Start with alpha testing. Then move on to beta testing. Then advance to the stages in product development when systems and interfaces are tested for viability. Then make adjustments and corrections before conducting additional testing until – finally — you reach the viable product stage.
Rarely will the initial incarnation of a new product design ever survive completely intact after a user testing cycle. Design staff and other personnel work out many of the major issues beforehand, but eventually user testing is needed to for a device to achieve the viable product stage.
User testing is a crucial step in product development. The more effective you can make it, the more relevant and valuable the data you collect will be. Without effective and efficient testing, the gathered data may be questionable. This has a direct effect on the product and — ultimately — the bottom line.
There are a few general guidelines which can be applied to virtually any user testing phase which will improve the effectiveness of that testing. These test parameters will help collect the most relevant and reliable information needed to improve your product design.
Include Your Target Demographic
It seems like such a simple concept, but it is sometimes overlooked. If the target audience is comprised of Asian males aged 18-38 with medium to high income, then including white women aged 55 and older in your test group is not going to provide much useful information.
What you want is a substantial cross-section of users within your target demographic. If you utilize a true variety of users from your target group, you will get the best data that will help you improve your design.
Outline Testing Tasks
In order to provide relevant data, users should have a set of tasks they need to perform. As the designer, it’s your job to try and determine the nature of those tasks. Your testers represent the future users of the product or service you plan to offer. Test tasks should reflect those which the target group will be performing during normal use of that product or service.
Utilize a Test System
A test system is exactly what it sounds like; a system used for testing. Test systems are useful in that they are specifically designed to capture data, and can be changed without affecting the final product.
Instead of testing the entirety of a product, certain portions of it can be tested and the data collected, then the system can be reconfigured to test a different section of the product.
Define Target Data
There are essentially two types of data to consider; Process Data and Bottom-Line Data. Process Data is that which shows the step-by-step procedure of accomplishing a given task. In other terms, it shows the ‘How” and possibly the “Why” a user reaches the completion of a task. Bottom-Line Data only shows the end results of performing the task; the “What”. How long the task took and whether or not it was successful are standard measurements for Bottom-Line Data.
As a designer, you will more often want Process Data over Bottom-Line Data, at least in the earlier phases of testing. How a person moved from one section to another and why that path was taken instead of another is exceedingly more useful in fixing an interface failure than simply knowing, “I can’t get there from here.”
Although it is not commonly used, allowing your users to provide verbal feedback as they actually perform the tasks can be tremendously useful. You or the system provides a cue for the start of a task and the user then explains aloud what is being done and his thoughts while performing that task. This information is recorded as it is spoken, providing invaluable data about design errors or user confusion.
These five techniques are not the only ones which can be used to improve user testing, but they are some of the most beneficial. There are multiple sources of further information if you want to learn more improving user testing.
You may think your design is completely awesome, but proper testing will reveal flaws in even the best designs. If you use these methods to improve your user testing, you are practically guaranteed to deliver a better product than if you hadn’t tested it at all.