Usability tests will be required if you wish to deliver a user interface that is effective. The likelihood of coming up with an effective user interface design without performing any testing and research is not very good even for an experienced designer.
Usability testing involves testing a site and evaluating the actions of users that comprise the site’s target audience. Our focus will be on interaction design that is centred on end users and how to proceed with testing when you are developing such a site.
Every site has a purpose, and the goal of testing its usability is to determine if this site fulfils its intended purpose with respect to its users.
Usability tests each have different goals that are associated with the specific objectives of the test. The test results will either be categorised as a baseline or control measurement. Since numerous tests will be conducted over a time period, the results are generally compared with the results of the baseline tests.
The basic principles governing usability tests are as follows:
What are the user’s feelings regarding the tasks he/she completed? Did the user feel confident or stressed? Is this a site the user would recommend to others?
The number of steps and the amount of time required for a user to complete a specific task is measured. For instance, finding a product, buying a product etc.
How often do users make mistakes when attempting to execute these tasks and what are the adverse effects of these mistake? Some mistakes are recoverable, others are not.
After a period of time in which the user does not access the site, how much information is retained regarding the user interface and navigation process?
These are the basic principles governing UI testing. However, each tester must have their own goals for usability. Based upon these goals, the tester can monitor the site’s users and update the user interface accordingly.
Some individuals have an improper interpretation of “usability testing”. Simply gathering data and opinions on an application is nothing more than performing market research. This kind of research is quantitative research and not usability testing.
Usability testing needs to have a means of systematic observation that is carried out within controlled conditions. These tests are designed to determine how site visitors utilise the site.
If 80% of site visitors say that the site works well, this does not infer that the site was tested for usability and had positive results. It only indicates that most of the site visitors that were questioned think that it works well, however, this is not useful information for improving the user interface.
A critical aspect of testing your user interface is getting users involved as much as you can. Rather than asking them about the appearance of your UI, get them to execute some actions on it.
There are numerous factors that have an effect upon the process of browsing. The majority of your users will not have the ability to name them, however, they will normally be able to demonstrate them to you when they are using your interface.
Several methods are available to perform usability testing. We will examine the most important methods to assist you in selecting a method that is suitable for your specific testing purposes.
When you are testing a site you need to have realistic test scenarios during which a user needs to perform a specific task or tasks using the website being tested. During the test, the actions of the user should be recorded using a product like CrazyEgg. You may also want to have a questionnaire to gather feedback information regarding the site being tested. Eye-tracing and co-discovery learning are testing techniques for usability that may be utilised in these methods.
This is a relatively simple, general methodology where you work with a small number of users (four to six). This test is called ‘Hallway Testing’ because the test subjects are random individuals who may be passing by in a hallway. This testing method may be employed when your site is not geared toward a specific target audience.
This type of testing should be used in the early part of the design phase. You should test fairly quickly and frequently. You will need to get several groups of participants. The testing process is fairly straightforward. Get five random test subjects, perform your tests on them and record the results. Solve any outstanding issues you may find and then find five more test subjects. Repeat this iterative process until outstanding issues have been solved. After several iterations the critical usability errors should be significantly reduced and you can then focus on the remainder of your design. Your site will need further testing down the road, but at least you will have solved some of the most pressing usability issues, so you can proceed with your development.
The concept behind using random test subjects is that you want your interface tested by people who have no familiarity with it whatsoever. You want to test with individuals who have never used or seen your user interface before. This will give them all a common starting point. If people who have never used your user interface before are able to execute all the assigned tasks easily, it indicates that most site visitors will be able to do likewise.
This type of testing may be employed when you anticipate having users from across the globe. Gathering these users into one location may not be feasible or cost-effective.
This type of testing facilitates remote testing with the tester and user separated by distance and time. This type of testing can be facilitated by using remote applications like WebEx and TeamViewer.
The tester will be able to obtain the click stream of users along with logs of user mistakes and other incidents that happen during site interaction. Subjective feedback can also be collected from users.
The good thing regarding these types of tests is that they are executed in the user’s own environment. This means that a real-life testing environment can be simulated. Of course, this type of testing also permits you to work with individuals from across the globe without incurring additional expense.
Executing a synchronous remote test can be more challenging than it may appear. Managing cultural and language barriers via remote access may reduce the efficacy of your test. Distractions and interruptions in the environment of the participant may be other challenges that are not possible for you to resolve remotely.
This usability testing approach involves getting experts in your market niche to evaluate your user interface. The issues involved with this type of testing are primarily logistical and financial as it may be expensive to gain access to experts in your field.
A/B or split testing is probably the best known of all usability and user interface testing methods. Its goal is to identify the site elements that increase the engagement and interest of users.
This technique is referred to as A/B, since two versions of a site are tested and then the results are compared. The sites are identical with the exception of a single variation that may have an impact upon the behaviour of users.
When testing is occurring the site is monitored with tools such as CrazyEgg and Google Analytics. When A/B testing is ongoing the two different versions of your site will change randomly so that you can visit the site once and see version A and the next time you visit you may see version B.
This testing technique is normally employed without users having any knowledge of it. It is often used to maximise conversion rates and improve the profitability of a site. Amazon was the first major company to use this method extensively, but it is now used by numerous organisations.
Although this technique has been used primarily by ecommerce sites. A/B testing may be employed to improve the design of user interfaces as well. It can be used to determine the efficacy of two competing interface designs.
Executing several tests with a limited quantity of test subjects is usually better than testing a single time with a large quantity of subjects. This will result in numerous quality tests as opposed to one large test of questionable quality. Approximately five subjects should be adequate for each test and provide enough data to work with for a time period.
The theory behind this concept is that after you identify that a few individuals are confused by a website feature, you will gain little from performing the same test on more test subjects, since the additional subjects will probably have problems with the same elements. Therefore, if you resolve the outstanding issues and perform the test again you will gain more valuable insights. Several iterations are required to derive the most from this type of testing.
Many experts say there are problems with this type of testing. They claim that usability applies to a large population sample and not a specific number of users. This indicates that problems with a user interface may not be detected with a limited sized test group. Nevertheless, this theory does not suggest that you should execute only one to two tests with a small number of test subjects.
Tests must be executed at least once per week, and possibly twice/week while the design process is ongoing. Long design processes will necessitate a larger number of tests. During the entire testing process this could add up to a significant number of test subjects.
The tests will have greater efficacy if you test users that have abilities that cover a wide spectrum during the early testing phases. During the final stages of testing the user interface will be improved and you should try to get test subjects that will be part of your site’s target audience.
When you are executing your usability tests, you should take note of the design elements that function well, as well as the problem areas. Keep testing these areas repeatedly. The theory is that design elements that do not work well need to be eliminated, and elements that are functioning well and appreciated by users should also be noted. Try not to alter the design elements in your site that are clearly working well. Your primary focus should be on the problem areas of your site rather than areas that are already working successfully.
Usability testing is a necessity if you are developing a user-interface that you want to be successful. Usability testing is particularly helpful if you are developing a user interface that will be used heavily on mobile devices, since usability is such an issue on mobile platforms.
For many sites, usability testing is not that costly, but it will pay great dividends in the future if it is done properly.
There are many differing usability testing methods that can be employed. We have referred to several of them above. You have to determine which testing methodology is best suited to your particular application. So, if you are in the midst of a design process, it would be to your benefit to execute some user testing. It will make your user interface much better and your users will appreciate it.