What Basic Concepts Should You Master for Learning Data Analysis?(Practical Data Analysis 3)
Learn how Target used data mining to predict pregnancy, the Apriori algorithm, and the roles of BI, DW, and DM in uncovering business insights.
Welcome to the "Practical Data Analysis" Series
A Target store in Minnesota, USA, faced a customer complaint accusing the retailer of sending baby product coupons to his daughter, who was still in high school.
However, the customer later called to apologize after his daughter admitted, under his questioning, that she was indeed pregnant.
Target's distribution of baby product coupons was no coincidence. They discovered that women's purchasing habits change during pregnancy. For example, they might switch from scented to unscented lotions and purchase large amounts of vitamins and other health products.
Using this insight, Target developed a "pregnancy prediction index" through which they could predict if a customer was pregnant and send relevant coupons accordingly.
But how was the connection between pregnancy and product purchases discovered?
This was achieved using the Apriori algorithm, introduced in 1994 by American scholar Agrawal.
By analyzing product sets in shopping baskets, the algorithm identifies associations between items. Merchants can use these hidden relationships to encourage specific purchasing behaviors, boosting sales.
This demonstrates the power of data analysis, as valuable insights often emerge from data. The story of beer and diapers is another classic example.
Today, supermarkets frequently bundle products for sale, such as Procter & Gamble's offerings: Rejoice shampoo paired with Olay body wash or Head & Shoulders shampoo with Safeguard body wash.
Bundled sales are an effective marketing strategy, driven by data analysis.