Updated September 20, 2023.
Product associations analysis is used to understand which items are either frequently bought together or if anything changes after a discount campaign or a change in the market.
Mostly the used cases are following:
- Building product bundles and using data to make sure bundles would have demand.
- Building campaigns, where product associations help to build campaigns with high demand.
- Planning the customer journey on the store and deciding which items should be highlighted in the store as example on the front page, top of landing pages etc.
Product Associations Dashboard
Product Associations dashboard shows product association data generated by Apriori algorithm.
Essentially, the algorithm looks for common product pairs that the customers have bought during the last 365 days. The products are not necessarily ordered at the same time.
Support tells the share of the customers who bought both products, i.e. greater support means that the association rule is applicable in more cases.
Confidence is the likelihood that product B is bought given that product A is bought.
Lift is the ratio of the observed support (likelihood that both products are bought) to the probability of both products being bought as if there was no association between the products. For example, value 5 suggests that the products A and B are bought together 5 times more often than they would be bought if the sales of the products was perfectly independent of each other.