparisons improve the analysis, although differences between the groups still need attention. Where practical, randomly assigning eligible customers to receive a new benefit or continue with the existing experience gives a stronger basis for evaluating the change. The result concerns the benefit tested, over the period measured. Treating such a trial as proof of the entire program’s lifetime return would go too far.
Revenue and return are different calculations
Consider a trial with 10,000 customers in each of two randomly assigned groups. Over the agreed period, the group offered a new benefit spends €110 per customer. The comparison group spends €100.
The estimated additional revenue is €100,000: €10 more across 10,000 customers. These are invented figures used to explain the calculation.
At a contribution margin of 30%, before the trial’s additional reward and operating costs, this produces €30,000. If those additional costs total €20,000, the estimated net contribution is €10,000.
Under a definition of ROI as net additional contribution divided by additional cost, the return is 50%.
Dividing the €100,000 revenue increase by the €20,000 cost produces five times. Both calculations use the same trial, but describe different outcomes. Calling both “ROI” without showing the formula invites confusion.
The calculation also depends on the measurement period. A reward might bring a planned purchase forward. If the trial ends before the following period, the team risks counting an earlier purchase as an additional one. Returns, reward fulfilment and spending after the promotion deserve attention before closing the analysis.
Agree the decision before choosing the costs
For a decision about continuing an established program, ongoing operating costs matter. For a launch decision, setup and integration costs also belong in the business case. For a small campaign test, the relevant question concerns the additional cost and benefit of the change.
These are different decisions. I would expect finance and the loyalty team to agree the scope before calculating the return, including how they treat rewards earned during the period but used later.
Where a controlled comparison is impractical, a model based on historical behavior and comparable customers still has value. The assumptions should remain visible, with a range showing how the result changes when those assumptions change.
Before approving an ROI claim, I would ask for a short explanation of the comparison group, additional spending, margin, costs and observation period. A reader should be able to follow the calculation from customer behavior to commercial result.
A published average helps frame the discussion. Your own investment decision needs a calculation your team understands.
References
[1] Antavo. Global Customer Loyalty Report 2026. Publisher overview, 12 March 2026. View source