Which is the best way to test in UX?

There are multiple ways to test. To decide what will suit your project best, consider these two correlating factors: confidence and effort.

The more confidence you require, the more effort you’ll need on your test, and vice versa. A good approach is to start with low effort and low confidence. Testing more options is easier at first, and you’ll narrow down to fewer options for more robust testing later in the project. It is not uncommon to test multiple iterations using different methodologies.

Broadly, you could do evaluative research, where the users declare or pretend they would do something. Or an experiment, where the user does something that signifies their intent. None are perfect, but we can learn a lot from both.

Common Research Practices

Intercept Testing (aka Guerrilla testing) The metaphor here is Guerrilla warfare, where the approach is frequent, less organised attacks. In UX terms, you intercept people on the street (usually a coffee shop) and you get them to role-play a scenario using the prototype or designs you built. You observe their usage and learn from the results.

In Lab Usability Testing In Lab Usability Testing is like the testing above; however, you recruit the participants in advance, prepare more and record the session. It’s a more controlled environment and you can get more confidence in the results.

Surveys or Click Test Studies If you want to know how many people would use something or which experience is better, you’re after quantitative results, where you can research a large number of individuals. You can validate your experience by showing your designs or prototype in a survey where you can ask the participants to answer questions about the designs, or attempt to do tasks.

Common Experimentation Practices

Landing Page Experiments: A website describing the experience you want to sell. The catch is that the service is not available yet and users can register interest. If they sign up, it means that they would likely use that service. Wizard of Oz: Remember the musical? The Wizard had no real powers, he made everyone believe he had them. Here, the user will believe he’s going through the actual live experience, but in fact it’s all fake. Think of a chatbot operated by a human in the back end. It’s very effortful way to test with few participants, but you can get real results.

It’s not over yet. UX Design is not a linear process!

Do not seek praise. Seek criticism.
Paul Arden, Creative Director, Saatchi & Saatchi