Usability testing

A usability test is conducted to verify the usability of a software or hardware with the potential users. He is one of the techniques of empirical software evaluation, in contrast to analytical methods such as cognitive walkthrough.

Within a development process, for example, a website design, a re-design process - classic, scenario-based usability tests ( as presented here in detail below ), mainly performed when clicking dummies or beta versions were created. In the foreground is the detailed optimization of interaction processes ( such as the purchasing process with an online store ).

In addition to scenario-based usability tests in the lab, there are other test variants, for example:

• Expert-based usability evaluations
• On -site surveys (including benchmarking)
• Dropout analyzes using user tracking tools and (event -related ) on-site surveys
• Konzept-/Design-Tests (online and offline)
• Online usability testing on task-based online surveys and tracking of click behavior.

They are suitable for different question or issue positions and phases of a development process.

Method

In a usability test subjects are induced to solve typical tasks with the test object, they would do later in a similar way with this product. It will examine the points at which problems occur during use.

The subjects are asked to read a say, so that the observer knows what the person is thinking.

A developer is easy to overlook weaknesses in their own product and tends to defend it. Therefore, it should occur in the usability test, neither subject nor as a tester. Instead, an independent evaluation team should be used. If an instruction of the subjects required by the test personnel, care must be taken that by doing the test results are not distorted.

Before and after the test surveys can take place. A first survey, the selection of suitable subjects serve. Before the usability test, the prior knowledge of the subject, for example, experience with similar products, queried. Sometimes this card sorting techniques or wording tests are used.

After the test information is requested for the investigated product. These are held semi-structured interviews. Their implementation makes high demands on the interviewer. As part of these surveys, comparisons with other products are often employed. To this end, we ask the subjects, which consist advantages and disadvantages compared to similar products. Since classical, scenario-based tests are usually done with a small sample (12 to a maximum of 30 subjects ), the results of such comparisons can only show the first signs regarding where appropriate, present valuation differences.

After the usability test the vulnerabilities are analyzed and the product is optimized so that it is perceived as easy to use by as many people and so the user experience is improved.

Recording

In order to usability testing, and ultimately to be able to monitor potential vulnerabilities as accurately as possible, various forms of recording are used:

• Sound and video and screen capture
• Tracking software, for example, mouse movements, mouse clicks, keystrokes and web sites visited record ( when web usability ) of the user
• Eye tracking
• Observation

After the test all protocols ( text logs, video, eye-tracking data) are evaluated.

Applications

Usability testing can be done in principle, whenever something of a man-machine interaction subject. In this case both software and a real object, a possible object to be examined.

Example are:

• Mobile phones
• Any such software is subject to the operation by people
• Video
• Websites
• Telephone schedule information
• Machines
• Toys
• Technical documentation ( manuals, etc. )
• Medical Devices

Other methods of testing

Usability testing can be carried out as benchmark tests, for example to find out if a product is already good enough to use. For example, could be a usability goal is to improve a web application for booking a flight until a user manages to book a flight within three minutes. In the usability test, the time is measured for this task and waives thinking out loud in this case. The usability test controls the prototyping process.

Further usability tests are carried out as a comparison test, for example, when software purchase. It is tested that product has better usability.

The procedure is often in usability laboratories. Meanwhile, there are also the providers ' remote usability testing "or offer " Crowd usability testing ". Here, the tester will not be invited to a lab, but most of home. As for the tests, no laboratory is necessary and the testers to work from home, can achieve significant cost savings via this method.

Eye tracking in usability research

Eye tracking or eye tracking is used among others in usability and concept tests to record the eye movements of subjects. This can be determined which areas of the user's screen has considered and how intensely he has looked at this. In addition, it is possible to identify regions or elements which have not, or hardly receive attention. Data from Eyetracking-/Blickverlaufsstudien provide valuable insights for the optimization of layouts and designs, and are thus used more and more in the early stages of a development process.

Eye tracking provides evidence to

• Which elements of the user 's attention is directed screen
• What elements of a user sees, which do not
• How important is an area for a user
• Whether a user oriented or disoriented
• Whether a text is carried

Special features for Web applications

Conventional test method, in which participants monitored or surveyed on site are very costly and resource intensive. Therefore, these methods are for small businesses or web developers hardly in question. For this reason, less costly methods such as remote usability testing or proxy methods are used here frequently.

For remote usability testing, the whole process of usability testing is done online. It is not necessary to recruit volunteers locally and invite them to a lab or your own office. In general, these tests are performed on a specialized supplier, at the same time a subject - panel displays, from which on the basis of various criteria such as age, gender, Internet experience, purchase behavior, etc. people can be recruited. The running time of such a test will vary according to target group between a few hours and several days.

Another method, the Web application is specially expenses poorer and less expensive, is the proxy method. The user requests are redirected to a website through a proxy. This saves call sequence and viewing time of each page, as well as all user input. The assumption that the users now tasks, for example to find out about the offer of the relevant page, both the clarity of the layout, as well as the clarity of the text can be judged based on the stored data. The advantage is that neither of the page code needs to be changed, nor their functionality is impaired. The origins of this method date back to the late 90s. As a pioneer, developed by AT & T 1999 web event logging tool WET and developed in 2001 WebQuilt apply.

Special medical equipment

Since medical devices are safety- critical devices in general, special requirements must be imposed on a usability testing for these devices. So in addition to the usability factors of efficiency, effectiveness and satisfaction is also the security note. This method " UseProb " combines the requirements of serviceability with those of risk management. Also the usability test in a special usability lab should (for example, an experimental surgery) are performed in order to simulate a typical usage environment and situation. Special requirements for the verification and validation of the suitability for use of medical devices are laid down in the standards DIN EN 60601-1-6 and DIN EN 62366.

Special features of modeling languages

In addition to the already existing approaches for determining the usability of a wide variety of objects of investigation such as applications, websites, etc. usability testing methods can also be used for the evaluation of modeling languages ​​. Goal, it may be the process of reviewing a newly developed modeling language or extension over existing notations with regard to their usability. Both the aspects of creating, as well as the interpretation of a model can be tested using direct usability attributes, usability influencing meta-properties of models play a role. The usability attributes of modeling languages ​​are measured concretely in the test metrics and calculated. Of 28 surveyed in 2013 publications on the topic "Measuring the usability of modeling languages ​​" put 18 the definition of usability ideally in the form of individual attributes fixed. The criteria learnability, effectiveness, efficiency and user satisfaction were mentioned in more than one -third of the sources examined as criteria of usability. The actual calculation of metrics by which these criteria can be measured vote, largely consistent across publication.