Topic-Based Segmentation in Email Marketing

Abstract

Email is one of the most common marketing channels businesses use to reach potential customers and advertise their products or services. To maintain the relevance of email messages and the performance of email marketing, personalization, or information from everyone is usually needed. This personalization is often derived from transactional data that reflects the behavior and interests of each customer. However, not all industries have enough transactional data for each of their customers. The challenge is how to use non-transactional data to discover customer interests and preferences. This study proposes to model customer preferences and build customer segmentation in email marketing based on topics and historical email interactions with customers. Biterm Topic Model (BTM) is used in this study for topic modeling because it is suitable for short texts such as email titles. Then segmentation is built based on the generated topics and customer email interaction history. Topic-based segmentation can increase the open rate by nearly two times on average. Marketers can use the findings of this study to develop new strategies and achieve higher email marketing performance.