The author worked with a client that markets coin and bullion products.  The customers in the numismatic industry are coin collectors and investors.  The typical customer who purchases from a big retailer or goes to the grocery store has a different motivation for purchasing a product than a numismatic customer purchasing a high-end collectible.  Their shopping behaviors and engagement levels differ from those of retail customers.  These collectors inherited their grandparents’ collection and could retell the historical or intrinsic significance of each coin in their collection.  Others see precious metals as an additional investment opportunity only.   

Most importantly, coin collectors enjoy being in a community.  Whether they are sharing the history of their collections, trading coins online, or asking a sales representative about a newfound hoard, they enjoy speaking about their passion: coins.  The author believes that understanding the customer’s calling behavior and how it affects retail sales will help improve outbound marketing and customer engagement strategies and enable the client to foster this collecting community.  The author analyzed a sample of call and sales data, clustered customers by their calling behavior, and trained multiple machine learning models to address the following objectives: 

• Segment the customers by their preferred length of conversation on the phone. 

• Segment customers by their preferred interval of calls (e.g., monthly). 

• Identify the customer’s preferred day of the week and time of day to receive a call. 

• Identify when to call a customer that leads to a sale.