User Testing: When & How to Test

User testing is such a vital part of finding the best solution. We can research until our heart is content, but what can be more valuable than getting critiqued by the people you’re designing for? That being said, the question no longer is, “Do we test?” but rather, “How do we test?”

Illustration of websites on multiple devices.
Published: 1.15.2016

This post is the second in a series on user experience. We'll be covering each facet of user testing, with each discipline giving their take.

Where to Start, What to Test
The first step is figuring out the objectives for the test: What are we trying to learn? Which design will users prefer? Is our proposed sitemap intuitive? What about the path to conversion? If you have questions about all of the above, you’re going to need multiple tests. Don’t try to cover too many topics in one test—keep them short and to the point.

Deciding what you want to learn is the key to determining when and how to test, and you’re going to be most effective when that test is targeted.

Types of Tests and Best Use Cases
The world of user testing is so broad that it can seem daunting to try and tackle. What type of test is best for this objective? When in the process is most efficient? There really isn’t one correct answer, as each project is different. But, this is what I tend to see:

Card Sorting/Tree Testing
When to Test: Before Design
What You’re Testing: Information Architecture

Web design is kind of like building a house, with information architecture being the foundation—if your foundation is broken, it doesn’t matter how pretty the paint is. Testing a site’s architecture needs to happen early in the process, and needs to answer the question: Is your proposed site map intuitive to your target users?

A quick definition:
Tree Testing is meant to test how you think your content should be organized. You set up the ‘tree’ and users can click through until they think they’ve completed the task, then move on to the next.

Animated illustration of tree testing

Card Sorting can work a few ways, but is overall more flexible. It’s meant to find out how users think your content should be organized. You create your cards (pages of content) and set up categories that they can be placed into. Then, the user organizes the cards as they see fit. You can also set it up for users to edit categories.

Animated illustration of card sorting

This type of testing works best for information architecture because it allows users to focus on the content without the distractions design, colors, photos, etc.

Optimal Workshop has some great tools for Card Sorting and Tree Testing.

A/B Testing

Animated illustration of A/B testing

When to Test: During Design
What You’re Testing: Design Preference

A/B Testing works best when you’re trying to find out if users prefer one option over another. It can be static homepage designs to get a feel for whether or not they like blue or green, or it can be the presentation and functionality of a form and how a user would complete it.

A/B Testing is like comparing apples to apples to decide if users like Red Delicious or Honey Crisp.

This testing type and its typical use cases tend to work better during the design process, before code. You don’t want to invest a lot of time developing something and then find out users prefer something entirely different—that’s a ‘bummer summer’ on all fronts, especially to your budget. You can always mockup interactive prototypes if you want to test functionality (we tend to like InvisionApp). Just remember, keep it targeted.

Prototype Testing

Animated illustration of prototype testing

When to Test: After Design, Before Development
What You’re Testing: Functionality

Prototype Testing is for when you’re ready to really test your site or app functionality; how the user interacts with the image slider, if the placement of your call to action generates conversions, etc. You’re testing medium-to-high-fidelity interactive prototypes to see how users are going to react to the final product.

You can conduct this at a couple different stages—during wireframing and/or after design, but before coding. Ideally you could test at both stages; wireframing to get feedback on the foundation of page structure and workflow and after design to tweak that foundation and get feedback on style treatments—i.e. does this call to action button get lost on the page?

There are several tools that allow for easy prototype iteration—again, we tend to like InvisionApp. It lets you upload JPGs or PNGs and link between them. They’ve recently added new features for transitions, sticky navs, pop-ups, etc. You can create realistic prototypes from static images—it’s pretty great! We’ve also used Axure, which allows for much more interactivity and is great for more heavy-duty functionality testing; you can have build in forms with real text fields that users can fill out and submit.

It’s Almost Over.
There isn’t a one-size-fits-all solution for when and how to test. Each project presents its own scenario and we—as designers, developers, and strategists—have to adapt accordingly. But, the one consistency in all of this is this: keep your tests targeted. Find out what you want to learn and conduct all the tests necessary—timing and budget permitting, of course.

We're perfecting our user testing process here at Sanger & Eby. Contact us if you'd like to know more. 

Keep reading.
Load All
View More