AskPolly Dashboard
AskPolly is a self service tool for market research. A person enters a query of what they want to measure, and the population they want to measure on. Polly then returns what conversations people are having, how people feel about each of the conversations, and the demographics of the people.
The overarching challenge is how people interact with Polly. There are two main areas of interaction:
1, Input to Poly: Giving Polly input about what a person wants to learn about and;
2, Output from AskPoly: interpreting the output to answer a specific question.
1, Input to Poly: Giving Polly input about what a person wants to learn about and.
Based on our testing findings, we observed an issue with the location field on the search page. Despite providing clear instructions to specify their interest in hearing from Ottawa residents, participants tended to overlook the field and leave the default value unchanged. In order to address this challenge, we implemented a sorted search screen design that prompts participants to enter the desired population location. To streamline the process, the other input fields are now disabled until the location field is filled.
To further assist users, we introduced the Getting Started UI, which includes helpful tooltips that guide users through the steps of sorting their search input. These enhancements proved successful in subsequent testing sessions, as all participants consistently entered population locations and queries.
2. Output from AskPoly: interpreting the output to answer a specific question.
We made several significant updates based on user feedback.
Firstly, we introduced a dynamic three-state button to enhance the selection process. This new button clearly indicates whether all cards, a few cards, or no cards are selected. By clicking the button, users can quickly select or deselect all conversations, streamlining the interaction.
To improve clarity, we introduced a distinct visual contrast between selected and unselected cards, accompanied by a hover state that provides a preview of the effect clicking a card may have.
Additionally, we standardized all graphs as bar charts to ensure consistency and predictability for users when interpreting the data. This change aims to facilitate a seamless reading experience across different graphs.
Furthermore, we redesigned the card interface, opting for a tabbed layout. This modification resulted in improved usability, enabling participants to easily locate and access various demographic data available to users.