METHOD
We took a mixed-method approach to study user behaviors and built user personas for Tiktok/Douyin content creators. We first sent out surveys (including screening questions) to quantitatively collect users’ high-level attitudes, opinions, and knowledge about Xingtu, as well as demographic data. We then conducted 23 semi-structured interviews and 23 card-sorting sessions in person. Then, we analyzed the data from the interviews, survey, and card-sorting results and created three content creator personas. After that, we closely worked with UX Designers to ideate potential solutions for the issues we found and ways to better assist content creators to grow up within the Xingtu platform. Finally, we wrote the research report and did a presentation for 150+ colleagues in our team.
Survey
Design
The survey is designed to screen potential participants, and quantitatively collect information about content creators’ high-level attitudes, opinions, and knowledge about the Xingtu platform.
(** Due to confidentiality concerns, I would not be able to show the actual survey here
Sample
We randomly sampled XXX users based on the following criteria:
1. Creator Type: self-served
2. Activity status: login >1 in the most recent 30 days
3. Number of followers: 10-500w
Interview
We used the survey response to further filtered out 50 interview candidates based on:
- Content creator category
- Number of followers
- Monthly income via Xingtu
- Location
In the end, we conducted 23 in-person interviews in Beijing, Shanghai, Guangzhou, Shenzhen, and Chengdu.
Card Sorting
Followed by the interview, we conducted 23 card-sorting with the same participants.
Analysis
After interview, I first wrote a report for each individual interview using notes and video recordings. After that, I used affinity diagram and qualitative coding to analyze interview data in order to create constructive answers to the questions we had before the research. Then, I wrote a research report to present my results and did a presentation to my colleages
** Due to privacy concerns, I'm not able to show the analysis process in detail