One of the best-known applications of data mining is the traditional market basket analysis, in which relationships between products are examined. Companies, especially in the retail segment, are intensely concerned with learning which products sell together. This allows businesses to promote and cross-sell products. Companies seek to identify relationships between products that may not appear intuitive, such as the classic example of beer and diaper sales being closely related, due to the fact that many men, sent to the store to pick up diapers, also bought beer while in the store.

Data mining helps companies examine all the products and determine the odds of other products being sold with that product. For example, if a business sells a jar of peanut butter, what are the other products that are likely to be sold? Specifically, what are the odds of each of those being sold? Are the odds high of a purchaser of peanut butter also buying bread or crackers? What about jelly? Those might be obvious, but market basket analysis should rank all items that are sold in conjunction with peanut butter, enabling companies to decide how to market and make product placement decisions.

Using SQL Server Data Mining, companies can analyze transactional data in either relational data warehouses or OLAP cubes to find frequent product combinations. The Microsoft Association Rules algorithm determines items that occur together and derives the rules indicating the strength of the correlations. Furthermore, beyond simple analysis, companies can apply models created with SQL Server Data Mining to produce real-time product recommendations either at the point-of-sale for coupon generators, or in an online shopping basket to increase sales of related products.