Eye-tracking can be an invaluable tool and I thought I would share our top 5 tips for using one as part of a usability project.

1 – Over-recruit for your desired sample size – Some participants will not provide sufficient eye tracking data for analysis. This means that you will need to recruit about a quarter more participants than is required for statistically sound analysis of heat maps or gaze plots. It is imperative that this is set as an expectation with stakeholders or clients at the beginning of the project.
2 – Allow plenty of time – In comparison with moderated user testing, more time is generally required for the set-up of the project, the running of sessions and subsequent analysis.
3 – Pilot your study -This will ensure that prior to seeing actual users you are confident the eye-tracker is set-up correctly and working to ensure data is being logged and measuring your stimulus. Without this important step all of the data collected may be useless.
4 – Use a dedicated eye-tracking set-up – This should include a chair on legs (opposed to an office swivel chair) and allowing the moderator to control the eye-tracking equipment (depedning on the eye-tracking hardware this could include the use of a second computer with data logging capabilities). Also, ensure the moderator sits well behind the participant and out of their line of sight.
5 – Is eye-tracking the right tool? – Lastly question whether eye tracking is going to add value to the project. Will it add more than just moderated think out loud user testing alone?

Example eye-tracking heatmap from UsabilityOne

Is there anything you would add to the list for getting the most out of eye-tracking?

For more on eye tracking see Ciaran’s article from the newsletter last year.

PDF’s in general should be avoided as they cause many usability problems. However there are times when PDF’s have their place on the web; for instance if a document is meant to the printed then PDF is the correct medium. If your site contains PDF’s it is vital that users are aware of when a link will open a PDF.

Indicating the presence of PDF links helps to avoid PDF shock; that is giving users a nasty surprise of being dumped into a PDF file. There are a number of different ways of indicating PDF links. One method is to include a PDF gateway page; a page which summarises the PDF content, warns users that they are getting a PDF file, as well as presenting the file size. Gateway pages are supposed to prevent PDF shock as well as let users determine if the PDF is worth opening. However in our labs we often see users experience difficulty with gateway pages. Users assume the gateway page is the ‘result’ and don’t continue to click the PDF file. It is in-fact breaking the link to the real file. Many government websites use this technique and I’m starting to wonder if it really is the best way to display PDF’s.

My preference would be to include the PDF symbol next to the link along with the file size. If this is included in the link anchor text then screen readers such as JAWS can easily determine the presence of PDF’s. Surely this is enough to determine the presence of PDF’s without the need for a gateway page.

What do you think is the best way of presenting PDF links?

We love when our clients give us exciting and innovative designs to test and often we agree that an interface that is not immediately intuitive is learnable and so can be considered to be usable for the target audience. Several factors affect whether an interface that requires learning time is usable.

1. Does the effort outweigh the benefits? The effort required to learn the interface sometimes outweighs the benefits of presenting that content in a complex interface. We feel that innovation is essential, however, sometimes the user is required to do a little too much to access content and they quickly lose interest in the content and the interface as a whole.
2. Will the initial effort required to come to grips with an interface put the user off? It is essential that the interface provides very simple and intuitive core functionality to ensure that users are not discouraged during initial use. This will allow users to use core functionality and learn more advanced nuances of the interface over time.
3. Does the interface hide the purpose of the website? When the interface is too complex or excessively visually based the purpose of the content it provides can get lost.
4. How often will the user visit the interface? Some interfaces will be heavily used and re-visited regularly. In this instance the user has the opportunity to learn the interface and become familiar with its features and nuances, however, where a user visits the interface irregularly or perhaps only once, the opportunity to learn a complex interface does not exist.
5. Who will use the interface? Different user groups are more willing to learn and explore than others. While it is true that younger and more technically savvy users will get used to an interface more quickly, these are often the ones that give up fastest. In comparison to older users who are often more cautious and systematic in their approach, more proficient users assume knowledge and are typically less forgiving of an interface that requires learning.

As web technologies advance we hope that this will enable designers to create increasingly interesting and innovative interfaces, however, we hope that the users don’t get lost in the process.

Back in the early 1990s I began exploring how task models could be used to support usability evaluation.  Twenty years on, and I’ve found a potential contender in the business world. But first, let’s explore what I mean by task model.

A task model is anything that enables you to graphically represent a task, its components, and the logical and temporal relationships between those components.

In the mid 90s the software engineering community adopted UML (User Modelling Language), along with tools such as Rational Rose transforming UML models into source code. But, what’s available in the User Experience community?

Many task models have been developed by Human-Computer Interaction researchers to support the representation of user interaction with systems. GOMS (Goals, Operators, Methods and Selections rules) and its extensions (Keystroke Level Model or Cognitive Perceptual Model) being the most popular. These models remained mostly in the academic world – complexity being their worst enemy. Other popular models are flow charts or mind maps.

However, you may ask, “What models are there to help us, usability practitioners, represent the interaction between users and systems?”

The potential answer is BPMN (Business Process Modelling Notation – http://www.bpmn.org/). It is a notation gaining adoption through the industry world. Its strength is its simplicity. With little explanation, the notation is readily understandable by all stakeholders. Although BPMN’s aim is to represent business processes, it can easily be adapted to represent user interaction with systems and processes.

The figure above presents an illustration of a recruitment process (from http://www.rickgeneva.com/wp/posts/swimlane-lane-or-pool-learn-to-swim-first/). The circles with envelopes represent events that trigger a task to start (the event is the email being sent or received), while the empty circles represent the start and end of the task.

So, why do I use models? I use them to communicate and validate user interactions or processes, to support usability evaluation, to help define the language to be used and/or to create training material.

Twenty years later, and I am still puzzled why models are not more widely used amongst the various members involved in the design and development of digital solutions.

As a new year begins we often reflect on the changes we have observed over the last 12 months. One of the more interesting changes that we have noted, is how people view the online world and the role it plays in their lives.

We have a set of introductory questions we use on most projects across numerous methodologies (interviews, focus groups, one-on-one user tests and card sorts) that break the ice and provide us with some background for each of our participants. These cover things like, common activities online, amount of time online each week, where they access the web and years using the web.

In 2010 we had to change these questions because the web is now so ingrained in people’s lives that our old questions had become difficult to answer.

In the past, people had a clear model of their online lives and were very conscious of being online. Today, being online takes many more forms and is often more subtle. For example, many people are either unaware or confused about whether they are online when using an app on their phone. Further, if we compare the experience of using an app to check your bank balance to logging into internet banking via your desktop the former is a much less involved process which at week’s ends you may have forgotten.

This shift began to hit home in the last 12 months when asking people about what they do online. Some really struggled with the question and responded by saying “well, everything really” . Often, we now have to prompt people to find out about online behaviours such as doing research, banking, buying tickets and booking holidays. In some ways asking about the web is like asking about how electricity impacts your life, it is there but we tend to take it for granted.

Possibly the most telling feedback is that we often hear people say “I don’t really surf the web, anymore”. I always have a chuckle when I hear this. Firstly, the term ‘surf the web’ takes me back to the 90s and secondly it is a term that the digital industry has long forgotten. The statement does emphasise the fact that (mostly) people go online to get things done.

So if online is just another, not too salient, aspect of people’s lives, how does this impact user research? We need to stop thinking of online as a thing (particularly as a mystical and magical land). And start focussing on the various topic areas such as banking or paying bills.

Do you think this is accurate and what trends have you seen?

As we are well aware, the internet has changed the way we do things. The way in which retailers ask us to sign up for loyalty programs or clubs is one of those things that has changed. It was not long ago that you would stand in the store and fill out a paper-based application in order to join a loyalty program. Now you are handed a card and asked to complete the application online in your own time. A recent experience highlighted the importance of a smooth integration between the off-line and on-line elements of the application process.

During a recent visit to Borders bookshop, at the checkout I decided to sign up for their new VIP loyalty program. I was handed the membership card and told to complete my application online. When I arrived home and logged onto the website for the life of me I could not find how to complete my membership application. Despite my expertise and experience in user interfaces, this task proved difficult. I became frustrated not being able to locate the call to action from the homepage, referencing both the instruction pamphlet and the membership card in search for further assistance. Finally I found the text link labelled ‘Join Boarders v.i.p’ in the top right of the screen.

If your company requires users to finalise what begins as an offline process via your website keep the following in mind:

Highlight promotional features prominently on the homepage: As I was coming to the site with the sole purpose of completing a process I had begun offline, I wanted the process to be as quick and seamless as possible. I was looking for a prominent call to action, something that grabbed my attention straight away. I was looking for a large advertisement, particularly as this loyalty program was a new feature.

User workflows differ when the task begins offline: In my example above, when I arrived on the website I was not thinking in terms of ‘joining’ as the link suggested, I thought I had already joined due to having received the card in-store. Instead I was looking for a way to finalise the process, looking for key words such as ‘Register your VIP card’. An additional call out box on the homepage which referenced my offline step (receiving the card) would have been extremely helpful.

Know your audience: Joining a bookstore’s loyalty program does not necessarily require users to interact with the company online. Many customers may join the loyalty program only to use the benefits in the physical store. You are therefore likely to attract more customers that may not be so familiar with the internet than if it was an online bookstore for example. Providing clear calls to action as mentioned above, would greatly assist these users.

Forty years ago not many people would have been able to accurately predict the advancements made in technology at the turn of the century and beyond. However, one prediction that has held true (relatively) is Moore’s Law conjured up in 1965.

Moore’s Law states that technology and computing will develop at an exponential rate. That is, every 18 months or so performance of technology is double that of what it was two years ago. In essence, technological growth is continuously accelerating.

But accelerating towards what? Exponential laws have no finishing point, they merely continue to grow at an enormous rate. So is this trend really to continue at the same rate or will there one day be a plateau? You would think there would be what with a limited number of resources in the world, but then again, I continue to be amazed by the latest products to come out.

Figure 1 The exponential growth of the number of transistors in computer processors (Taken from www.intel.com)

So without an end in sight, what does all this new technology mean? Well, put simply, those of us working in the computer and technology industry will have jobs for many years to come.  There is always something to push us forward, whether it be a competitor’s latest release, or a new idea that has only just taken off (3D television comes to mind). For us at UsabilityOne it means we will always have something else to look forward to testing, and for those of you who are thinking of updating websites, softwares or other technologies, it means you have a range of options that are always changing and always improving. So which one do you choose?

Although it can sometimes be overwhelming to try to take in the seemingly limitless amount of technology around us, we should not despair but should embrace. The most important thing we should take away from Moore’s Law is that we should always be on our toes, because as the nature of the law suggests, now more than ever before there will be new gadgets and gizmos coming at us.

Now, you may be sitting there thinking “I have all these proposals on my desk for website improvements and new designs but how do I know what is going to work best”. Of course there is no one simple answer – it depends on your business, your goals, your budget, etc. One tip we can give though is to work from the top down: What does your company do? What are your goals? It is crucial to have a clear sense of what you are aiming to achieve by implementing a new design technology. A couple of examples are listed below:

• A company wants to increase its sales so it implements an eCommerce feature on its site
• A university wants to improve its student interaction with its online learning facilities so they aim to implement a more interactive website using javascript and videos

To make the most of the myriad of available technologies, first decide what you want and/or need, then conjure a plan to best implement what is available. Identify the goals and objectives of the business, any boundaries which impact on your decision, and assess your timelines to choose a suitable direction in which to proceed. You may not need the best technology, you just need the best technology for you.

UsabilityOne Managing Director, Shef Bey was interviewed earlier this week by Power Retail. See a video of the interview here…
http://www.powerretail.com.au/insights/usability-testing-essential-for-optimised-online-user-experience/

At UsabilityOne we have extensive experience, not only in evaluating search functionalities across a wide range of websites, but also in helping our clients gathering user requirements and designing the user experience for their search functionality.

Search is an important user interface element in any website. Research shows that around 10% of users will elect to use search as their primary means of locating content within a website. However our experience indicates that most users will turn to search when they have been unsuccessful in locating content via a site’s navigation structure. In addition, when it comes to search, Google sets users’ expectations to a high standard.

So, how do we evaluate the performance of a site search, from a usability perspective? Well, this question can be addressed from two separate angles: firstly, from a classic usability evaluation perspective we can compare how the site search performs against search best practice and observe users searching the site. Secondly, we can look at Key Performance Indicators (or KPIs) to monitor how the site search is performing over time.

At UsabilityOne we suggest the following seven KPIs that relate to search:

1. Average searches per visit
2. Percent of visitors using search
3. Percent of zero results searches
4. Percent of zero yield search
5. Results Page views per search
6. Search result to site exits ratio
7. Percent of search refinements

Unfortunately, the above KPIs do not monitor how well suited to the query the search results are. Some qualitative methods need to be put in place in order to monitor both the appropriateness of the search terms used and also the relevance of the results returned.

The story behind each of these quantitative and qualitative measures will be published in our next newsletter.

Recently, I have been working with a calculator that deals with finances. I’ll be the first to admit I’m not a financial genius and at times during this testing I found myself double-checking over the meanings of various aspects and outputs throughout the calculator. I wasn’t the only one; throughout the testing a number of participants could not properly articulate what the output actually meant.

Initially to combat this, I had the genius thought of adding ‘?’ symbols where there were issues for the participants, but the range of areas where these could be added were huge. By the end of the testing there could have been ‘?’ symbols all over the place, which in itself creates clutter and hence the usability of the application deteriorates.

So I pose the question: “How much help is too much help?” At what point do we draw the line on how much information we can give to users and just say “if you don’t get it by now, you’re probably not ever going to get it”? In this case it is important to assess who you want to use your interface. Should it really be a case of teaching whilst users are interacting, or should the ‘?’ icons merely be there for guidance?