Chapter 2:
Validating Top-Level Categories with a Paper-Based Tree Test
Introduction
Before building the clickable navigation prototype, I ran a low-fidelity tree test on paper to ensure my five top-level categories truly match shoppers’ mental models—and to catch any ambiguous labels.
1. Objective
Validate which top-level categories guide shoppers directly to the item they’re seeking on their first click.
2. Approach
Format
Paper-based tree test of my navigation hierarchy (no visuals, just plain text).
Top-level category labels presented:
Dresses
Tops & Bottoms
Footwear
Bags & Accessories
Hats & Headwear
Procedure
Participants mark their first choice for each task on the sheet of six labels.
I reveal only that category’s children (e.g. under Footwear: Sandals | Loafers | Espadrilles) and they mark their final pick.
I record whether their first-click matched the intended category.
Tasks
“Browse T-Shirts”
“Shop Raffia Sun Hats”
“Find Loafers”
Participants
| Persona | Behavior Pattern |
|---|---|
| Efficient Browser (A) | Clicks confidently on the correct category first time. |
| Distracted Shopper (B) | Picks a sibling category, then self-corrects. |
| Detail-Obsessed Explorer (C) | Peeks under the wrong category before backtracking to the right one. |
3. Initial Results
When I first ran the paper-based tree test, overall task completion was 100 %, but first-click accuracy revealed a hiccup in two of the three tasks:
| Task | First-Click Success |
|---|---|
| Browse T-Shirts | 3 / 3 (100 %) |
| Shop Raffia Sun Hats | 2 / 3 (67 %) |
| Find Loafers | 2 / 3 (67 %) |
The combined “Bags & Accessories” label proved too broad: by grouping carry-items and wearables together, it misled two out of three participants when they tried to find Sun Hats or Loafers.
All other categories—Dresses, Tops & Bottoms, Footwear, and Hats & Headwear—performed flawlessly on first click, demonstrating that their labels aligned well with shoppers’ mental models.
Splitting the one ambiguous label into separate “Bags” and “Accessories” categories promised to eliminate these misclicks and create a clearer, more intuitive navigation structure.
4. Label Update & Retest Results
I separated “Bags & Accessories” into two distinct categories—”Bags” and “Accessories”—and reran the three tasks.
Bags (crossbody, totes).
Accessories (scarves, jewelry, hats, etc.).
| Task | First-Click Success |
|---|---|
| Browse T-Shirts | 3 / 3 (100 %) |
| Shop Raffia Sun Hats | 3 / 3 (100 %) |
| Find Loafers | 3 / 3 (100 %) |
After making the split, every participant achieved 100 % first-click success on all three tasks—confirming that this one label change removed the only remaining friction in the navigation.
5. Key Takeaways
By isolating labels in a low-fi, paper-based tree test, I confirmed that my top-level categories map directly to shoppers’ mental models—achieving 100 % first-click success once “Bags & Accessories” was split into discrete “Bags” and “Accessories” labels. This validation not only removes a critical navigation friction point but also gives me confidence that my taxonomy-driven labels will scale as the catalog grows.
Next, I’ll leverage these validated category names in a clickable prototype, ensuring that interactions and visual affordances reinforce this clarity and lead users straight to the products they’re seeking.