UX researchers have developed many techniques over the years for testing and validating their ideas.
Here are ten essential methods to learn and employ on your next project.
We cover many of these in detail at our UX Bootcamp in Denver.
- Moderated In-Person Usability Testing: This fundamental technique is used by usability professionals for obtaining feedback from live users interacting with everything from paper prototypes to fully implemented applications.Despite improvements in technology that allow for asynchronous and remote testing, there will always be a demand for co-located facilitation for testing complex applications, hardware and mobile usability testing, which is more difficult to reliably test with remote methods.
- Moderated Remote Usability Testing: Interacting with participants and probing on comments and behaviors provides rich insights into users’ motivations. Using ubiquitous remote meeting services like GoTo meeting, it’s easy to have participants from around the world participate in a facilitated usability test. This works best with website and web-based apps but we’ve also successfully tested desktop software using multiple computers connected to the same meeting in our usability lab.
- Unmoderated Usability Testing: If you’re looking to detect more than just major differences between designs or to estimate the prevalence of an attitude or behavior, you’ll need a reasonably large sample size. It’s difficult to conduct more than a few dozen moderated sessions so using software like our MUIQ® tool, Loop11 and UserTesting.com is essential when gathering and analyzing data from samples of 30 to 1,000+ participants. Many of the options available have become more advanced in recent years, allowing users to walk through tasks, record their screen and even think-aloud while attempting tasks.
- Card Sorting : One of the more popular ways to create and test a taxonomy or navigation structure is to have actual users, not product developers, organize a representative set of items into groups and then label the groups. Software from Optimal Workshop provides electronic “cards” that participants can drag and drop and name allowing for quick studies with large samples.
- Tree Testing: A lesser known method for testing website navigation, tree-testing, is like a reverse card-sort. Instead of having users place items into categories, you have them look for items in established categories. This approach is more similar to actual browsing behavior than card sorting and we strongly recommend it for anyone looking to improve the navigation or taxonomy of a website or application.
- Voice of Customer Surveys: Surveys are an efficient and cost effective way of understanding who users are and what they are trying to accomplish on websites or software. A VOC survey can be emailed to prospective or current customers or set to pop-up on a website. Using just a few questions and a top task analysis can build the necessary foundation of information for knowing what to fix and what your users are trying to accomplish.
- Heuristic Evaluations: With a set of tasks, user profiles, and a set of interface guidelines (called heuristics), researchers can identify many of the problems users will likely encounter during a usability test and in actual use of an application. This staple of discount usability engineering is best performed with multiple evaluators who have knowledge of the domain and of good interface design.
- Keystroke Level Modeling: When interface efficiency is important you’ll want to measure how much a new design can improve task times. You can actually estimate time saved (or lost) using some well tested approaches that are based on average human performance for typical computer based operations like clicking, pointing and typing.
- A/B Testing : The usability lab is great for simulating experience and testing more prominent design changes, but can make it hard to collect enough sample for analyzing subtler alterations. To truly put your design to the test you’ll want to subject it to actual website traffic by randomly assigning visitors one of two design options (typically referred to as A and B). A/B testing is especially important when trying to detect smaller differences in designs. Even a modest difference of 1 percentage point on product purchases can translate into millions of dollars of profit or loss over the course of a year on many high-traffic website. To determine if these differences are beyond what we’d expect from chance fluctuations you need to test each design with thousands of users and see which one wins.
- Click Testing: A user’s first impression can matter a lot, especially when interacting with navigation structures and homepages. Understanding what design elements users are attracted to and where users click first provides quick quantitative guidance on what’s of interest and what’s ignored.