B2E DASHBOARD
Detect & handle bad open-end answers
Quick summary
Potloc is a market research company fighting against poor quality and fraud in the industry. As a product designer there, I focused on quality improvement initiatives.
Open-ended questions are crucial in surveys, requiring respondents to provide answers in plain language.Our technical and user research revealed we could automatically identify the quality of open-ended responses.
8 hours/week
Time saved per team members.
98%
Accuracy of the algorithm.
100%
User satisfaction
Open-end analysis
When you're doing a market research, open-ended questions are unique because they allow respondents to articulate their thoughts in their own words, offering valuable insights into their engagement and authenticity.
Unlike closed-ended questions, they have no pre-defined options, making them excellent indicators of genuine participation.
Overview
Objective
Our objective is to rigorously evaluate open-ended responses to ensure datasets contain only authentic, high-quality feedback. This will make our insights more reliable and truly representative of audience perspectives.
Key benefits
- Eliminate 8 hours/week of manual work for each project.
- Reduce by 3 the number of tools used.
- Speed up delivery to our final customers.
- Demonstrate our commitment to team members.
Discovery
To understand and design the exact tool our teams needed, I collaborated closely with a software engineer, our AI team, and our product manager. We organized several workshops with teammates in both France and Canada.
Throughout this process, our users were fully onboarded in the feature development and actively participated in breaking down the different versions. I also worked with our technical teams to understand our constraints as we prioritized functionality based on user validation.
This discovery phase resulted in:
- User Research,
- Technical feasibility (handled by the software engineer),
- AI provider assessment (handled with the AI Manager).
Solution
Using Figma and our Design System, we delivered a beta version to test our solution with real data in real conditions with actual users. Based on this beta version, I made several iterations to optimize the tool, ensuring it precisely met our users' needs and effectively displayed both macro and micro data about open-ended response quality.
Deliverables included:
- Detailed user flows,
- Screen flow maps,
- UI mockups,
- Interactive prototype,
- Precise documentation.
Results & Added Value
- Delivered a V0 that added value to our internal users' daily workflow.
- Iteratively improved based on beta tester feedback and Amplitude data.
- Aligned with company vision by empowering customers to clearly understand their data quality.
User Journey
I organized a workshop with the project support team to understand the different steps they go through during a project. This was the starting point for the feature—we identified many switches between tools (Potloc app, Forsta, Google Sheet, Slack), most of them due to open-ended issues.
User Research
To fully understand the open-ended analysis process, I conducted meetings with our users to collaboratively define our goals, identify both user and technical requirements, and establish boundaries for the project. I compiled all feedback in a shared Notion page, ensuring all stakeholders remained informed and aligned throughout the process.
- Objectives: Why users need to perform open-ended analysis
- Needs: What users require to achieve their objectives
- Risks: Potential threats to survey projects or technical implementation
- Opportunities: Elements we could leverage to streamline the process
Screen flow maps
To ensure our users and engineers were aligned with our solution, I created several screen flow maps detailing our framework:
- Show a macro view of open-end quality (including number of answers, minor/major open-end issues, and similar answer rate)
- Add a new open-end check
- Edit an existing open-end check
- Delete an existing open-end check
- Access open-end details
- Access similar answer details
UI Mockups
High-fidelity interfaces designed in Figma, optimized for quick access to information, easy data comparison, and helping our project support team make the right decisions.
The emphasis is on clarity, speed, and accessibility.
This project demonstrated how user-centered design can transform internal workflows. By collaborating closely with team members and technical partners, I delivered a solution that eliminated manual work, reduced tool switching, and improved data quality visibility.
The iterative approach, from discovery workshops to beta testing, ensured the final product met user needs while aligning with our company's commitment to quality and transparency in market research.
Let's work together
We'll begin around a call to understand your expectations and the problem we've to solve.
