Introduction

Quid is a data visualization tool that allows users to search through millions of documents in order to aggregate them in broader themes based on language similarity, providing a high level view of how the data is segmented. The first step in this process is to collect the most relevant set of documents through Search.

The goal of this project was to redesign the Search experience, in order to provide the users with the best set of tools to curate their data and end up with the most relevant documents that they wish to analyze.

Team

UX Design: Radhika Sawhney | Product: Paul Chong | Developemnt: Derek Gathright

Previous slideNext slide

Problem Statement

To improve the Quid Search experience by providing user tools to easily express their intent in order to capture the most relevent documents for a clean, noise-free analysis.

Project Goals

Problem Definition

Before creating a solution, we spent a considerable amout of time gatering data that would help us understand the user painpoints and frame the probem that we were trying to solve for. The steps we took to frame the problem were as follows:

User Pain Points

This is what the search experinece looked like before the redesign. The onus of creating a robust boolean query was entirely on the user with no guidance from the platform. The only feedback avaliable to the user was a list of articles with no curation capabilities.

Ideation

Understanding the Landscape

Creating Concepts

After understanding the user problems and the technology constraints, the next step was to create some initial sketches and the desired user flow.

Wireframes

Search Query Builder

One of the core problems identified was the fact the onus was on the user to create a robust boolean query in order to create a good search, without any guidance from the tool. I worked on the concept of a ‘Query Builder’ that would help the users frame their boolean query in a more conversational manner. The idea was to have Quid provide guided questions to understand the operators the user wanted to use (OR, NOT, AND).

Previous slideNext slide

Evolution of Topic Cards

The other goal for us was to provide a visual way to search. For this concept, I worked closely with the backend team to understand various ways in which the articles could be aggregated. We used the metadata such as the Topic the articles belong to, People mentions, Company mentions, Source publishers, Time etc. The concept was to provide the ability to curate these aggregations in order to remove noise. We experimented with different visualizations, such as treemap, timeline, bar charts, etc. to represent the volume.

Next we worked on defining the curative actions that a user could take to remove noise. The idea as to first have the user expand her search by adding recommended terms or synonyms (OR) and then refine by removing unwanted topics, people, companies, sources, etc (AND NOT topic).

Previous slideNext slide

Formative Usability Testing

In order to improve on the design, we conducted Formative Usability Testing by continuously iterating based on the feedback we recieved. We interviewed close to 20 users in cohorts of 4-5 people and went back to redesign the experience based on the feedback from each session. This helped rapidly iterate and improve on the initial solutions.

Setting Success Metrics

In order to measure success, we established a baseline by measuring the usability of the current design. This was done by rating the current expereince on a 7 point Likert scale, and calculating an average baseline.

We also conducted a SUS (system Usability Scale) analysis to get a benchmark metric of how usable the current search experience is, and set an aspired goal for improvement.

Iterative Testing

Through the usability tests, we measured the overall improvement from the baseline and made the following iterative changes:

Design Solution

The final solution consisted of the following additions:

Previous slideNext slide