Creating impact while mitigating risk: The strategic value of user research, Adie Margineanu
Summary
Adie Margineanu, User Experience Lead at the University of Toronto Scarborough (UTSC), presents a detailed case study of how a disciplined, multi-phase user research program drove the redesign of the university's admissions website over a 10-month period. The project involved five distinct research studies engaging 194 prospective student participants in total, with Margineanu as the sole dedicated researcher — representing less than 9% of the delivery team.
The talk is structured around three recurring objections from leadership to user research — lack of time, money, and resources — and systematically dismantles each. Research consumed approximately 20% of the project timeline and less than 10% of total project costs (recruitment only, since Margineanu was an internal resource), and it ran concurrently with design and development without causing delays.
The research program followed the product lifecycle with four key decision points. First, a tree test with approximately 20 prospective students validated the proposed information architecture, revealing that users preferred a journey-based global navigation (programs → applying → finances → campus) over one organized by institutional departments. Second, two rounds of prototype usability testing — one low-fidelity and one medium-fidelity — with roughly 10 participants each surfaced critical usability failures in the program finder. Users were confused by advanced filters placed above the fold and assumed they had to fill them out before seeing programs. Findings led to simplified filters, progressive disclosure of prerequisites, and a surfaced co-op checkbox. International students specifically did not understand co-op, prompting a tooltip addition. Third, sentiment testing with 60 participants (split across two visual design options) used a validated word bank developed with the brand team to determine which concept was more on-brand. Though stakeholders preferred Option A, data clearly showed Option B was far more brand-adherent, and leadership accepted the data without pushback. Fourth, post-launch usability testing with 12 participants (~$1,500 recruitment cost) validated that the program finder worked at scale with real programs.
Outcomes were measurable: Google average position improved from 10.7 to 6.5; sessions increased 10% year-over-year despite sector-wide declines attributed to AI; session duration rose 3.8% site-wide and 19% on program pages; program page sessions increased 18% compared to the same period the prior year; and conversion (clicks on 'Apply Now') increased 62% in the first eight weeks post-launch during peak admissions season.
Margineanu also surfaces lessons learned: wireframes should use real or directional content (not Lorem Ipsum); vendor discovery phases need tighter integration; institutional content governance gaps limited the specificity of eligibility and fee information; and development starting before research was complete created feature constraints. She identifies the initiative as a minimum viable research program — card sorting was skipped, visual design testing was less robust than ideal, and proxy users were recruited under time pressure. The session closes with confirmation that the pilot established a repeatable research model with executive buy-in and dedicated budget going forward.
Key Takeaways
- •User research can be executed by a single internal lead and still cover the full product lifecycle — it does not require a large team or external research vendor.
- •Budget objections can be countered concretely: research cost less than 10% of total project spend (recruitment only) and consumed about 20% of timeline without halting other workstreams.
- •Tree testing should precede information architecture decisions to prevent organizing site structure around departmental divides rather than actual user journeys.
- •Wireframes and prototypes should always use real or directional content, not Lorem Ipsum — dummy text limits user engagement and hides design constraints during testing.
- •Advanced filters placed above the fold before visible results create confusion; users assume they must complete them before content appears — test and simplify filter UX early.
- •Sentiment testing with a brand-validated word bank provides objective, defensible data that can override stakeholder visual design preferences without generating conflict.
- •Post-launch usability testing is essential because live environments with real program inventories expose failures invisible in prototypes — budget recruitment costs (~$1,500) for this phase.
- •Progressive disclosure of filters (e.g., only showing high school prerequisites when an undergraduate program type is selected) reduces cognitive load and irrelevant information.
- •International student segments may not understand institution-specific terms like 'co-op' — identify and address these comprehension gaps during prototype testing.
- •Skipping upstream research methods like card sorting can lead to poor initial tree test results and iterative rework — plan the full research sequence from the start.
Transcript
Um, so I'm so I'm very excited to uh for this talk. I know a little bit about it, so I'm excited for everyone else to learn more about this project. So uh but it's about the strategic value of user research and um Addie uh she's been working in user exper experience research design and strategy for the past 14 years and has been leading user experience at user experience at uft scarboard for the last three and this talk about making user research feasible sound investment for teams to adapt to k...
