Redesigning Enterprise Search at Mayo Clinic (Part 1)
In the summer of 2022, I was brought into a project that was long overdue. I had been working as a UX practitioner at Mayo Clinic for the Intranet Team for about 8 years and one of the biggest complaints from our Intranet (enterprise) users was that “search sucks”. I wrote a very detailed research report/case study internally, but I wanted to distill what the team (and in particular what I personally contributed) into a more practical and public version here.
We obviously cannot “fix search” in just one project phase, and so we approached it like a MVP (minimal viable product) for the first phase. The team used an evidence-based approach to our design decisions. I am excited to share our work with you.
My Role(s)
I was involved heavily in both the research and design. For research, I worked closely with another researcher to facilitate (Micrsoft Teams) and conduct the research (Qualtrics), analyze the results (in Excel), and share the findings to the design and development team. I also worked closely with another designer on the wireframes and prototypes (Figma), and made design decisions based on the research conducted. Finally, I worked with another developer to share our design decisions and handoffs.
What We Knew
- We needed to upgrade to the current version of the vendor’s search engine software.
- We cannot integrate all enterprise databases and platforms into our search engine results. For launch, we limited the result set to only WordPress sites.
- Based on Web Analytics, the top searches were related to: finding people, Human Resources, help desk, research, education, and parking.
- Secondary Research — Nielsen Norman Group published an article titled “Intranet Search Essentials” which served as a best practices guide for designing enterprise search.
What We Didn’t Know
- How to quantitatively measure if our new design is actually an improvement?
- What users liked and disliked about the current search box and results page?
- What users liked and disliked about the new search box and results page?
- How would mobile and tablet (screen size constraints) affect our new design?
- How to prioritize the different UI elements on the page layout? E.g. People results, Intranet results, Filters, Search For ‘term’ On external sites.
How We Found Out
- The UX Research team conducted a baseline UMUX survey of current search box and search engine results page (SERP) design.
- The UI team created two different design concepts for the Search Box.
- The UX Research team conducted a UMUX survey and preference test questions to validate the two new Search Box designs.
- The UI Design Team created two different design concepts for the Search Engine Results Page.
- The UX Research team conducted a UMUX survey and preference test questions to validate the two new Search Engine Results Page designs.
- The UX Research team conducted follow-up 1–1 user interviews to clarify user preferences of the new Search Engine Results Page designs.
UMUX
The UMUX (Usability Metric for User Experience) is a simple 4-question assessment based on the System Usability Scale (SUS). We chose the UMUX as a standardized scale to measure a user’s perception of usability and usefulness. We asked this survey in all our user testing sessions in order to quantitatively measure if our new design is actually an improvement.
Baseline UMUX
In an intercept survey, we showed (N=852) users the current search experience.
The score of 52.10 tells us that the existing Search experience was not very good.
The New Search Box
Now that we had a baseline to work with, the design team created four different design variations for the search box. We wanted to test whether users wanted search box always visible or if they were okay with a more clean look (hidden until they click the search icon). Also, we wanted to see if users wanted to keep existing functionality of a categorized search or if they were okay with a more minimalist design (minus the search categories).
We showed (N=162) users the new search box designs in random order.
Design 1a
Design 1b
Design 2a
Design 2b
The winner
After showing each design option, we asked each participant the four UMUX questions. The results of the UMUX study showed that users preferred Design 1a (Search Box with categorized search dropdown options always visible) over the other three options.
Redesigning Enterprise Search at Mayo Clinic Part 2
There’s a lot of content to cover, so I’ve broken it down to multiple parts. In the continuation of this article (Part 2), I will go over the design decisions that the design team came up with for the new Search Results Page.
Follow up articles and keeping in touch
Thanks for taking the time to read this. If you want to stay in the loop with what I am writing, please follow me on this platform, as well as on LinkedIn.
