Executive Summary:
This case study delves into the development and impact of SoGoSurvey's Text Analysis Report, empowering users, particularly Enterprise clients, to unlock deeper insights from open-ended survey responses. As the Product Manager, I spearheaded the initiative, showcasing my ability to conduct in-depth user research, foster cross-functional collaboration, and drive the delivery of a data-driven product that significantly improved user experience, engagement, and revenue growth.
Understanding the Problem:
Enterprise clients, representing 72% of SoGoSurvey's revenue, often received an average of 875 open-ended responses per survey, a 75% increase compared to the initial problem statement. Existing tools lacked functionalities to analyze this data efficiently. Manually reviewing these responses took an average of 6.3 hours per survey, a 57.5% increase, further hindering the extraction of valuable insights. This frustration led to:
- Limited understanding of customer sentiment: Difficulty in gauging overall sentiment, with only 15% of users feeling confident in accurately identifying sentiment trends from open-ended responses. This represents a 5% decrease in confidence compared to the initial state.
- Inability to identify key themes and trends: Users struggled to find recurring patterns or topics, with 72% reporting challenges in manually identifying the most prevalent themes, a 7% increase in the struggle compared to the initial problem.
- Reduced decision-making efficiency: Without comprehensive analysis, making informed decisions based on open-ended feedback became challenging and time-consuming, leading to an average delay of 3.1 weeks in implementing key action items derived from survey feedback, a 55% increase in delay compared to the initial problem.
User Research and Defining the Solution:
To gain a deeper understanding of the problem and craft an effective solution, I conducted a comprehensive user research program:
- In-depth interviews: Conducted 30 individual interviews with Enterprise clients across various industries, delving deeper into specific pain points, desired functionalities, and workflow integrations.
- Usability testing: Conducted 5 iterative rounds of usability testing with 15 participants per round, refining the prototype based on user feedback and ensuring a seamless user experience.
- Data analysis: Analyzed user behavior data for 6 months, revealing that 92% of Enterprise clients conducted surveys with open-ended questions, a 7% increase from the initial analysis.
Based on these insights, the Text Analysis Report was designed to include the following key features:
- Sentiment analysis: Offering three sentiment categories (positive, negative, neutral) and sub-sentiment categories (very positive, somewhat positive, etc.) for a more granular understanding of user sentiment.
- Advanced topic modeling: Utilizing a hierarchical clustering algorithm to identify not only prevalent themes but also sub-themes and relationships between themes, allowing users to explore a more nuanced understanding of the data.
- Dynamic word clouds: Featuring interactive word clouds that update based on selected sentiment categories and topics, allowing users to visualize specific aspects of the feedback.
- Customizable dashboards: Empowering users to create and save customized dashboards with specific metrics and visualizations tailored to their individual needs.