Posted in Website Usability Testing	January 9, 2008

The difference between heuristic and cognitive usability studies, and how to conduct both.

When conducting a usability study with real users (as opposed to an expert review), your main duty to the test is to pay attention to the user. But above all, remember that you are scoring the website, not the user.

• Make the user feel comfortable
• Be careful how you phrase your questions
• Do not lead the user
• Tell the user to “think out loud”
• Pay attention to facial gestures
• Identify frustration and emotion
• Keep track of hesitation and time delays
• Refrain from attempting to assist the user

There are two types of testing methods when it comes to real users, cognitive and heuristic. During a heuristic evaluation, the study is usually conducted one-on-one with a moderator and user with an advanced working prototype of the website. With a cognitive walkthrough, you may have many more users involved and may be more of a theoretical prototype evaluation.

Depending on your test, the people who should be in the lab during the process may vary. You will always have the lab moderator, but you may also have an assistant who is most familiar with the site, and possibly the stakeholders if requested. Be careful not to invite anyone who is “close” to the design such as the web designers and architects, as there is more opportunity for distraction and intimidation for the user.

Only 43 seats left at the Mequoda Summit Napa Valley 2008 where you can learn more about conducting usability labs, and how to optimize your website architecture. Register now.

Paper Prototyping

If you are only testing usability AFTER your site has been developed, you are only looking for validation, not actual usability. But how can you test a website that has not even been written in a line of code so far?

Paper prototyping includes a mockup of the user interface using paper, and sometimes cardboard (for PDA, phone and other three dimensional interfaces). The user is able to arrange topics, articles, and other critical elements of your website in a way that makes sense to them. Using a paper prototype, you and your team can often "walk-through" your design concept and how well it will work with actual users and tasks.

The paper prototype is a "low-fidelity" model of your website. Because no one could possibly mistake it for the real thing, you are likely to get a wealth of feedback about what process might work better or what labels and functions make more sense to the user.

Using a paper prototype, you and your team can often "walk-through" your design concept and how well it will work with actual users and tasks. Nielsen Norman offers a $68 DVD on Paper Prototyping.

Image source: Nielson Norman Group

Only 47 seats left at the Mequoda Summit Napa Valley 2008 where you can learn more about conducting usability labs, and how to optimize your website architecture. Register now.

Card-Sorting

This test is especially helpful if you have a lot of categories or sections and you want to know how users expect to see those organized.

As the name implies, a card sort requires the test participant to sort cards, each with a word or statement printed on it, according to the user's mental model of the relationships between the words or statements. The test facilitator then records the sort in a spreadsheet or in one of the online sorting tools. A cluster analysis of the results from several test participants gives designers an excellent view into what is important to the user and how they refer to it in their own language.

Card Sorts can be conducted with individuals and groups depending on what kind of information you are trying to elicit. In either case it's important to leave some blank cards available, as users might want to use words you have not provided.

Image source: markboulton.co.uk

Only 47 seats left at the Mequoda Summit Napa Valley 2008 where you can learn more about conducting usability labs, and how to optimize your website architecture. Register now.

Link Affordance Test

A great affordance test is to print off web pages in grayscale, give a test user a yellow highlighter and ask them to highlight anything on which they think they could click. The first thing you will notice is that anything that is underlined will get highlighted. Those words that may or may not be colored but which are not underlined will most likely not look like links to your test user. Headlines that are not linked might also get highlighted because they are different in size, weight and location to the main content, all possible visual cues to a desperate information seeker. Users may also highlight images that are not linked while overlooking images that are linked based solely on subtle cues like outlining, drop shadows and 3-D effects.

Heuristic Evaluation (one-on-one)

A heuristic evaluation is one of the more popular methods of usability testing and takes place with a functional website, likely in the last stages of development or right before a redesign. This is the type of lab setting that we showed you in Mondays tip How to Conduct a Usability Lab, and requires a computer with a webcam and usability software like Morae, a moderator, and a user.

In this usability test, you are required to come up with 5-10 major tasks such as:

• Subscribe to the email newsletter
• Download a free report
• Subscribe to the RSS feed
• Purchase a book
• Unsubscribe from email list

The user is then recorded via webcam and the usability software while being timed performing each task. You will want to set a time limit for each task, from 60 seconds to three minutes. In general, if a user takes three minutes to find and subscribe to an email newsletter, the site has failed this task. Come up with a scorecard to document these times and scores.

Why is usability testing important to your website? When designers are confronted with these test results they often resort to the "But they will learn where things are by using the site!" defense. After all, one of the driving forces for designers is to create something that's never been done before or a look that is intriguingly unconventional. Unfortunately, it is not at all important to a user to find out how to use a clever design. What is important to the user is to get in, get the information they want, and get out.

In Friday’s article, we will go over how to evaluate the feedback you receive and construct it into a deliverable for you or your client. If you are interested in hiring Mequoda to conduct a usability study on your website, please fill out our contact form.