One afternoon session on day three of Customer Expo 2019 featured a presentation from Jim Leone, Data Science Vice President at cxLoyalty. cxLoyalty provides loyalty and customer engagement solutions to approximately 3,000 clients and partners. During his presentation, Leone unlocked insights into the essence of personalization and the necessary processes for achieving it.
Leone first emphasized the challenge that personalization represents. He said, “We consider personalization as a gigantic monster.” Leone noted that cxLoyalty has always striven to help brands understand every aspect of this “gigantic” concept.
Leone then discussed the personalization process that his team has been working on. It features three different phases: pulling all compiled customer information, integrating that information with contextual data, and then running everything through an AI engine. In order to optimize the process, Leone stressed the importance of omnichannel content selection. This means extracting, transforming, and using all sources of data in the engine, from operational and locational to behavioral and content-oriented data.
Leone then addressed the difference between identity data and content data, which is a concept brands need to understand to deliver personalization. While identity data only includes such information as location, gender, income, and transactions, content data includes the marketing insights that result from analyzing such things as what customers look for in products and what items compliment those products. With the integration of both, brands will be able to gain in-depth views of their customers and customize their offers.
The highlight of the presentation was Leone’s discussion of a cxLoyalty case study. This discussion relayed how the company worked to revamp a client’s mobile reward app. Leone said that the process started with simple steps of initiating and categorizing different target rewards options. Next, he used a weighting system to decide the priority of a reward display. The determinants of the weighting system included level of promotion, latest trends, customers’ interests, and their segments. The points added up from each determinant ranked how each reward would be displayed to each customer, ranging from a Starbucks gift card to a sporting event ticket.
Leone made a point about the necessity of constantly measuring and testing any machine-learning model brands plan to build. He said that brands collect data to predict business patterns and bring more personalized offers to customers. So, if the machine-learning model a brand uses does not bring any values, that company has to take extra steps to meet customers’ needs.
Leone concluded by emphasizing the simple, effective AI method he extrapolated from the case study mentioned above. In short, the proper extraction of the right data combined with weighted rankings can create strong customer affinity for the rewards programs they enroll in.