The figures that show not that the program is growing, but that it is working
The most frequently presented indicators of loyalty programs are the number of members and the amount of points distributed; yet neither shows that the program is working. The program’s real questions are different:
Do members shop more often, more, and more profitably than before they joined, and than comparable non-members? Which tier, which segment, and which channel generate the highest return from the program?
Which rewards change behavior, and which merely create cost? What do the redemption rate and expired points say about the vitality of the program? Which members are at risk of churn, and
is it possible to see this in advance? What is the net return of the program when its total cost — rewards, points liability, communication, and operations — is compared with the incremental margin it creates? Loyalty analytics
is answering these questions regularly, consistently, and across all channels, and continuously improving the program’s design according to these answers.
Without analytics, the program appears before management only as a cost item, and has to be defended in every budget period. It is seen that members shop more,
but whether this stems from the program or from already-loyal customers joining it cannot be separated. The reward catalog is changed by intuition, tier thresholds are set by debate,
and the points liability creates surprises in finance. Because channel data sits in separate systems, the member’s behavior in the store, in e-commerce, and at the call center does not form a single picture, and
how the program affects channel preference is unknown. Lost members are noticed only after they are lost.
In Minerva, loyalty analytics is fed directly by the fact that membership, points, reward, campaign, sales, return, and finance data reside on a single customer identity and a single database. Member versus non-member
comparison, behavior change before and after membership, frequency, basket, profitability, and redemption rate by tier, segment, channel, and region, cost and effect by reward, the trajectory of expired points,
and the points liability reconciled with financial accounts are monitored on real-time dashboards. The program’s net return is calculated with the same yardstick as campaign return on investment,
deducting reward and points costs and communication and operations expenses. Artificial intelligence and machine learning capabilities learn from past member behavior to score churn risk in advance, identify members close to a tier transition, and
recommend the most effective reward and campaign type by segment; the results feed directly into segmentation, loyalty campaign, and reward catalog decisions. All analyses can be extended with user-defined reports and
dashboards according to the company’s own indicators.
Evaluate your loyalty program not by the number of members, but by the behavior members have changed and the net margin they have created; defend the program with data, improve it with data.