Loyalty glossary · 7. Data and analytics (31)
Average Order Value
Average Order Value is the mean spend per transaction across a member or segment, used to benchmark engagement, set earning thresholds, and model liability.
Average Order Value is a simple division, but the loyalty industry treats it as a strategic number. Programme managers use it to set accrual rates and to decide which members are worth a premium. The problem is that a single average erases the shape of the underlying spend distribution, and a programme built on the average is built on noise.
Work the sensitivity. An actuarial model that assumes a 50 dollar average order value and a 1 point per dollar accrual will book 50 points per transaction. Change the assumption to a 60 dollar average order value and the model books 60 points, a 20 percent increase on the same member set. No behaviour changed, only the assumed mean did.
Activity-based qualification makes the trap explicit. If a programme rewards members only when an order exceeds a threshold set from the average, members learn to split baskets. The average rises, the threshold rises, and the programme pays more for the same total spend. Qualification built on average order value is a target, not a filter.
The median would be a better benchmark. A segment with a few small transactions and one enormous one has an arithmetic mean that flatters the group. Any tier or partner offer priced off the average overinvests in the many low-value members to chase one outlier. Use percentile bands instead of a single mean.
Accrual design compounds the error. A flat earn rate per dollar makes average order value irrelevant to points earned per transaction, but a graduated rate that pays more above a threshold turns the average into a pricing lever. Operators who recalibrate accrual tables on a shifting average are pricing their liability on a number that moves with every outlier.