Empowering Access
Redesigned the University of Washington's Disability Resources website, consolidating 70+ pages across three campuses into 27 and raising accessibility scores from 82 to 98.
Overview
Project Type
- Website Redesign
- Information Architecture
- Content Strategy
My Contribution
- Product Design (sole designer)
- Research & Facilitation
- Component Development
- WordPress Implementation
Tools
- Figma
- WordPress
- Ollama (local RAG pipeline)
- Bootstrap
- Dubbot
Timeline
- Jun 2024 – Ongoing
- 12 months active design
Collaborators
- DRS Tri-campus Management
- DRS Students
- DRS Instructors
TLDR;
DRS gets ~5,400 calls a month about information that's already on their website. The instinct is to add more content. The real problem is the opposite — 70 pages across three campus sites, organized around internal processes, not user tasks.
I restructured the entire information architecture into 27 unified pages with audience-specific pathways, designed a callout card system that satisfied both legal reviewers and overwhelmed students in the same move, and built every component myself within a locked-down WordPress theme. The key insight: structure is the first interface. Get the site map wrong and no amount of UI polish fixes findability.
My Impact
67%
Improvement in Tone, Findability and Terminology vs baseline.
+18pts
Improved scores in overall accessibility scores as measured on Dubbot, automated WCAG compliance.
61%
Page count reduction (70 → 27) while expanding content coverage across all three campuses.
What was Broken
DRS is the University of Washington’s department responsible for disability accommodations across its three campuses — Seattle, Bothell, and Tacoma — under the Americans with Disabilities Act (ADA) and Section 504.
~5400 calls and emails in one month.
450 Hours or 18.75 Days every Month
People were calling about information the website already had. It wasn't a content gap, it's a findability failure.
The Problem. Visualized.
Every dot represents a single interaction.In September 2024, I partnered with the DRS front desk and testing center to log every incoming call and email for 30 days. About 3,000 interactions were tracked and tagged by topic and intent.
The pattern was consistent: people were calling about information that already existed on the website. Testing center hours, documentation requirements, accommodation timelines. All of it was published. None of it was findable.
If we do quick napkin math, its 5400 interactions x ~5 mins per interaction, which is 27000 mins or 450 hours.
That means out of 22 working days 18.75 days are spent answering questions. A full-time employee works roughly 160 hours/month. At 330 hours/month, DRS is burning through 2+ full-time employees' worth of capacity on responding to calls and emails alone. The testing center accounts for 61% of the total load.
This wasn't a content gap. It was a findability failure.
That distinction changed the project's direction. If information is missing, you add content. If information is unfindable, you restructure. These are opposite solutions. Getting this wrong would've wasted months.
A disability services department split across three sites, 70 pages, and zero shared navigation.
When I joined, students with disabilities were navigating three separate websites, 70 pages total, to find information about services they're legally entitled to under the ADA and Section 504.
When this site fails, it's not like a shopper picking a competitor. These are students already experiencing barriers who can't access support they need. Every failed interaction has real consequences.
Seattle
Most comprehensive; the main resource even for students elsewhere.
Tacoma
Sparse content; some pages had just one paragraph.
Bothell
Key info available as downloadable PDFs.
Key Constraints
I had very specific constraints on this project that I had to work within.
Legal review on every word
Past litigation means content must align with ADA, Section 504, FERPA.
Mandatory WordPress theme
No control over typography, spacing, or interactive patterns.
No developer
I coded every component myself.
No analytics
Built my own measurement through focus groups and structured surveys.
Tri-campus Collaboration
Three leadership teams. Decisions as granular as campus name ordering carried political weight.
An instructor with a PhD, describing the existing DRS website. The site was built for DRS, not for the people using it.
That's an instructor, someone with a PhD, describing their experience with the existing DRS website.
The site was organized around DRS's internal processes, not user tasks. Instructors couldn't find their responsibilities. Students couldn't locate key information. The language assumed insider knowledge and presumed non-compliance.
Focus group feedback surfaced six problems, but they all traced back to one root cause: the site was built from DRS's perspective, not the user's.
Problems in a nutshell
I had very specific constraints on this project that I had to work within.
Unfindable information
Content existed but was scattered across 70 pages with no task-based navigation.
Unclear Language
Acronyms, jargon, and interchangeable terms with no definitions.
Opaque Processes
Users couldn't identify what to do, in what order, or by when.
Walls of Text
Dense numbered lists within a narrow WordPress column. Psychologically exhausting to scan.
Missing Guidance
No proactive info on issue prevention (e.g., accessible materials).
Combative Language
Language presumed non-compliance; instructors felt unsupported.
What Changed
The biggest decision on this project wasn't a component or a color. It was the site map.
The site's core problem was structural. So the most impactful decision wasn't a component or a color; it was the information architecture.
I tracked 71 individual content topics through a mapping spreadsheet: 12 direct mappings, 12 merges, 10 redistributions, 7 net-new pages, 6 restructured items, 5 expansions, and 3 removals. Every piece of content from the old sites was accounted for.
Information Architecture
Old
New
Coding Examples
Every word has to survive a lawyer. Every reader is stressed and navigating a complex process. I needed one solution that satisfied both.
This was the project's hardest tension. DRS has faced litigation. Every piece of content must align with ADA, Section 504, and FERPA definitions; language can't be "simplified" in ways that alter legal meaning.
But the students reading this content are in high-stress states, navigating complex processes while managing conditions that make dense text harder to process.
These two requirements are in direct conflict.
The resolution: a callout card system.
I designed four types of callout cards (Tip, Note, UW Policy, Legal Policy), each with a consistent color and label. The cards visually separate supplementary and legal information from the primary user-facing content.
Legal reviewers could verify that required language was intact. Users gained the clarity that research demanded. Neither side gave up anything.
This is a design solution and a political solution. It resolved the user experience problem and the conflict between the legal and design teams in one move.
Callout Cards
What Shipped
The WordPress template didn't change. How information reaches people changed entirely.
The WordPress theme, fonts, and colors remain unchanged. Instructors and students open the site and see the same UW shell they've always seen — same navigation, same footer, same brand. What changed sits underneath: the information architecture, the content hierarchy, and the component system they use to navigate it. Nobody has to learn a new visual language to get the benefit.
Changes, Results and Validation
Original
Post Redesign
“I don't know how to read this page.”
Summer 2024
“The biggest improvement over the original DRS website is that navigation is much improved, and I think that is directly a result of your work.”
Winter 2025
Supporting Evidence
from the Winter 2025 (Jan 2025) feedback survey
“…It's clear that a ton of work went into this, and you all should be really proud!…”
“…We addressed all the areas that could use improvement…”
“…Much much much improved from previous version…”
There was no developer on the project. I built the full component toolkit myself and made it reusable for staff after I leave.
Each piece followed the same pipeline: Figma design → focus group testing → AI-assisted code generation (Bootstrap + UW branding guidelines) → documentation as copy-paste WordPress shortcodes for DRS staff to maintain after I leave.
Custom Components
Terminology
67% of instructors rated “Clear” or “Very Clear”
(vs. baseline where dominant theme was confusion and undefined terms)
Findability
67% rated “Easy”
(vs. baseline where participants couldn’t locate timeline information)
Tone
67% rated “Collaborative”
(vs. baseline “combative”)
Accessibility
82 pts → 98 pts
Dubbot, automated WCAG compliance
Page Reduction
61%
(70 → 27 pages, expanded content coverage)
Content Topics
71 mapped — 12 direct, 12 merged, 10 redistributed, 7 new
Design Files
4+ major iterations
Research
35 participants across 8 sessions, 4 rounds
See my other projects
If you liked this project, my mom thinks you might like these too!
