6.1 Uncovering the Initial Mental Model
discovering how people initially interpret your sketched interface from its visuals
When people see your system for the first time, they will immediately start forming a mental model from its visuals: what your system is for, what it lets them do, and how to operate it. This initial model influences how they use and explore the system. As they use the system over time, they will (hopefully) correct any misconceptions they had, and ‘fill in the blanks’ for the parts of the system they either did not fully understand or did not know about.
This initial mental model is important. An incorrect initial mental model leads to almost immediate frustration, errors and inefficiencies. Because incorrect understandings linger (perhaps for a considerable time), they interfere with ongoing learning and use. They may also cause people to abandon your system, especially in cases where system use is discretionary. This is why it is so important to ensure that your design portrays a good mental model.
Fortunately, you can uncover this initial mental model in the very early stages of design – even in the sketching stage. The method is quite simple. After briefly introducing your design, just ask people to explain, in detail, their understanding of every visual element on the screen. If their explanation doesn’t match your intention, then you have spotted a problem in people’s mental model that you should try to fix in a redesign.
Case Study: Usability of a Fax Machine
The Situation. I once had problems using an unfamiliar fax machine to send a time-critcial document. Even though I was doing the most basic task expected of a fax machine (to send a fax), I was unsure of its operation, or if the fax was actually sent. Frustrations piled onto frustrations. Afterward, I decided to use this fax machine as a case study for teaching various usability testing methods. Using PowerPoint, I sketched the front panel of the fax machine, which appears on the next page. Back in the classroom, I projected this sketch onto a large display. I then asked for a participant volunteer – a student who had occasional need of a fax machine – to help uncover problems with this system.
The sequence that follows uses this sketched fax machine to illustrate how you can uncover a person’s initial mental model. Later chapters will also use this fax machine to illustrate other methods.
There are advantages to having your participant work over a large sketch projected onto a wall vs. a small sketch on a table or desktop computer.
First, if you are videotaping the session, it is much easier to set up the camera on a tripod and point it at the wall versus trying to position it over a table.
Second, if another observer is taking notes, that person can sit away from both of you but still see what is going on.
Third, you can easily see what the person is looking at and pointing at because the distance between that person and the screen elements is larger.
Uncovering the Mental Model
1. Preparing for Data Collecting
You need to set up the room and your materials so you can record what happens. Ideally, you should set up a video camera pointing at the person and the sketch so you can record the session for later review. Alternately, you could have another person present whose job is just to observe and take notes. You could try to take notes about what you see, but there is usually just too much going on (and at too fast a pace) for you to both give instructions and record activities in real time.
You need to tell the person what you are going to ask him to do. You also need to be very clear that you are looking for any problems he may have doing some of the things you are going ask him to do. Gommol and Nicol (1990) suggest the following dialogue as a standard opener to a test involving users.
You need to introduce the system to set the context. That is, give your participant just enough information to help him get into the right frame of mind. You may want to add information about the interface that is not apparent in the sketch, but would be apparent in the final system. Avoid disclosing information or hints about the actual system operation, as this would unduly influence how the person forms his initial mental model.
Now tell your participant what you want him to do and how to do it.
Before reading on, try the above exercise yourself. That is, walk through the interface and note down what you are certain about, where you are just guessing, or where you haven’t a clue.
The caveat is that ‘trying it yourself’ wouldn’t work if you were the actual system designer, as designers already have a mental model of the system. The reason why designers use test users is to see whether the users’ mental model as acquired by the visuals matches the designer’s intended mental model.
A sampling of typical things that a person may say is illustrated here. The first segment shows the person explaining the labels. Note how the ‘experimenter’ is taking notes.
Somewhat later, the person explains the dial pad…
…and then the central buttons.
The description above is only part of what the participant would say. In actual tests we did with people, almost all of them were quite confused about the cluster of buttons on the right, where at best they made wild (and usually wrong) guesses about what they did.
While we explained how the above process can be used to uncover someone’s initial mental model, it also works to uncover that person’s mental model after a period of use. This can reveal persistent misconceptions a person may have about the system, as well as holes in his or her knowledge.
Before reading on, try the above exercise with a test user. Have him walk through the entire visuals: it should take between 5-10 minutes. Take notes. From those notes, try to identify particular problems or classes of problems. Then redo the design to fix these problems. To make it ‘real’, try to keep to the constraints of the existing fax machine, where your new design could be done at little or no extra cost. Our solution is detailed below and on the next page.
6. Identifying Problems Through the Mental Model
Even the brief fragment above tells us a lot. While we saw that people quickly recognized the dial pad as a way to enter phone numbers, that is about it for the good news. Considerable problems are revealed with this sketch design, where people were unable to form an accurate mental model of it. For example, the meaning of the labels are cryptic, in part because people did not know the meaning of the abbreviations (e.g., ‘HQ’ actually stands for ‘High Quality’), and because it is not in their own language (e.g., what does ‘PRINTER INTERFACE’ really mean?). They were unsure about the primary buttons at the bottom that triggered dialing the number and sending (or cancelling) the fax. The central buttons are similarly cryptic. Even if people guessed at the correct function of some controls (such as the arrow buttons to save numbers), they could not say how they would actually go about using them (e.g., the actual sequence of operations necessary to save a number). And some controls are just plain mysterious, such as the cluster at the right.
7. Iterating the Design to Solve These Problems
Our next step is to redo the sketch to ‘fix’ some of these problems. Consider our solution on the next page. We redid the labeling: we removed all abbreviations, we rewrote them in the user’s language, and we added labels to the large button icons. We spatially and visually grouped related controls together (the phone directory, the printing boxes, the dial/send buttons and the pause/ cancel buttons). We put a plastic cover over the cryptic panel on the right, which visually tells the person that these controls are not needed for basic tasks but are available for advanced uses. Finally, we moved a few things unnecessary to the basic operation into this advanced control panel, e.g., the ‘hold’ buttons. The total cost of this improvement is minimal (the plastic cover adds costs, but removing some of the buttons saves costs). This is not to say that this design is perfect – it isn’t. However, it’s probably better than the previous version. And we can test it again – remember, our previous test only took about 10 minutes to do!
The sequence illustrated here is an example of the Narrative storyboard method introduced in Chapter 4.4.
It didn’t take long to do. We used PowerPoint to create the fax machine (Chapter 3.6), and photo-traces to capture three poses each of two people (Chapter 3.9).
We made our narrative visually dynamic simply by altering the cropping of the scene, and by changing the size and position of the photo-traced people within it.
We also horizontally flipped the photo-trace of a person. This let us reuse a pose while still making it look somewhat different across panels.
Repeat the above process with one of your own sketches, or an unfamiliar application, or even a web site (try a travel web site or an airline site). Do this more than once. What is important is that you get familiar with this process, and that you discover how many problems (and successes) you can uncover in a short amount of time by looking for mismatches between your participant’s initial mental model and your own model of what the system design was supposed to do.
You Now Know
Your visuals tell a story about your system, and people will form an initial mental model of how your system works from it. You can quickly uncover this model simply by asking them to explain the visuals. By looking for mismatches and gaps between their model and how the system is actually supposed to work, you can identify problems that you can repair in your next design iteration.