Loyalty glossary · 6. Retail and grocery (20)
Basket Analysis
Basket analysis is the study of what members buy together in one transaction, used to set earn rates, design tier thresholds, and choose redemption offers. It separates items that drive margin from items that only travel with them, and it should change the loyalty programme.
Basket analysis is not a loyalty programme's academic exercise. It is the method for separating items that earn their reward from items that merely travel with them in the same transaction. Most programmes reward every item identically, which is the first error in grocery and retail.
Margin variation inside one basket is large enough to break flat earn. A 60 dollar basket often holds a 5 percent margin product next to a 25 percent margin product. The programme that gives the same earn rate on both is treating both as equally profitable, which no finance team believes.
Work the numbers. A 50 dollar basket with 20 percent low margin. Flat earn at 2 cents per dollar costs 1 dollar per basket. Weighting by margin gives 0.4 cents per dollar on the low margin portion, 10 dollars, and 0.9 cents per dollar on the high margin portion, 40 dollars. The total cost is 0.4 dollars, a 0.6 dollars saving per basket without changing a single member behaviour.
Activity-based qualification turns this arithmetic into policy. Instead of requiring 500 dollars of spend to reach a tier, a programme can require a basket score built from margin weighted purchases. A member who buys 500 dollars of high margin goods earns qualification faster than one who buys 500 dollars of loss leaders, and the liability follows real economics.
Once earn rates vary by category, the accrual for unredeemed points no longer equals a flat percentage of issued points. A programme that wants to know its true liability must use an actuarial model that forecasts redemptions by margin cohort, not by total balance, because 1 million points issued on a 30 percent margin basket and 1 million points issued on a 5 percent margin basket carry different future costs.