Bringing Design to Software
© Addison-Wesley, 19965 The Consumer Spectrum Paul Saffo
We are a frighteningly adaptable species. A good tool should adjust itself to the user, but good tools are scarce, and so we have learned to adapt ourselves to all but the most awkward of gizmos.... Of course, there are limits to what computers can expect their human companions to put up with. We do not use tools simply because they are friendly. We use tools to accomplish tasks, and we abandon tools when the effort required to make the tool deliver exceeds our threshold of indignation.
The software designer approaches a design task with a background of concerns and assumptions. She may, for example, want to make the software obvious and easy to use for novices. Or she may want it to be compatible with a well-known application, or maximally effective and efficient for the skilled user who is tackling a complex and specialized task. Much of the discussion in this book, and in the literature on human-computer interaction that preceded it, addresses one or another of these concerns.
There is a prior question, however, in designing a product for consumer use. What does the consumer care about? It is easy to let slogans such as user friendly and easy to learn blind the designer to the diversity of the software marketplace. Different people have different concerns. It takes a broader kind of perspective-that of an industry observer-to interpret the complex and often conflicting demands and forces in the market.
Paul Saffo began his career as a lawyer, specializing in technology law. He soon decided that the technology was more fun than the law, and spent his free moments as a software designer. Eventually, he gave up his law practice and joined the Institute for the Future, a consulting firm that tracks the long-term trends of technology. Saffo is particularly noted for his analysis of the current and future developments of cyberspace-the networked communication, computing, and information world that is rapidly expanding and is entering our everyday lives.
Saffo's perspective cuts across the usual boundaries between developer and user and between technology concerns and business concerns. He offers insights into how different aspects of software design are relevant in determining what will succeed in the software world. He argues that the traditional software-design concerns of functionality and ease of use are not always positive in a given product, but rather play a complex role in a larger tradeoff space of consumer concerns.
— Terry Winograd
§ § § Consumers are notoriously fickle when it comes to purchasing high-tech products. So much so, in fact, that computer and consumer-electronics industries attach great importance to user friendliness as a critical success factor. Yet even the most cursory glance at the last two decades of hits and failures suggests that this characteristic is only part of the whole user equation. Consumer history is littered with multiple instances of user-friendly products that were market duds, including RCA’s DiscoVision (an early laser-disk system), numerous home computers, and a host of early consumer-electronics gadgets, such as voice-controlled telephones, smart appliances, and pocket-sized photocopiers.
At the other extreme, there is no shortage of instances of complex products that have gone on to market success. The majority of consumers today have no clue how to use even a fraction of the features on their VCRs, or to navigate the maze of buttons on the average TV remote control. The most glaring exception is the personal computer itself. If ease of use is so important, then how did the first DOS-based PCs— notoriously cranky, unfriendly, and arbitrary devices—ever establish themselves in the market? If ease of use and user friendliness were all important, DOS would have been abandoned in favor of the Macintosh or some other operating system long before Windows was finally ready to ship.
DOS flourished because user friendliness encompasses only part of the usefulness equation, not only in computing, but also in the entire high-tech product sector. DOS PCs persisted in the face of easier-to-use alternatives because we are a frighteningly adaptable species.
A particularly dramatic example of our adaptability comes from the early years of Douglas Engelbart's research at the Stanford Research Institute (SRI). His team was testing multiple input devices, including a gizmo comprising a stylus at the end of a sliding rod that in turn rotated around a central pivot. Engelbart reminded one of his programmers to write a polar-to-xy coordinate conversion program before setting out this gizmo for other people to try. The programmer forgot, but set up the demo, which his teammates began to use on a walk-up basis. The omission was discovered by Engelbart a few days later when, walking past, he noticed a user staring at the screen, drawing an arrow-straight line by unwittingly pulling the stylus through a perfect arc!
A good tool should adjust itself to the user, but good tools are scarce, so we have learned to adapt ourselves to all but the most awkward of gizmos. Computers are especially ungainly devices, so manufacturers count on users to meet their incomplete inventions more than halfway. The happy captives are said to be computer literate—tame and tractable, and expert at making up for the manufacturers’ design failings.
Of course, there are limits to what computers can expect their human companions to put up with. We do not use tools simply because they are friendly. We use tools to accomplish tasks, and we abandon tools when the effort required to make the tool deliver exceeds our threshold of indignation—the maximal behavioral compromise that we are willing to make to get a task done.
The Threshold of Indignation
This measure of threshold of indignation completes the usefulness equation, capturing both the importance of the task that we are trying to complete and the user friendliness of the tool with which we are trying to complete it. In the case of the PCs in our offices, computers do not have to be friendly at all if the task is important enough and we have no alternative. Knowledge workers suffering from CRT eye strain and repetitive-motion syndrome understand intuitively that this extreme is the unfortunate norm in the business world. Our business culture forces on office workers relatively high thresholds of indignation. Businesses often have the luxury of being able to provide training that can do much to turn annoying techarcana into acceptable business tools.
In contrast, individual consumers have extremely low thresholds of indignation. Even the friendliest of information appliances will gather dust if the task bores us. The majority of VCR owners never venture beyond the play and record buttons on their machines, because the fancy features offer little benefit when weighed against the inconvenience of the fiddly command sequences. Similarly, the market for home computers before 1993 was confined largely to students and businesspeople overworking at home in the evenings—individuals who were bringing home their high business thresholds. And hobbyist early adopters have even higher thresholds yet!
The threshold-of-indignation measure also has more general application beyond information technologies. Experience with other devices can help us to understand the full range of potential consumer behavior. For example, cultural familiarity with a technology also influences the threshold. Automobiles are even more user hostile than are PCs. They are complex to maintain and hazardous to operate, regularly killing and maiming scores of innocent citizens. Yet automobiles seem user friendly, because we have spent a lifetime assimilating a host of automobile skills, including pumping gas, conversing with mechanics, learning the intricacies of the vehicle code, and understanding automotive-loan financing. Similarly, our telephone system has become steadily more complicated since divestiture, yet we have barely noticed, because our learning curve has kept pace, raising our collective threshold of indignation. We have all become operators, internalizing the complexities of direct dialing, calling cards, and access codes.
As this latter example suggests, thresholds of indignation can change over time. A 1920s time traveler would be lost in the touch-tone complexities of our 1990s telephone system. Yet, if we traded places, we would be equally confused and infuriated by endless and time-consuming transactions with human operators.
Also, generational responses to technologies differ markedly, with different-age cohorts exhibiting differing levels of comfort with distinct systems. Baby boomers prefer the efficiency and anonymity of bank automatic-teller machines, whereas their parents avoid these banking robots to conduct their transactions through a human teller. More recently, younger users have taken to cyberspace environments such as world wide web, while their elders are left watching from the sidelines. Eventually, some portion of the laggard elders will follow, but only after the younger generation has blazed the trail.
Reinventing the function of the device can alter thresholds of indignation. PCs became a consumer hit when CD-ROMs reached critical mass, and faster modems combined with a growing Internet to deliver applications that are more exciting than mere word processing and spreadsheets. In short, the PC was reinvented from an original role of standalone processing tool to a new role of access window on a larger information world.
The Universe of Tools and Users
One additional dimension completes the consumer-adoption equation: ability and willingness to pay. As depicted in Figure 5.1, we can model the entire consumer universe with willingness and ability to pay on one axis, and threshold of indignation along the other. The full range of consumer responses and the products that consumers are willing to purchase are indicated in each of the four corners, described next:
Figure 5.1 The Consumer Universe The range of consumer behavior in purchasing products can be bounded by two dimensions: threshold of indignation, and willingness to pay. Each of the extremes in the four corners has its natural purchasers and archetypal products.
• The Air Force general—infinite wealth, matched only by an infinitely high threshold of indignation. Imagine a device that is infinitely expensive and infinitely complex, and you have described the B-2 Stealth bomber, with all the latest in high-tech trickery, pound-for-pound costing more than its weight in gold. There is only one consumer with both the cash and the requisitely high threshold of indignation to own one of these. Even NASA cannot match the buyer or the product with its nearly as complex high-tech gizmo, the Space Shuttle.
• The teenage hacker—a general’s love of complexity, but no cash. Now imagine a device that is as impenetrably complex and confusing as a B-2, but is accessible to someone who has no money at all. Until the early 1990s, the Internet was the quintessential technology for this quadrant by virtue of the nerdy complexity of the access tools of the time. Back then, the Internet was so confusing that only two kinds of individuals were willing to commit the time to understand its intricacies—computer scientists and teenage hackers. For the latter, the Internet's sheer mind-numbing complexity made the environment an irresistible electronic playground accessible with a small investment in a PC and modem. Even though new tools have made the Internet ever more accessible, cyberspace remains a license for teenage obsession. Teenagers explore its nether reaches even as other consumers await the Holy Grail of the "Internet in a box."
• The Hollywood starlet—lots of cash and utterly no patience. Who has endless wealth, but no patience with even the simplest of devices? Zsa Zsa Gabor’s on-screen persona nicely matches this intersection, so we generalize this quadrant as that of the Hollywood starlet. The perfect consumer toy for such bored rich people is, of course, a Rolls Royce. And if it is too complex to drive, then so much the better. Just throw in more cash for a chauffeur, at once simplifying the interaction with the device and adding to the price! High-end home theater systems today are the Rolls Royces of consumer electronics.
• The average consumer—limited cash and limited patience. Finally, there are the rest of us, the average consumers with limited cash and limited patience. Consumers love novelty as much as do the other three, but a combination of cash and time constraints limits our purchasing flights of fancy. If a product delivers benefits (either functional or entertainment) without too much work and does not cost too much, we will make it an integral part of our lives. The telephone is the model product for this quadrant: it is cheap (as low as $10); is easy to use (just plug it in); and delivers real, immediately understandable benefits. The challenge for consumer manufacturers is to tame their products and to drive down costs to the point at which their products sell into this corner.
A Snapshot of High-Tech Products Today
By filling in the matrix with an assortment of products sold today, we can quickly get a sense of where one zone makes a transition into another. Figure 5.2 shows such a high-tech spread for the world of the consumer (left half of diagram) and business (right half of diagram).
Figure 5.2 The High-Tech Spread Products can be arrayed within the matrix, revealing a continuum of information devices suitable to a variety of populations. A distinct—but permeable—line in the center divides consumer-oriented devices from business devices.
On the business side, PCs are the simplest and cheapest of the devices depicted, and are the ones most easily understood by the individual worker. Workstations are slightly more expensive and complex, because they must be connected into a network managed by system professionals. Mainframes are more remote and complex yet, whereas cyberspace until recently (when new tools made it radically more accessible) was all but invisible to everyone except the anointed few. In fact, we can describe each of these products along purely functional lines that have nothing to do with the technology. A PC is a device that an individual user can turn on and off, to which she can add and remove software at will. With a workstation, the user can control the on–off switch, but must ask someone else’s permission before messing with the software. And a mainframe is a device that she can approach only through bureaucracy.
The same principles apply on the consumer side. A telephone is solidly in the consumer quadrant by virtue of its low cost, clear benefits, and ease of use. TVs are a bit farther out on both axes: They are more costly and difficult to use, because we must hook them to AC power and an antenna or cable, and then navigate through an ever-growing number of channels. VCRs fall yet farther out. Although VCR prices are about the same as TV prices, a VCR's complexity falls well beyond the few features that most viewers actually use. Component stereo systems fall at the edge of the consumer quadrant. Not only are the pieces expensive, but also setup involves puzzling out a tangle of wires, setting custom preferences, and then learning how to use obscure and confusing remote controls. Fortunately, many consumers live in families that typically include individuals farther along one axis than the other members. As a team, they can make product desires a reality. Mom and Dad do the purchasing, and their more technology-tolerant children install the stereo or perform the intricacies of VCR programming.
The model in Figure 5.2 can help to explain consumer trends, and, in specific instances, can even predict a product’s chances of success or failure. Note a few of its specific implications:
• High-tech products tend to move downward and left over time. It is axiomatic that the cost of high-tech components drops rapidly over time, and the performance of new components increases. For example, Moore’s law (named after Intel cofounder Gordon Moore) states that the cost of a given unit of processing power on a chip decreases by a factor of 10 every 2.5 years, whereas the number of circuits that can fit in a given space (an indirect measure of performance) increases at the same rate. The dropping cost factor means that high-tech products will move downward on the map over time. The rising performance factor tends to push products left along the indignation curve, because it takes extra power to make devices user friendly.
• Consumer thresholds tend to rise over time. As mentioned earlier, consumers become accustomed to particular technologies, and their threshold of indignation rises. Newfangled telephones once frustrated our grandparents; today’s more complex systems rarely annoy us. In short, we evolve with our technologies, our rising sophistication matching the new capabilities of technology. Generational factors play a role here—children raised with Ataris and Nintendos have little difficulty grokking the complexities of Windows-based PCs and Internet gateway services. This factor adds some force to the leftward motion of products on the map—time and rising sophistication alone will make devices more user friendly.
• Product manufacturers tend to be tuned to particular zones on the map—and moving is remarkably difficult. We can also place industries on this map. For example, the consumer-electronics industry is focused on the lower-left (consumer) quadrant, whereas the computer industry centers on the upper-right quadrant. Each industry—and often each segment within each industry—is finely tuned to the specific threshold of indignation of its core customers, and to the price point of its mainline products. This orientation makes it extremely difficult for companies to move out of their habitual areas, because doing so requires acquiring new understandings about unfamiliar levels of indignation thresholds, and reorienting entire manufacturing and sales processes to a different price granularity of product.
In this latter regard, companies are like villages that hunt different kinds of game for their food. One village might hunt mammoths, depending on one kill every 6 months. Another might depend on weekly deer hunts; a third might be inhabited by rabbit eaters, who must come home with game every day to survive. In high-tech industry today, aerospace firms are mammoth hunters, personal-computer companies are deer hunters, and consumer-electronic firms are rabbit hunters. Just as mammoth hunters might find rabbit hunting demeaning and exhausting, aerospace companies are unlikely to invent the consumer-electronic successor to the VCR, even though they pioneered many of the components that VCR makers incorporate into the products.
• It is easier for companies to move right or up than down or left. All things being equal, it is easier to move into customer populations that have higher indignation thresholds than to move in the opposite direction. Selling more expensive products, although difficult, is far easier than selling cheaper products. IBM's woe-filled attempt to dominate the PC market is a case in point.
Moving down or left is like trying to swim to the bottom of a pool while wearing a life vest. By paddling furiously, you can get part way down; eventually, however, you will bob back to the surface. Similarly, a few highly aggressive companies can move down one level, but they rarely get farther. IBM’s move into PCs in the 1980s represented a heroic—but ultimately futile—attempt. Although IBM made "big iron"—mainframes and the like—the infant PC sector was a close neighbor that the company was able to greet by virtue of good executive decisions and visionary managers. But IBM’s next step down—into home PCs, with the PC Jr—was a singular disaster because the consumer-level prices and threshold of indignation were too alien to the men in gray suits.
• Adaptive interfaces are an important opportunity. User friendliness is profoundly relative: What is friendly and helpful to a novice user may be highly intrusive to a seasoned expert. Given that users progress through stages of mastery, it is only logical that interfaces should be capable of identifying a user's skill level, matching to it, and then systematically revealing more capabilities in a way that encourages the user to learn. Of course, this matching of interface to levels of skill is even more important outside the consumer environment, where professional effectiveness depends on having the right tool for the job, and moreover a tool matched to its user's evolving skills. As the Japanese craftsperson Soestu Yanagi (1972) once observed, "The craftsman is most free when his tools are proportionate to his needs."
• Ads and hype are important influencers of consumer thresholds of indignation. "It's new and I don't like it." The sentiment was that of a change-weary 6 year old, and it echoes the sentiment of most of us at one time or another. Advertising and hype—overenthusiastic, often credulous enthusiasm for a new product—can be a potent weapon in moving us all into a frame of mind open to new ideas. Far from being an impediment, this sort of excitement is an essential feature in the technology-diffusion process.
• The most important product opportunities are the gaps immediately adjoining a company’s core markets. New companies can look anywhere on the matrix that seems to fit their capabilities, but established companies should seek gaps close to home—a pattern reflected by Apple’s and IBM’s respective entries into personal computing. Success is most likely when a company is following already-familiar customers, or is leveraging already-understood technology. Moving into an area where both factors are unfamiliar amounts to playing corporate Russian roulette. The biggest gap in the high-tech sector today lies in the center of the matrix, as depicted in Figure 5.3. That is where the new media industries in the making are appearing.
Figure 5.3 Where Industries Fit In The gap between the zones of the computer and consumer-electronics industries is becoming ground zero for competition in the 1990s. Consumer-electronic players are trying to grow up into it, while computer players, such as Apple and Microsoft, are trying to reach down.
When we view this matrix, it becomes apparent that new media are not merely an extension to personal computing, but rather provide the obvious collision point between upward- and rightward-moving consumer-electronics companies and downward- and leftward-moving computer manufacturers. If the model has any predictive value, it is a sure bet that the consumer-electronics players are the most likely to grab most of the action in the long run, although the ultimate winners will have to understand and master the needs from both sides of the gap if they are to succeed. The players who find creative ways to team up across this gap are likely to dominate these new industries in the decade ahead.
Suggested Readings
- Ivan Illich. Tools for Conviviality. New York: Harper & Row, 1973.
- Jun'ichiro Tanizaki. In Praise of Shadows. New Haven: Leete's Island Books, 1977
- Soetsu Yanagi. The Unknown Craftsman: A Japanese Insight into Beauty. Tokyo and New York: Kodansha, 1972, 1989.
About the Author
Paul Saffo is a director of the Institute for the Future in Menlo Park, California. He writes about the social and economic aspects of emerging digital technologies for numerous magazines and newspapers. A book of his essays, Dreams in Silicon Valley, has been published in Japan.
