How should you use Market Basket Analysis?
Use Market Basket Analysis to identify products customers often purchase together, then treat each pairing as a hypothesis—not proof that one product caused the other sale. Validate the idea with product relevance, contribution, inventory, customer intent, and a controlled merchandising or campaign test.
Amazon lists Market Basket Analysis among Brand Analytics dashboards. Access, data coverage, aggregation, lookback options, and available dimensions can vary by account, marketplace, brand role, and time. Confirm the current dashboard definitions before comparing periods or teams.
Build a defensible pairing shortlist
- Record the focal ASIN, paired product, marketplace, period, available co-purchase metric, and product ownership.
- Remove pairings caused by substitutes, duplicate purchases, incompatible variants, reseller activity, or temporary promotions.
- Check whether the products solve one shopper mission and can be explained without unsupported claims.
- Verify stock, lead time, return rate, contribution, price stability, and eligibility for the intended format.
- Prioritize pairings where both products create value even if the test does not increase total order count.
Choose the smallest useful test
Test the pairing in one controlled surface such as a Brand Store path, Sponsored Brands landing experience, approved bundle format, detail-page comparison, or external campaign. Preserve a comparable baseline and avoid changing price, inventory, listing content, and advertising at the same time.
A higher basket value is not enough. Measure whether the pairing creates incremental contribution after discounts, advertising, fulfillment, returns, cannibalization, and operational cost.
Measure the outcome without overclaiming causality
- Track qualified visits, units per order, combined contribution, return reasons, stockouts, and product-level conversion.
- Compare paired-product behavior with similar periods or cohorts while recording promotions, ads, seasonality, and availability.
- Check whether the second product gained incremental demand or merely shifted sales from another ASIN.
- Stop pairings that confuse shoppers, increase returns, create inventory imbalance, or require an unprofitable discount.
- Promote a pairing to a larger test only after the result repeats across a representative period.
Respect access, privacy, and catalog constraints
Brand Analytics uses aggregated customer data, so do not attempt to identify individual shoppers or infer sensitive characteristics. Use the dashboard for product and merchandising decisions within the account's permissions and applicable policies.
If Market Basket Analysis is unavailable, the brand role is not eligible, or the data volume is too thin, use order-level business records that the seller is permitted to analyze and clearly label the limitation. Do not create a bundle or variation relationship solely because two products appear together.