Azure DevOps, previously known as Visual Studio Teams Services (VSTS), is a collaborative environment that supports Agile software development tools for planning, testing and tracking your work in the cloud. Enterprise-wide search is at its core a search engine like Google but searching in a Software as a Service(SaaS). This gives many use cases that includes documents search and access, retrieval and reporting of enterprise data which relies heavily on its search engine.

I was UX designer for the Search & Socials team of Azure DevOps, Microsoft's cloud offering for planning and tracking software development work. I was working on Search experiences of the platform. I was able to make a big impact by improving the experience's System Usability Score from 66 to 81 percentile which translates to: most user loved the new search experience and will recommended it to friends.

We conducted interviews with internal users to gather their feedback on the existing search experience. The major concerns were on the volume of the results and filtering. After initial user feedback, I tore down the experience into chunks of information which user goes through when he/she is looking for something, there were three main panes in a UI:

Pane 1 Pane 2 Pane 3

Typeahead Dropdown on Azure DevOps for code search which our users loved.

Navigation + Filter: Combining two important aspect of search created cognitive overload

Results pane was cramped up leading to hidden and confusing information

The gaps in above mentioned areas has led to the complex interactions and poor navigation. The hierarchy of the information flow was mismanaged by providing navigation and filters on the same pane. The image below shows the 13000ft view of the search experience, it clearly made the results page into a complex and cramped up UI, where sieving to the required information was difficult.

Both navigation and filter together in a single pane, presented the user with lot of options and confusing hierarchy. Jumping filters was a major concern while changing the search context.

In the Result pane, user also struggled with redundant and indiscernible information. We also tested the current system for its Usability Score, I performed scripting and use case documentation while moderating some of the test along with a User Researcher. The 66 percentile found the system below average.

To simplify the UI we focused on the information flow and went with simplistic approach for searching information. While maintaining the information hierarchy, my focus was keeping the end users always aware of the context.

Initially, we looked at both the above approaches and built low-fidelity mockups to discuss about the pros and cons. First one is common in e-commerce websites and is widely used by the all user, while the second one presents more hierarchical way. After conducting A/B testings and looking patterns in searching for our persona, we were able to narrow down to second approach for overhauling the search experience This method also gave an extra edge by providing search as a query builder, a more popular method to search in advanced users.

The SUS score for overhauled experience went from 66 percentile to 81 percentile which means, the end user will most likely recommend the system to their friends.