UT Libraries is the nation’s fifth largest academic library system. The organization’s web presence represents its 11 physical locations, services, and extensive collections. The site serves as a landing platform for users to search thousands of academic databases, journals and e-resources.
The website first launched in the beginning years of the internet in the early 1990s. Though the site has changed shape over the course of 25 years, legacy content was never seriously addressed in previous refresh efforts.
The Web Refresh project started in the beginning of 2017 to address legacy content, ownership and a necessary platform upgrade. The project’s goal is to redesign the website in a responsive, UT-branded theme with updated information organization and visual design.
For this project, I work as a developer on a team using agile methodology in my role within the Libraries IT department.
Our team uses an agile workflow for this project. Over the course of a 2 week sprint, we participate in the typical ceremonies such as daily scrum, sprint planning, backlog grooming, stakeholder demo, and a sprint retrospective.
Throughout the project, we have adapted agile to our team’s needs and introduced other recurring meetings into our work cycle. For example, we have added in an internal demo to loosely plan out the presentation we provide in our stakeholder demo. We have also added in bi-weekly code review sessions to ensure smooth merges/transitions between our coding environments.
I work as a developer on this team and primarily pull stories that focus on user experience, Drupal site building, front-end web design, and content strategy. In the absence of our team’s product owner, I have also played the role of PO proxy where I’ve lead sprint planning, backlog grooming and stakeholder meetings.
Research has played an important role throughout this project from looking at analytics to speaking with our site users.
Google analytics reporting
I have used data from Google Analytics throughout this project to answer questions about our audience and site usage. Some of our questions focused on finding out the most used pages & areas of the site, how users access our site, and for organic search, knowing what words people use to find the site.
Early on in the project, reports from Google Analytics helped us to prioritize certain areas of the site first, provide stats justifying a new responsive website, and develop a list of features for our backlog.
With the launch of the new website, I have focused on setting up Google Analytics click events thru Google Tag Manager to track essential metrics such as navigation, chat button, donate button, and search type usage. The analytics reports generated from these events have been essential in evaluating progress and how the new design serves users.
We examined the competitive websites and online features of several tier 1 research institutions. All of these universities have recently launched redesigned websites and provided us with an overview of trends in the field.
We conducted various focus group sessions at different stages in this project. Early on in the project, I planned and facilitated a session with the Ask a Librarian staff who field the research desk and chat service for the Libraries. They are the frontline staff who interact with students and faculty regularly and receive questions about how to navigate both our physical and online services. In the focus group session, we gathered information on common user behaviors and the chat staff’s experience navigating users through the site.
In the phase of the project when we began to focus on the development of a new search feature, we planned 2 sessions. One focused on our undergraduate student audience, the second focused on PhD level researchers and subject specialists. We presented the participants with the concept of a unified search results feature and asked for their feedback. We also guided the participants thru tasks where they individually used competitor search interfaces. Following the completion of the individual tasks, we reconvened as a group to discuss feedback on competitive features.
Interviews with Content Editors
As part of developing a content strategy for the new site, we interviewed content editors and contributors on our existing site. In these sessions, we asked about their experience, likes and pain points with our current process. We also captured any wishes they have for the content editing experience on the new site. Following the sessions, we created an affinity diagram to help us identify future stories for our content strategy epic.
Information Architecture validation through Optimal treejack testing
We launched an A/B Optimal treejack test in the spring semester to seek insights on how to structure our new site’s navigation. We placed a pop-up on our existing site that routed a participant to one of two tests. The A/B test featured the same 7 tasks, but slightly different navigation options. The test included a demographic survey to capture which audience we reached (e.g., undergraduates, graduates, faculty, staff, public).
We gathered 100 total responses between the two tests (50 each). We compared task success rates between the two tests in order to determine preferred language for our navigation labeling.
UX Cafe - Usability Testing
I have performed evaluative usability testing on several key features of the new site design with UT students, faculty and staff. These usability testing sessions are conducted within a bi-monthly research series called UX cafe. With the findings from these sessions, I synthesize the results to provide key insights, highlighted user feedback + stats, recommendations and user stories.
Read more about my work organizing this series in my portfolio piece.
Secondary User Research for Content Development
In order to drive consistent content development for the new website, I compiled secondary research reports for key content sections of the new website throughout the project. I gave these reports to the subject matter experts responsible for these areas of content to help provide rationale behind content design decisions.
The goal of this user research was to understand how our current site content was used and accessed, explore common user issues, and see how competitor examples treated similar content. I used these methods: analytics reporting, heuristic evaluation, competitive evaluation, chat log and email keyword analysis. Each report concluded with a set of recommendations and a draft architecture for the new content section.
A major challenge in this project was the bulk of content on our existing site and legacy platforms, which constituted over 6,000 pages. Going forward, we have tried to identify a responsible way to migrate and refresh existing content, while making the content production process easier for our site editors.
Generating content inventory of existing site
Through the use of a Drupal content audit module and the export feature in Google Analytics, we were able to generate a content inventory of our existing site. Since our site lives across various platforms, through a simple SQL join query, we were able to create one combined inventory for reference in our audit.
Conducting rolling content audit by prioritizing site sections based off our MVP goals
Performing a full content audit where you address your current content in a meaningful way on a site with 6000+ pages is not very agile in nature. Fitting all of this work within a 2 week sprint is not possible. We also didn’t have the benefit of having a complete content audit before the project was kicked off. However, we did feel it was important to consider our current content in order to intelligently plan our new site structure.
We decided that instead of conducting a full content audit, we would perform the audit on a rolling basis, divided into our 2 week sprints. The content we decided to audit each 2 weeks was based off of our MVP goals. For example, since location and library space info were decided as top priority, we would flag related pages in the inventory, created a filtered view, and then audit these pages. Content was assessed for its usability (using a set of criteria we developed), its relevant audience (using personas we developed), and the proposed action for the page (e.g., move as is, move to another container, do not move).
Establishing a content production workflow for the new site
We worked with our stakeholder group to create a content production workflow that would work for the needs of library staff who would be editing the site. This involved leveraging a Drupal module suite called Workbench to make the workflow part of our CMS and work for a variety of content types. See a presentation detailing the setup below.
Creating content templates to gather content from subject matter experts
For each new content section on the site, I created templates to gather content from subject matter experts using Google Drive docs. These templates were created based off of the recommendations and draft architecture generated from user research. For each major content section, I held a kick-off meeting to create alignment for the content design that would be focused on user needs, instructed SMEs on how to fill out the templates, and answered questions.
Drupal Site Building and Front-end Web Development
Through both content changes on our Drupal database and code changes through our git repository, I have completed several stories focused on Drupal site configuration and front end development.
These stories include:
- Using UT branded Drupal kit to create pages on the site
- Prototyping directly in Drupal for quick deployment and delivery
- Designing custom features outside of the UT Drupal Kit through interactive mockups, using tools such as Adobe Illustrator, Marvel, and Axure
- Implementing custom features through code changes
- Creating menu structure for site and global navigation
- Creating custom content types and views such as our explore library spaces and find equipment feature
- Setting up taxonomy terms for news, events tagging and locations
- Researching and installing existing Drupal module solutions
I co-presented a talk about how we used an agile process in our library website redesign project at the Designing for Digital Conference in Austin, TX in March 2018. This presentation demonstrates how we used agile methodology, worked within our given limitations, and engaged our stakeholders to approach this project.