Information Architecture | UX Research
Mindbridge is a Toronto-based enrichment school with strong pedagogy and a website that hid it. I restructured the content model, taxonomy, and navigation so prospective parents could evaluate programs and start enrollment without contacting the school first.
Outcomes
Three competing navigation systems reduced to one,
One catch-all FAQ decomposed into reusable content chunks.
A metadata schema for programs.
Context
Responsive Web
Methods
Problem framing
user research
wireframing
prototyping
user journey mapping
Timeline
FAll 2025
Role
Information Architect
Ux researcher
UX Desigenr
The problem
Parents were the site's primary user, and the site was built for everyone but them.
The homepage stacked three navigation systems at once: a utility bar, a main menu, and a vertical sticky column with no indication about which was primary.
The most prominent sticky button, "Join our Staff," targeted job seekers, but the client confirmed that recruitment was not a business priority, nor primary or secondary users of the website.
Three navigation bars and no primary path.
Fifteen competitor sites benchmarked
14 of 15 had a top-level About page; Mindbridge's equivalent content was split across two labels ("Staff" and "Meet Our Teachers").
Parent forum discourse
showed the recurring complaints were pricing opacity and hard-to-locate schedules, not missing programs.
The FAQ was a catch-all
holding enrollment steps, tuition policy, and general questions — parked on the About Us page.
Solution
One navigation, one hierarchy.
Collapsed three bars into a single global nav with five categories and one high-priority CTA, so every page has one predictable entry point.
Content chunks, not pages.
Broke the FAQ into tagged, reusable units — enrollment, programs, policies — and placed each where the question actually gets asked.
Metadata-rich program cards.
Surfaced grade, schedule, outcomes, and instructor on the card itself, so parents compare without opening five pages.
Instructor name on the card was a direct response to the client's insight that staff quality drives enrollment.
Core Flows
The flow that carried the value was evaluate → shortlist → enroll. Under the current site, a parent hit a decision point — is this program right for my child's grade? — and had no way to answer it without opening every program page and reading prose.
I moved the branch earlier. Filters resolve fit before the parent commits attention to a description. Admission criteria and the registration calendar sit inside the enrollment path rather than in News, where date-related content had previously been filed.
Dead-end states were the second problem: blog posts, news items, and event pages ended with no path upward. Breadcrumbs and a reclassification under News & Events gave every deep page a return route.
[VISUAL] Asset: Flow diagram, current vs. proposed, with screen thumbnails at nodes Shows: Where the grade/program-fit decision moves earlier in the path Format: 3:2, standard flow notation, the relocated decision point highlighted in accent Caption: Moving program-fit filtering ahead of the description page removed the read-everything step parents were doing by default. Alt text: Two-lane flow diagram comparing the original path, where filtering happened after reading, against the proposed path with filters first. Source: Needs to be created
Research
Research question: where does a prospective parent's evaluation break down, and is the cause missing content or misplaced content?
Method selection. Direct user recruitment wasn't available inside the project timeline, so I used three converging proxies rather than claiming interviews I didn't run: competitive benchmarking (n=15), stakeholder interview with the client, and discourse analysis of parent forums. Benchmarking was chosen over a heuristic walkthrough because the question was about conventions — what parents have been trained to expect elsewhere — which a heuristic audit can't surface.
I ran the content inventory and audit across every page, splitting content into chunks and tagging each with the metadata a parent would need. Synthesis mapped audit findings onto a current-state content model, then a proposed one.
Insights
Convention, not preference, sets expectations. 14/15 competitors used the same top-level categories; Mindbridge's variations read as errors, not personality.
Schedules are an unmet need across the category. Few competitors made calendars easy to find, despite parents naming feasibility as a first-pass filter — a competitive opening, not just a gap.
Instructor credibility drives choice. The client identified staff quality as the top enrollment driver, yet instructor profiles were unreachable from the nav.
Competitive Analysis
15 competitor sites benchmarked
Selected for direct overlap with Mindbridge's category—private enrichment and after-school programs. I measured pattern frequency, not aesthetics: how often each site used a given navigation label, where CTAs sat, and whether search existed at all.
Competitor sites overwhelmingly use shallow, predictable hierarchies, typically presenting global navigation categories such as About, Programs/Courses, Schedule/Calendar, Contact, and News/Events. This reflects established IA principles: shallow hierarchies reduce cognitive load and prevent disorientation
Analysing Global Navigation in 15 Competitors and Mindbridge’s Current Website.
Assumption killed: I entered assuming the site needed search. The audit showed the real failure was indexing — one grade filter — and that search over an unstructured catalog would have produced empty results and lower trust.
[VISUAL] Asset: Benchmarking matrix, 15 competitors × navigation elements, with frequency counts Shows: How convention was quantified rather than asserted Format: 3:2, real data, Mindbridge's row highlighted Caption: Counting label frequency across 15 sites turned "this feels unconventional" into a defensible 14-of-15 argument. Alt text: Comparison table of fifteen competitor sites showing which global navigation labels each uses, with tallies per label. Source: Exists in report (Figure 1) — needs re-typesetting for legibility
7. Design Strategy
Principle: modular and extensible over tidy. Content had to be separable from its container so a single chunk could appear in multiple places. This ruled out the obvious fix — reorganizing pages — because it would have re-created the same catch-all problem one level down.
Prioritization. I ranked changes by effect on conversion, the client's stated primary objective, against implementation cost. Navigation consolidation and FAQ decomposition ranked highest: high effect, near-zero build cost.
Two hard tradeoffs.
The first was accepting more pages. Nineteen pages instead of six runs against the instinct to reduce, but depth with a coherent hierarchy beats shallowness with orphans. I made this bet because the audit showed the problem was ambiguity, not volume.
The second was deferring search entirely — the most visible recommendation, and the one I cut — in favour of the metadata layer that makes it viable later.
Success measure defined before designing: number of distinct paths a parent must traverse to reach an enrollment decision.
[VISUAL] Asset: Prioritization matrix, effect-on-conversion vs. implementation cost, with the real recommendations plotted Shows: Why search was deferred and nav consolidation went first Format: 4:3, plotted items labeled, deferred items in a muted treatment Caption: Search scored high on visibility and low on value at this catalog size — the clearest cut of the project. Alt text: Two-axis matrix plotting eight IA recommendations by conversion impact and build cost, with search positioned in the low-value quadrant. Source: Needs to be created
Wireframes
Low fidelity throughout, deliberately. The argument was structural, and greyscale wireframes kept the client's review on hierarchy rather than color.
Each wireframe set tested one question: can a parent orient on the homepage in one screen; can they compare programs without opening pages; does the enrollment path hold together end to end.
[VISUAL] Asset: Annotated wireframe set — homepage, About, Programs Shows: IA decisions rendered as layout, with the numbered rationale visible Format: 3-up, consistent frames, one background Caption: Keeping wireframes greyscale kept client review on structure — every round of feedback was about placement, not palette. Alt text: Three annotated greyscale wireframes for homepage, About, and Programs pages with numbered design rationale callouts. Source: Exists in report (pp. 20–21)
Accessibility
Not done during the project, and I'm not going to pretend otherwise.
Two decisions had accessibility consequences I can defend. Removing the sticky button stack cleared content it had been covering, including the school's address. Specifying a labeled search field over an icon-only control follows the pattern that performed best for visibility and non-native-speaker access in the benchmark set.
Neither was verified. There was no contrast audit, no focus-order review, no screen reader pass, and no usability test.
The audit was the whole project. Every recommendation traced to a row in the content inventory. I now start structural work with the inventory rather than treating it as documentation.
I over-indexed on search early and spent time on a solution the catalog couldn't support. Recognizing that indexing was the real problem came from the audit, not from intuition.
The tradeoff I'd defend hardest is adding pages. It looks like the wrong direction until you see the orphan map.
What I'd change: I'd have pushed for five parent participants and a single findability task. Without it the work is well-argued and unproven, which is a weaker position than it needed to be.
If I had another month, I would run that test on the current site and the prototype, and publish the delta — even if it were flat.