When reward costs outrun referred-user quality

Generous rewards can fill a referral leaderboard and still leave the business worse off. Quality shows up later: week-four retention, second purchase, or a second invite from the same referred user. If those signals stay flat while payouts climb, the program is buying activity rather than relationships.

We encourage clients to list every reward type — credit, cash, merchandise, status — and attach a true unit cost including support time and fraud review. Many Bangkok-based apps underestimate the staff hours spent validating unusual claim patterns after a campaign spike.

A mild but useful test is to lower a reward for one invite channel for two weeks while keeping creative and copy identical. If volume collapses and quality stays the same, the previous reward was papering over a weak invite story.

Reward design also needs a clear exclusion list: employees, multi-account patterns, and self-invites. Without it, performance reports mix genuine advocacy with noise.

Back to field notes · Ask about an assessment