Improving Search Usability for 100,000 users

SaaS
Search
Mobile Design

Project Overview

Digital Thread Engineering (DTi), an internal SaaS product for industrial operations, is a database of CAD drawings, part schematics, CIDs and other manufacturing data that connects stages of the product lifecycle such as production, operations, testing, and maintenance with all departments in GE. As lead designer on the search experience, I was tasked with identifying and designing improvements to the experience.

Design

UX & UI Design

Strategy

Evaluative Research, Usability Testing

Tools

Sketch, Invision, Microsoft Teams

Industry

Saas

Research & Analysis

Our team, including a senior PM, technical PM, and engineers, identified inefficiencies in DTi search through stakeholder interviews and user research. The root issue was the lack of granular search control, forcing users into broad, slow searches with irrelevant results and frequent timeouts. Users adopted workarounds like running multiple tabs or memorizing part numbers to save time.

Concepts & strategy

To address these issues, I proposed two enhancements. Data Type Search, enabling searches by specific data types for more targeted, faster results and autocomplete suggestions, adding real-time keyword and content suggestions to guide users toward relevant results and refine searches efficiently.

Hypotheses

Implementing these changes would lead to: Higher Search Success Rates: Users find information more efficiently. Reduced Search Time: Real-time suggestions and targeted searches minimize wait and refinement time. Improved Satisfaction: Faster, relevant results enhance the user experience.

View Live Prototype

Outcomes

Before development, I conducted a usability study with 5 engineers to quantify the new design’s impact on overall user satisfaction for general and specific search tasks.

The study consisted of two search tasks, one general, one specific, and a follow up questionnaire using a likert scale to understand initial perceptions, and satisfaction. Engineers were asked to speak out loud during each search. My goal was to capture initial evidence for the design direction.

Using a 'Wizard of Oz' approach to displaying real time search suggestions and autocomplete, feedback was overwhelmingly positive:

  • Real-time suggestions and autocomplete were met with exasperated enthusiasm as users immediately understood how much time they would save with more precise search
  • User satisfaction: Using satisfaction and ease of use as the anchor, participants rated their experience with the new design 'Strongly agree'. Users remarked that interface simplicity and improved usability made the process feel more intuitive and easy.

These findings validated the designs and allowed our team to feel confident in the impact on user productivity and sentiment.

After presenting my designs and findings to our team, the PMs decided to add the improvements to their product roadmap and commit to building a pilot search experience using the proposed enhancements.