by Don Norman
For everyday things, conceptual models need not be very complex. After all, scissors, pens, and light switches are pretty simple devices. There is no need to understand the underlying physics or chemistry of each device we own, just the relationship between the controls and the outcomes. When the model presented to us is inadequate or wrong (or, worse, nonexistent), we can have difficulties. Let me tell you about my refrigerator.
FIGURE 1.9.Refrigerator Controls. Two compartments— fresh food and freezer—and two controls (in the fresh food unit). Your task: Suppose the freezer is too cold, the fresh food section just right. How would you adjust the controls so as to make the freezer warmer and keep the fresh food the same? (Photograph by the author.)
I used to own an ordinary, two-compartment refrigerator—nothing very fancy about it. The problem was that I couldn’t set the temperature properly. There were only two things to do: adjust the temperature of the freezer compartment and adjust the temperature of the fresh food compartment. And there were two controls, one labeled “freezer,” the other “refrigerator.” What’s the problem?
Oh, perhaps I’d better warn you. The two controls are not independent. The freezer control also affects the fresh food temperature, and the fresh food control also affects the freezer. Moreover, the manual warns that one should “always allow twenty-four (24) hours for the temperature to stabilize whether setting the controls for the first time or making an adjustment.”
FIGURE 1.10.Two Conceptual Models for a Refrigerator. The conceptual model A is provided by the system image of the refrigerator as gleaned from the controls. Each control determines the temperature of the named part of the refrigerator. This means that each compartment has its own temperature sensor and cooling unit. This is wrong. The correct conceptual model is shown in B. There is no way of knowing where the temperature sensor is located so it is shown outside the refrigerator. The freezer control determines the freezer temperature (so is this where the sensor is located?). The refrigerator control determines how much of the cold air goes to the freezer and how much to the refrigerator.
It was extremely difficult to regulate the temperature of my old refrigerator. Why? Because the controls suggest a false conceptual model. Two compartments, two controls, which implies that each control is responsible for the temperature of the compartment that carries its name: this conceptual model is shown in Figure 1.10A. It is wrong. In fact, there is only one thermostat and only one cooling mechanism. One control adjusts the thermostat setting, the other the relative proportion of cold air sent to each of the two compartments of the refrigerator. This is why the two controls interact: this conceptual model is shown in Figure 1.10B. In addition, there must be a temperature sensor, but there is no way of knowing where it is located. With the conceptual model suggested by the controls, adjusting the temperatures is almost impossible and always frustrating. Given the correct model, life would be much easier.
Why did the manufacturer suggest the wrong conceptual model? We will never know. In the twenty-five years since the publication of the first edition of this book, I have had many letters from people thanking me for explaining their confusing refrigerator, but never any communication from the manufacturer (General Electric). Perhaps the designers thought the correct model was too complex, that the model they were giving was easier to understand. But with the wrong conceptual model, it was impossible to set the controls. And even though I am convinced I knew the correct model, I still couldn’t accurately adjust the temperatures because the refrigerator design made it impossible to discover which control was for the temperature sensor, which for the relative proportion of cold air, and in which compartment the sensor was located. The lack of immediate feedback for the actions did not help: it took twenty-four hours to see whether the new setting was appropriate. I shouldn’t have to keep a laboratory notebook and do controlled experiments just to set the temperature of my refrigerator.
I am happy to say that I no longer own that refrigerator. Instead I have one that has two separate controls, one in the fresh food compartment, one in the freezer compartment. Each control is nicely calibrated in degrees and labeled with the name of the compartment it controls. The two compartments are independent: setting the temperature in one has no effect on the temperature in the other. This solution, although ideal, does cost more. But far less expensive solutions are possible. With today’s inexpensive sensors and motors, it should be possible to have a single cooling unit with a motor-controlled valve controlling the relative proportion of cold air diverted to each compartment. A simple, inexpensive computer chip could regulate the cooling unit and valve position so that the temperatures in the two compartments match their targets. A bit more work for the engineering design team? Yes, but the results would be worth it. Alas, General Electric is still selling refrigerators with the very same controls and mechanisms that cause so much confusion. The photograph in Figure 1.9 is from a contemporary refrigerator, photographed in a store while preparing this book.
The System Image
People create mental models of themselves, others, the environment, and the things with which they interact. These are conceptual models formed through experience, training, and instruction. These models serve as guides to help achieve our goals and in understanding the world.
How do we form an appropriate conceptual model for the devices we interact with? We cannot talk to the designer, so we rely upon whatever information is available to us: what the device looks like, what we know from using similar things in the past, what was told to us in the sales literature, by salespeople and advertisements, by articles we may have read, by the product website and instruction manuals. I call the combined information available to us the system image. When the system image is incoherent or inappropriate, as in the case of the refrigerator, then the user cannot easily use the device. If it is incomplete or contradictory, there will be trouble.
As illustrated in Figure 1.11, the designer of the product and the person using the product form somewhat disconnected vertices of a triangle. The designer’s conceptual model is the designer’s conception of the product, occupying one vertex of the triangle. The product itself is no longer with the designer, so it is isolated as a second vertex, perhaps sitting on the user’s kitchen counter. The system image is what can be perceived from the physical structure that has been built (including documentation, instructions, signifiers, and any information available from websites and help lines). The user’s conceptual model comes from the system image, through interaction with the product, reading, searching for online information, and from whatever manuals are provided. The designer expects the user’s model to be identical to the design model, but because designers cannot communicate directly with users, the entire burden of communication is on the system image.
FIGURE 1.11.The Designer’s Model, the User’s Model, and the System Image. The designer’s conceptual model is the designer’s conception of the look, feel, and operation of a product. The system image is what can be derived from the physical structure that has been built (including documentation). The user’s mental model is developed through interaction with the product and the system image. Designers expect the user’s model to be identical to their own, but because they cannot communicate directly with the user, the burden of communication is with the system image.
Figure 1.11 indicates why communication is such an important aspect of good design. No matter how brilliant the product, if people cannot use it, it will receive poor reviews. It is up to the designer to provide the appropriate information to make the product understandable and usable. Most important is the provision of a good conceptual model that guides the user when thing go wrong. With a good conceptual model, people can figure out what has happened and correct the things that went wrong. Without a good model, they struggle, often making matters worse.
Good conceptual models are the key to understandable, enjoyable products: good communication is the key to good conceptual models.
The Paradox of Te
chnology
Technology offers the potential to make life easier and more enjoyable; each new technology provides increased benefits. At the same time, added complexities increase our difficulty and frustration with technology. The design problem posed by technological advances is enormous. Consider the wristwatch. A few decades ago, watches were simple. All you had to do was set the time and keep the watch wound. The standard control was the stem: a knob at the side of the watch. Turning the knob would wind the spring that provided power to the watch movement. Pulling out the knob and turning it rotated the hands. The operations were easy to learn and easy to do. There was a reasonable relationship between the turning of the knob and the resulting turning of the hands. The design even took into account human error. In its normal position, turning the stem wound the mainspring of the clock. The stem had to be pulled before it would engage the gears for setting the time. Accidental turns of the stem did no harm.
Watches in olden times were expensive instruments, manufactured by hand. They were sold in jewelry stores. Over time, with the introduction of digital technology, the cost of watches decreased rapidly, while their accuracy and reliability increased. Watches became tools, available in a wide variety of styles and shapes and with an ever-increasing number of functions. Watches were sold everywhere, from local shops to sporting goods stores to electronic stores. Moreover, accurate clocks were incorporated in many appliances, from phones to musical keyboards: many people no longer felt the need to wear a watch. Watches became inexpensive enough that the average person could own multiple watches. They became fashion accessories, where one changed the watch with each change in activity and each change of clothes.
In the modern digital watch, instead of winding the spring, we change the battery, or in the case of a solar-powered watch, ensure that it gets its weekly dose of light. The technology has allowed more functions: the watch can give the day of the week, the month, and the year; it can act as a stopwatch (which itself has several functions), a countdown timer, and an alarm clock (or two); it has the ability to show the time for different time zones; it can act as a counter and even as a calculator. My watch, shown in Figure 1.8, has many functions. It even has a radio receiver to allow it to set its time with official time stations around the world. Even so, it is far less complex than many that are available. Some watches have built-in compasses and barometers, accelerometers, and temperature gauges. Some have GPS and Internet receivers so they can display the weather and news, e-mail messages, and the latest from social networks. Some have built-in cameras. Some work with buttons, knobs, motion, or speech. Some detect gestures. The watch is no longer just an instrument for telling time: it has become a platform for enhancing multiple activities and lifestyles.
The added functions cause problems: How can all these functions fit into a small, wearable size? There are no easy answers. Many people have solved the problem by not using a watch. They use their phone instead. A cell phone performs all the functions much better than the tiny watch, while also displaying the time.
Now imagine a future where instead of the phone replacing the watch, the two will merge, perhaps worn on the wrist, perhaps on the head like glasses, complete with display screen. The phone, watch, and components of a computer will all form one unit. We will have flexible displays that show only a tiny amount of information in their normal state, but that can unroll to considerable size. Projectors will be so small and light that they can be built into watches or phones (or perhaps rings and other jewelry), projecting their images onto any convenient surface. Or perhaps our devices won’t have displays, but will quietly whisper the results into our ears, or simply use whatever display happens to be available: the display in the seatback of cars or airplanes, hotel room televisions, whatever is nearby. The devices will be able to do many useful things, but I fear they will also frustrate: so many things to control, so little space for controls or signifiers. The obvious solution is to use exotic gestures or spoken commands, but how will we learn, and then remember, them? As I discuss later, the best solution is for there to be agreed upon standards, so we need learn the controls only once. But as I also discuss, agreeing upon these is a complex process, with many competing forces hindering rapid resolution. We will see.
The same technology that simplifies life by providing more functions in each device also complicates life by making the device harder to learn, harder to use. This is the paradox of technology and the challenge for the designer.
The Design Challenge
Design requires the cooperative efforts of multiple disciplines. The number of different disciplines required to produce a successful product is staggering. Great design requires great designers, but that isn’t enough: it also requires great management, because the hardest part of producing a product is coordinating all the many, separate disciplines, each with different goals and priorities. Each discipline has a different perspective of the relative importance of the many factors that make up a product. One discipline argues that it must be usable and understandable, another that it must be attractive, yet another that it has to be affordable. Moreover, the device has to be reliable, be able to be manufactured and serviced. It must be distinguishable from competing products and superior in critical dimensions such as price, reliability, appearance, and the functions it provides. Finally, people have to actually purchase it. It doesn’t matter how good a product is if, in the end, nobody uses it.
Quite often each discipline believes its distinct contribution to be most important: “Price,” argues the marketing representative, “price plus these features.” “Reliable,” insist the engineers. “We have to be able to manufacture it in our existing plants,” say the manufacturing representatives. “We keep getting service calls,” say the support people; “we need to solve those problems in the design.” “You can’t put all that together and still have a reasonable product,” says the design team. Who is right? Everyone is right. The successful product has to satisfy all these requirements.
The hard part is to convince people to understand the viewpoints of the others, to abandon their disciplinary viewpoint and to think of the design from the viewpoints of the person who buys the product and those who use it, often different people. The viewpoint of the business is also important, because it does not matter how wonderful the product is if not enough people buy it. If a product does not sell, the company must often stop producing it, even if it is a great product. Few companies can sustain the huge cost of keeping an unprofitable product alive long enough for its sales to reach profitability—with new products, this period is usually measured in years, and sometimes, as with the adoption of high-definition television, decades.
Designing well is not easy. The manufacturer wants something that can be produced economically. The store wants something that will be attractive to its customers. The purchaser has several demands. In the store, the purchaser focuses on price and appearance, and perhaps on prestige value. At home, the same person will pay more attention to functionality and usability. The repair service cares about maintainability: how easy is the device to take apart, diagnose, and service? The needs of those concerned are different and often conflict. Nonetheless, if the design team has representatives from all the constituencies present at the same time, it is often possible to reach satisfactory solutions for all the needs. It is when the disciplines operate independently of one another that major clashes and deficiencies occur. The challenge is to use the principles of human-centered design to produce positive results, products that enhance lives and add to our pleasure and enjoyment. The goal is to produce a great product, one that is successful, and that customers love. It can be done.
CHAPTER TWO
THE PSYCHOLOGY OF EVERYDAY ACTIONS
During my family’s stay in England, we rented a furnished house while the owners were away. One day, our landlady returned to the house to get some personal papers. She walked over to the old, metal filing cabinet and attempted to open the top drawer. It wouldn’t open. S
he pushed it forward and backward, right and left, up and down, without success. I offered to help. I wiggled the drawer. Then I twisted the front panel, pushed down hard, and banged the front with the palm of one hand. The cabinet drawer slid open. “Oh,” she said, “I’m sorry. I am so bad at mechanical things.” No, she had it backward. It is the mechanical thing that should be apologizing, perhaps saying, “I’m sorry. I am so bad with people.”
My landlady had two problems. First, although she had a clear goal (retrieve some personal papers) and even a plan for achieving that goal (open the top drawer of the filing cabinet, where those papers are kept), once that plan failed, she had no idea of what to do. But she also had a second problem: she thought the problem lay in her own lack of ability: she blamed herself, falsely.
How was I able to help? First, I refused to accept the false accusation that it was the fault of the landlady: to me, it was clearly a fault in the mechanics of the old filing cabinet that prevented the drawer from opening. Second, I had a conceptual model of how the cabinet worked, with an internal mechanism that held the door shut in normal usage, and the belief that the drawer mechanism was probably out of alignment. This conceptual model gave me a plan: wiggle the drawer. That failed. That caused me to modify my plan: wiggling may have been appropriate but not forceful enough, so I resorted to brute force to try to twist the cabinet back into its proper alignment. This felt good to me—the cabinet drawer moved slightly—but it still didn’t open. So I resorted to the most powerful tool employed by experts the world around—I banged on the cabinet. And yes, it opened. In my mind, I decided (without any evidence) that my hit had jarred the mechanism sufficiently to allow the drawer to open.
