Confluence Overhaul

I overhauled a Fortune 25 healthcare company's design team Confluence — rebuilding the information architecture from scratch to serve both daily users and new hires learning about the team

company

Fortune 25 Healthcare
Company

ROLE

UX Researcher,
UX Architect

team

Design Operations team

timeline

May - Aug 2024
12 weeks

🌿In a nutshell…

Problem:
A Fortune 25 healthcare company's Confluence directory structure for the Design Org is difficult to navigate, for finding resources and past project information

Impact:
I reduced a Fortune 25 healthcare company's design team Confluence by 65%, improving navigation confidence from 2.38 to 3.13/5

Problem description

Picture yourself as an employee of the company's Digital Design organization. You are told to find previous research for a new redesign on a healthcare application in Confluence. However, the cumbersome organization makes it difficult to find the documentation and you feel frustrated. You end up contacting other researchers to finally find the documentation deep in layers of folders.

Now picture yourself as an employee of the company not part of the Design organization, but you're curious on what they do. You are provided the link to the Confluence space by the Design org, however the overwhelming amount of directories and non-descriptive naming of the pages makes it difficult to quickly know what content the page holds.

This is the problem - while Confluence has served the organization for the team to organize projects, initiatives and resources, the lack of organization, clarity and non-alignment with employees' needs makes it difficult for someone to navigate the space, find specific resources or even understand what the Design organization is about.

Users struggled to navigate pages quickly, locate relevant documents, and maintain clean structure when adding new pages. Overall, the Confluence directory structure did not meet their needs.

How can we redesign Confluence's information architecture to ensure intuitive organization, clarity and better alignment to employees' needs to reduce confusion with better user experience and accessibility with the space?

My project for this internship for 12 weeks was overhauling of Confluence to create a better information architecture for the Design org's employees. This process employed UX research methodologies and facilitating conversations to ensure the best organization of our Confluence. While I cannot go over specific details/content due to confidential data, I can provide a high level overview of this project.

Project Goals

What I wanted to accomplish with the project for the Design team:

Clarity

Clarity

The directory structure should make it immediately obvious what the Design team does and where to find it — for both daily users and people outside the org visiting for the first time

Usability

Usability

Navigation should feel intuitive enough that employees can find what they need without asking a colleague or guessing which folder it lives in

Alignment

Alignment

The structure should reflect how the team actually works — organized around the disciplines, projects, and resources that matter to the people using it every day

Research and Design Process

Initial Insights

Initial Insights

  • Conducted interviews to understand primary use case for our Confluence space + pain points

  • Satisfaction survey to gather quantitative data on different components of Confluence

Usability Testing

Usability Testing

  • Used UserZoom to test navigation of structured proposals based on initial insights

  • Did A/B testing to find best proposal that met user needs

Stakeholder Facilitations

Stakeholder Facilitations

  • Had a conversation with UX managers to figure out the best method to organize projects in Confluence space

Launching new directory

Launching new directory

  • Visual landing pages

  • Communication with design organization

  • Content migration

  • Post satisfaction survey

Part 1: Initial Insights

To understand how the team actually used Confluence and where it was breaking down, I interviewed 8 members of the Design team — spanning interns, UX researchers, UX designers, content/conversation designers, design systems, and UX managers. I also separately interviewed 2 interns and 1 new hire specifically about the onboarding experience, since their needs were distinct.

Interview — I structured the interview into three areas — current usage patterns, pain points, and how people actually organize and create content — because I needed to understand not just what was broken, but how people's mental models of the space differed from the existing structure.


For new hires specifically, I focused on clarity and team integration rather than navigation, since their primary need was orientation rather than finding existing content.


If you would like to see the interview questions, contact me.

Survey — I used a modified SUS Questionnaire rather than a custom satisfaction survey because it gave me a standardized baseline I could compare against post-launch results — making improvement measurable rather than just qualitative. 21 responses were collected via UserZoom.

The results painted a clear picture of a space that was functional but frustrating. Employees used it frequently (avg. 4/5) but found it poorly organized (2.33/5), hard to navigate (2.38/5), and unnecessarily complex (3.67/5). Confidence in using the space was middling at best (3.33/5).

If you would like to view the questions with its respective average ratings, contact me.

Insight Results

Based on the interviews and the initial survey, here is what I learned:

Organization is a challenge

87.5% of interviewees used the word "unorganized" to describe the space. Pages were inconsistently tagged, parent directories were non-uniform, and 50% of employees defaulted new pages to the bottom of the tree instead of filing them correctly

Lack of Visual Appeal

Confluence was built for developers, not designers — and the older version we used had limited visual capabilities. One particular visual I discovered during research became the inspiration for the landing page designs in Part 5

Existing folders not meeting user needs

Old, unused archives cluttered the directory and made it harder to find relevant content. UX Designers specifically noted there were no folders to house design briefs or resources, so content had nowhere logical to live

Lack of onboarding information

New hires felt lost — there was no clear starting point for learning about the Design team, no admin policies, and no explanation of how their role fit into the larger organization. Onboarding content was previously on Trello, but an expiring license made migrating it to Confluence a priority

3 Choices of Confluence Structure
3 Choices of Confluence Structure

Part 2: Tree Testing

I tested three structural proposals with employees across different roles to find which felt most intuitive to navigate. I chose the hybrid proposal because it balanced two needs that the other versions handled separately — organizing by discipline and by workload — without forcing users to choose one mental model over the other.

Testing ran in two moderated rounds. Round 1 used manual FigJam testing; midway through I switched to UserZoom after a colleague's suggestion, which significantly improved the reliability of the data.

Round 1 surfaced two main issues: confusion between "Resources for Digital Design" and "Tooling and Enablement," and a need for deep linking since some content logically lived in more than one place. I resolved both before Round 2

83% of Round 2 participants successfully navigated at least 8 out of 10 tasks — a meaningful improvement. The one persistent challenge that data alone couldn't resolve was the projects section, which led directly to Part 3.

Part 3:
Battle of Data vs. User Feedback —
Project Structure Organization

This was the most complex challenge of the project — and the most interesting.

The usability data said organizing projects by discipline was intuitive. But managers pushed back hard, wanting projects organized by year and PI quarters instead. The problem was that organizing by time alone had already failed in the old structure, and adding another time layer wouldn't fix that. Designers and researchers also told me they looked up past projects by medical plan or function — not by when they were done.

I first tried resolving this through individual conversations, hoping to find overlap. There wasn't any — every manager had a different preference.

So I facilitated a 30-minute group conversation, presenting two options where manager opinions had clustered most. My manager and I asked managers how their designers access information from previous projects, sparking a conversation amongst them. We wanted to see if any of their conversations pivoted towards one of these workflow organizations.

Part 3:
Battle of Data vs. User Feedback —
Project Structure Organization

But surprisingly, the conversation led to organizing the projects by pillars.
This was the most complex challenge of the project — and the most interesting.

The usability data said organizing projects by discipline was intuitive. But managers pushed back hard, wanting projects organized by year and PI quarters instead. The problem was that organizing by time alone had already failed in the old structure, and adding another time layer wouldn't fix that. Designers and researchers also told me they looked up past projects by medical plan or function — not by when they were done.

I first tried resolving this through individual conversations, hoping to find overlap. There wasn't any — every manager had a different preference.

So I facilitated a 30-minute group conversation, presenting two options where manager opinions had clustered most. My manager and I asked managers how their designers access information from previous projects, sparking a conversation amongst them. We wanted to see if any of their conversations pivoted towards one of these workflow organizations.

But surprisingly, the conversation led to organizing the projects by pillars.

The company supports various "pillars" where projects can be categorized under that were related to the services this company provides. Each manager was also in charge of 1-2 pillars, which means that designers and researchers can (1) easily categorize their work into this specific pillar and (2) there is general clarity on who is in charge of what pillar as well.


This was a good reminder that sometimes the right answer isn't in your data or your proposals — it emerges when you get the right people in the conversation.



The company supports various "pillars" where projects can be categorized under that were related to the services this company provides. Each manager was also in charge of 1-2 pillars, which means that designers and researchers can (1) easily categorize their work into this specific pillar and (2) there is general clarity on who is in charge of what pillar as well.


This was a good reminder that sometimes the right answer isn't in your data or your proposals — it emerges when you get the right people in the conversation.



Part 4: Data-Driven Solution —
Putting it all together

From preliminary interviews, creating proposal, doing A/B and usability testing and talking to multiple stakeholders, this was the final proposal created, all backed up with user research data.

After talking with accessibility experts on our team, we also decided to order the contents of each larger folder alphabetically (except for Welcome to Design Org)

Part 4: Data-Driven Solution —
Putting it all together

From preliminary interviews, creating proposal, doing A/B and usability testing and talking to multiple stakeholders, this was the final proposal created, all backed up with user research data.

After talking with accessibility experts on our team, we also decided to order the contents of each larger folder alphabetically (except for Welcome to Design Org)

Part 5: Visuals

The layout of this page is similar to all the other pages that introduce the specific domain in the Design Org's company. This visual helps others know about the particular domain's work, the team and what help they can provide
The layout of this page is similar to all the other pages that introduce the specific domain in the Design Org's company. This visual helps others know about the particular domain's work, the team and what help they can provide

Results

Three months after launch, I sent the same modified SUS questionnaire to the team to measure improvement. We received 8 responses — lower than the pre-overhaul 21, which I attribute to PI planning season at the company.
The results showed meaningful improvement across the board. Perceived complexity dropped from 3.67 to 2.13, organization improved from 2.33 to 3.75, and confidence in using the space rose from 3.33 to 3.88. Navigation ease — the core pain point we set out to fix — improved from 2.38 to 3.13.
Three months after launch, I sent the same modified SUS questionnaire to the team to measure improvement. We received 8 responses — lower than the pre-overhaul 21, which I attribute to PI planning season at the company.
The results showed meaningful improvement across the board. Perceived complexity dropped from 3.67 to 2.13, organization improved from 2.33 to 3.75, and confidence in using the space rose from 3.33 to 3.88. Navigation ease — the core pain point we set out to fix — improved from 2.38 to 3.13.
🍃

Reduced directory size by 65%

🍃

Improved accessibility to relevant resources + more intuitive organization for experienced and novice Confluence users

🍃

Post-overhaul survey showed measurable improvement in user satisfaction and navigation confidence

Results

🍃
🍃

Reduced directory size by 65%

🍃
🍃

Improved accessibility to relevant resources + more intuitive organization for experienced and novice Confluence users

🍃
🍃

Post-overhaul survey showed measurable improvement in user satisfaction and navigation confidence

Create a free website with Framer, the website builder loved by startups, designers and agencies.