1
UNDERSTANDING USERS THROUGH PERSONA IDENTIFICATION The Human Experience of Music Information Retrieval in Digital Music Services John Fuller & Lauren Hubener, MLIS Special thanks to sponsor Jin Ha Lee & Rachel Price Review prior research Identify and summarize trends METHODOLOGY ADDICT ACTIVE CURATOR GUIDED LISTENER MUSIC EPICUREAN MUSIC RECLUSE NON-BELIEVER WANDERER Collect data through survey Analyze and categorize data Actively seeks new music Enjoys curating playlists Known-item searcher Prefers easily accessible music Prefers to hand over control to the streaming service Spends significant time learning about and discovering music Believes listening to music is a personal experience Skeptical of music service recommendations Enjoys serendipitous discovery of music and many genres 22% 7% 23% 3% 9% 25% 11% RESEARCH QUESTIONS 3. What are the implications for system design for each of the different user personas? 2. What are the expressed needs and behaviors of each user persona? 1. What kinds of user personas can we identify from real users of commercial music services? User Preferences Across Platforms SUMMARY AND NEXT STEPS 1. Approximately 35% of respondents fall into the user personas derived from epirical user interview data and observation sessions defined in previous research. 2. Further work includes understanding users who exhibit clusters of characteristics representing multiple personas, as well as the development of additional methods and measures of persona identification in order to verify the accruacy of personas. A Selection of Users’ Primary Streaming Service Preference Broken Down by Persona Source: RIAA 3 billion (US$) 2 1 Music Industry Revenues by Sector: 2013 vs. 2014 2014 2013 9.5% Streaming services Physical formats 22.4% - 7.1% Digital downloads Preferred level of customization Utilization of personal info Preferred level of interaction Desire for social features Demand for cross- platform compatibility Willingness to pay Low Neutral High 12345 12345 12345 Characteristics of Personas Across Platforms: Music Recluse vs. Music Epicurean 1. Likeliness to recommend an artist, song, or genre to a friend 2. Willingness to share personal information or listening history with a music streaming service 3. Regularity of making playlists 4. Level of organization and management of music collection 5. Likelihood to rate songs when using a music streaming service 12345 12345 12345 low medium high 15% 30% 17% 10% 13% 15% 31% 4% 9% 26% 26% 4% 33% 19% 26% 6% 10% 3% 3%
![[ADMA 2017] Identification of Grey Sheep Users By Histogram Intersection In Recommender Systems](https://static.documents.pub/doc/80x56/5a65e4417f8b9a6f448b48e9/adma-2017-identification-of-grey-sheep-users-by-histogram-intersection-in.jpg)
