Constructing Data-Driven Personas for Effective Utilization in Social Network Service Activity Data

In today’s rapidly changing world, service operators must quickly grasp user requirements and modify services. Personas, which are virtual user images, are used to plan, modify, and develop software content. Previous studies have proposed data-based persona construction methods, but their applications to essential services are limited. Herein we propose a method to systematically utilize personas for planning content in social network services. Our method creates personas based on service log data and applies them to essential services by dividing users into groups using K-means clustering and building average and outlier user personas in the service. Then our method compares personas of average and outlier users. To evaluate the proposed method, we measured the click-through rate of users to the content that the service producer planned based on personas.