Modernizing the Archives of Ontario
Rebuilding how people search for history.
Taking scattered and unclear information and making it into a search-first website built around how people actually look for records.
Role
Lead content designer, co-UX designer
Organization
Ontario Digital Service
Client
Archives of Ontario
The Archives of Ontario website is how citizens, genealogists, and legal professionals search and request historical records. This includes wills, estate information, land records, and photos.
The problem with the site was that:
it hadn’t been updated in years and became bloated with no clear information architecture and user flow
the homepage buried the important search function and search results didn’t show accurate or clear information
research guide content was scattered across webpages and inaccessible PDFs with no consistent pattern
Citizens couldn't find what they needed and staff were stuck manually walking people through a frustrating process.
A complete update was needed.
Challenge
Before: old home page with outdated look, 2 navigation menus, no clear call to action.
Approach
My product team and I applied a phased Discovery-to-Live approach.
As lead content designer and co-UX designer, I:
ran a content audit of all webpages, combing through the results with archives staff and subject matter experts
developed the content strategy for the site and for the redesigned research guides (structure, plain language, in-context help)
built a new information architecture using the audit and card-sorting/clustering exercises
ran tree testing on the proposed navigation and labels, iterating on the words and structure based on what users actually searched for
I contributed to:
in-person and virtual interviews with citizens, researchers, archivists, and staff
a comparative analysis of 8 other archive websites
rounds of prototyping using Figma and usability testing
1,380
existing pages evaluated for content audit
281
people completed unmoderated tree and first-click testing
4
weeks of prototyping sprints to ideate concepts using Figma
Focused on:
clarity of content and instructions
search and navigation usability
accessibility and discoverability of support resources
Key considerations included:
Do users have everything needed to accomplish their task at each stage of the user journey?
Is the user interface and are the search results self-explanatory and can users anticipate next steps?
Do the features and layout of the platform meet digital accessibility standards?
User research goals
User research findings
I identified recurring themes across interview transcripts and the comparative analysis of 8 other archive websites (at times using AI to help synthesize and group).
I found that the:
help text and calls to action were not prominent enough
pages were difficult to scan and understand
search experience did not align with users’ expectations or mental models
content was not as informative as it should be and contained too much internal jargon
Genealogist persona developed from user research and interviews.
Historian/researcher persona developed from user research and interviews.
Before jumping into layout or navigation, I audited the site's pages to understand what content existed and if it was doing its job.
After using AI-assisted analysis to score reading level and flag jargon-heavy pages, I reviewed the pages myself. I then collaborated with archives staff and subject matter experts to cross check and decide what to keep, update, consolidate, or delete.
I then assigned a “job to be done” so I could better group them by user need, which uncovered:
duplication (the same content repeated across multiple pages)
orphaned content (that no longer matched what users were looking for)
Content audit
Content audit spreadsheet with columns of data, from pageviews to reading difficulty.
Assigning “jobs to be done” to the pages’ subjects.
Information architecture: tree testing
281 people completed unmoderated tree and first-click testing using Optimal Workshop (recruited over social media and direct outreach).
The goal was to see whether users could find information in the proposed architecture and navigation menu.
The types of task-based questions asked included:
You want to find a digitized birth certificate of someone in your family. How would you find their birth certificate?
You'd like to add a historical image to a lesson plan. Where would you look for these images?
A sticky note exercise to start clustering groups of pages to begin building the information architecture.
People search by topic, not by formats.
Fix: structured the navigation around what someone is trying to find and not how the collection is organized behind the scenes.
Overlap between sections.
Fix: renamed certain menu items to be more precise, and moved around others. For example, removing “Learn” (folded lesson plans and stories into “Topics”) and simplified “Research”.
Even if it means changing the original structure, it’s better to move things to where a user expects to find them.
From a stakeholder engagement perspective, this data was important to make the case for certain parts of the redesign. These projects are often big shifts away from the current model, so it’s crucial to:
show unbiased information
centre decisions around data
take out personal preferences
Tree testing: results and takeaways
Proposed architecture and navigation menu used for tree testing.
Result of a high-scoring task (success of 88%).
Taking user research and interview data, I applied:
plain language throughout, replacing complicated archivist language on entry pages, while still maintaining meaning and nuance (increasing comprehension)
a big challenge was striking the balance between novice, first-time users and professional archivists, making sure unfamiliar terms are explained
content structures like subheadings and bullet points (improving scanability)
contextual hint text at the point when users need it (improving navigation and task completion rates)
I tested plain-language rewrites with AI assistance, then edited for accuracy and had staff review terminology to get it right. These were then further tweaked after more rounds of user testing.
Content strategy
Figma prototype of the home page that we used for testing.
A redesigned and improved Archives of Ontario website, launched with a:
streamlined and user-centred navigation
search-first homepage
new consistent and repeatable content model for pages
It is now:
built around what users actually came to do, with task-based navigation that doesn’t centre internal department structures
more plain language, with less internal jargon
average reading grade level across the site lowered from “college level” to “grade 8”
around 200 pages, down from around 1,380
taking away unnecessary, duplicate, or outdated information
Outcome
2 navigation bars, multiple call-to-action buttons, “AIMS” is confusing (it is the internal name of the search function).
Home page before
Modern, not overwhelming, with clear call-to-action search button (not surfacing advanced search options right away).
Home page after
Topic page before
Dense, harder to read, inaccessible, many links out to different places.
Clear hierarchy and numbered/stepped process, more plain language and important information up top.
Topic page after
At 1 year post-launch:
bounce rate on the homepage dropped by 9%
attributed to the search-first homepage redesign
inquiries that archives staff receive for help on finding records decreased by 15%
attributed to users being able to better navigate the site, with help from in-context tips, plain language information, and the step-by-step guide pattern
helping to save resources that can be better used in other ways
Impact
We tested iteratively so we could pivot when something didn't work.
For example, when early designs for search filter options underperformed in testing, we moved to a left-side filter bar instead of one above/horizontal. This pattern is what users preferred and expected. We then applied it to the other pages.
The revised information architecture and menu navigation was retested with the same participant pool to confirm that the changes made met user needs and closed the gaps.