Mini Deadline 8

USABILITY TESTING: CARD SORTING, TREE TESTING & A/B TESTING 🌳

Summaries of the Methods

1. Card Sorting

What it is: A user-centred research method used during the Information Architecture (IA) phase to discover how people naturally organise, categorise, and label information.

How it works: Participants are given a set of items (represented physically on index cards or digitally) and asked to sort them into groups that make sense to them.

The Types:

• Open: Participants create and name their own categories (great for early-stage design).

• Closed: Participants sort items into predefined categories (great for testing an existing setup).

• Hybrid: Participants use existing categories but can create new ones if needed.

Why it's used: It helps designers understand users' mental models, the way people naturally organise and think about information. This ensures website navigation, menus, and content structures match user expectations rather than the designer's assumptions.

2. Tree Testing

What it is: Often called "reverse card sorting," this method evaluates how effectively a site's text-only menu structure supports information retrieval.

How it works: All visual design elements are completely stripped away. Users are presented with a simplified, text-only hierarchy (the "tree") and asked to locate specific items.

Why it's used: It isolates navigation flaws, showing if content is buried too deeply or placed in unexpected categories before visual design begins.

3. A/B Testing

What it is: A quantitative, data-driven method used to compare two different versions of a live digital interface.

How it works: Users are randomly assigned to either Version A (the Control), which uses the existing layout, or Version B (the Variant), which includes a specific modification.

Why it's used: It tracks hard metrics, like click-through rates, task completion, or time spent, to replace subjective design assumptions with empirical evidence.

THOUGHTS/OPINIONS ON THE METHODS:

On Card Sorting & Tree Testing 

The Value: I think these two methods are fantastic because they focus on the foundational layout before you get distracted by colours or fonts. So many massive usability issues aren't actually visual failures at all, they occur simply because labels are a bit confusing or because information is buried where a user wouldn't naturally look.

The Application: Card sorting is great early on because it strips away all the UI noise and forces you to align your design vocabulary with a user's instincts. Tree testing is a really solid reality-check follow-up. It serves as an intentional stress test for your menus, letting you catch navigation flaws and fix a broken structure before you waste weeks designing or coding a layout that doesn't quite hit the mark.

On A/B Testing 

The Value: I think A/B testing is fantastic when you need empirical proof to get stakeholders on your side. If you're trying to see whether a more prominent button or a cleaner layout helps users complete a task faster, a split test cuts right through personal opinions and delivers reliable numbers.

The Critique: At the same time, it's easy for it to strip away the actual human context. A data dashboard tells you exactly what percentage of people clicked a button, but it doesn't show the emotional or physical frustration they might have felt while doing it. A lost user might click a poorly placed link simply because they have zero other options, and the system flags that frustration as a "successful conversion." Quantitative metrics are incredibly useful, but pairing them with qualitative insights keeps us from reducing a user's real-world struggle to a cold data point.