TESTER SEARCH

Internal Web App | Applause Quality App | 06 -2018

UX, UI, IxD

The Problem

Applause’s internal platform is used by test engineers to search and invite testers to join testing cycles. Old tester search did not include all device types and other newer profile fields. The goal with redesigning search was to create a reusable component that improves accuracy and efficiency in a more robust search.

How Might We…provide test engineers with a search tool so they can perform complex queries across all tester profile fields?

THE CHALLENGES

  • Reduce usage of 3rd party tools to supplement search

  • Maintain consistency in UI for wide range of data types

  • Provide system feedback to represent complex Boolean logic

THE PROCESS

I followed the UX process and started with user interviews while I explored the problem. Next I sketched various search builder tools and created wireframes to gather feedback. With each search criteria, I clarified or simplified UI patterns as I began building a component library. Feasibility meetings with Engineering also helped shape the final solution as technical constraints limited the scope for MVP.

 

WIREFRAMEs

I created digital wireframes for the new search experience and presented to internal stakeholders. The basic concept is a drag & drop interface where users can build their search query in a visual way. Early user feedback helped refine the interface, with additions like Boolean connectors and dotted drop areas.

HIGH FIDELITY

After stakeholder buy-in, I designed a high fidelity prototype that brought the interactions to life. I conducted usability testing to understand whether the UI was intuitive to the end user and if they could conduct more complex use cases.

Reviewing with Engineering led to more fine tuning, breaking out the multi-level cascader select for Devices into a separate control for Operating System.

 

THE SOLUTION

Built on the backend with Elastic Search, the new tester search would allow for all new tester profile data to be searchable. Each profile field translates to a search criteria, available on the right rail. Offering both autosuggest and a dropdown in a single control saves time while giving the user freedom. Similar criteria sharing the same UI patterns were added to a comprehensive library. Additionally, separating the search results from the search builder helped to reduce loading time dramatically.

LEARNINGS AND success metrics

  • Progressive disclosure helps reduce cognitive load when the query contains criteria with if/then relationship

  • Drag and drop allows for user control and connectors offer visibility of complex searches

  • Covering 80% of all use cases was a good approach when evaluating MVP scope

Feedback from test engineers were very positive, noting the adoption and speed of the new search. Total search times decreased from 5 minutes to 30 seconds, with a total of 8,500 searches conducted in the first month (vs. historic of 5,024). This signifies the reduction of list imports and use of 3rd party tools.

“The updated search functionality looks and work great. I no longer need to use Domo to narrow down searches. The additional filter is truly helpful.”

“This really is spot on, I just ran a search that we would normally need to do to fill slots and it was no time at all. All the annoyances of wait then selecting are gone.”