Application of Statistical Analysis of Extreme Values to Customer Journeys

• Activity Number: 129
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economic Statistics Section
• Author(s): Rainhard Bengez*+
• Keywords: statistical analysis of extreme value ; customer journey ; Poisson process ; non-stationary models

Abstract

A 'Customer Journey' is an important web 2.0 marketing concept. This concept embraces all points of contact a possible customer has with a certain brand or product. For a web based company such points of contact are search engines or search phrases, web sites and even parts of advertising banners on web sites. To be in and on the right (online) place, and to react quickly, when a customer's preference for certain channels vary, are essential for these companies. In the present work we have applied statistical analysis of extreme value to these near real-time customer behavior (clicks and request of web sites stored, e.g. in cookies). We have modeled the exceeding and shortfall of thresholds. The exceeding has been specified for each point of contact with a 2-dimensional Poisson process and non-stationary models for parameters of location and scales. For many points of contact we could identify positive trends with high evidence. However, spatial analysis of the combined trend estimations within a model similar to a hierarchical model, showed big (seasonal) differences for the trend estimation. Hence, we have applied different concepts to model this behavior with higher accuracy.