I am not your typical appliance user. I have a doctorate in human factors and I am a Certified Professional Ergonomist. But, more importantly, I grew up in parallel with the emergence of consumer digital technology. I had an Apple II computer in the early 1980s and a Radio Shack TRS-80 even before that. So you might think that I’d be ahead of the curve when it comes to figuring out how to use complex appliances and products. In fact, I’m not. Sure, I am adept at setting an oven clock or storing a number in a cell phone, but it still requires a decent amount of effort and learning for each new product. And I certainly can’t assemble anything out of the box very well.

As a practitioner, I’ll be the first to admit that designing usable products is not easy, particularly in the consumer appliance market where the range of user characteristics such as physical capabilities, usage patterns, and technical skills is so broad and variable. There are also internal challenges to creating an easy-to-use product. Product design teams that are inexperienced with user-centered design methods may be intimidated by the process and unsure where to start to holistically improve a design. At the opposite end of the spectrum, organizations with in-place user-research processes may be overloaded with interpreting qualitative and quantitative data to drive actionable design decisions.

Most importantly though, there are misconceptions about the relationship between product complexity and usability. There’s an implicit assumption among many in the appliance design world that, all other things being equal, a product with more features will be a product that is more difficult to use. Some have portrayed technology as the enemy of usability, as innovations in technology drives the growth in features and complexity.

Automaker BMW recently redesigned the iDrive controls to better accommodate interaction in the context of driving. (Image: BMW.)

In reality, we have seen some products that utilize technology to improve usability, such as the Apple iPhone, and others where technology gets in the way of usability, such as BMW’s original in-car iDrive system (currently going through a re-launch with significant design improvements). But, the difference between the iPhone and the iDrive is not one of complexity, at least in a technical or functional sense of the word. If you calculate the number of options and actions supported by both the iPhone and its similarly named automotive foil, they are roughly equivalent in their complexity. The usability differentiator between comparably featured products is not simplicity versus complexity, but clarity, the users ease of perceiving and interacting with the provided functions.

The iPhone delivers clarity in a number of ways, most notably via direct manipulation of the interface — touch to open, pinch to reduce size, and so forth. The iDrive failed to be effectively clear, in part because it requires significant coordination between spatially separated tactile (controller) and visual (display) interaction points — all while competing with a driver’s already taxed visual and tactile resources. In other words, it’s not the quantity of technology that the user interacts with, but the quality of its implementation that differentiates a product that is quickly understood and adopted from one that is returned because it is perceived to be “broken” or requires special training.

So how does one achieve clarity in products that deliver an increasing number of modes, options, and configurations? A helpful way of thinking about the problem is to consider clarity as a measure of the complexity of the user experience, rather than the product functionality. In some cases, making the operation of a device clearer to the user is achieved by transferring the experiential complexity to technical complexity, rather than eliminating complexity from the user-product relationship.

For example, an automatic transmission is easier to use than a manual transmission from the driver’s perspective. The ease of use is improved by relocating the complexity of the gear-shifting task from the user to an automated system. But from a technical perspective, the automatic transmission is significantly more complex than the manual version. Accommodating both modes of interaction adds even greater technical complexity, as automobile manufacturers are now providing transmissions than can work in either fully automatic or a clutch-less manual shifting mode. On the other hand, some recent products have reduced experiential complexity by avoiding technical complexity. For example, the Flip Video Mino surprised the competitive consumer video camera market with its lack of features, reducing the number of interaction touch points.

The Flip Video Mino took the unconventional approach of reducing features to improve ease-of-use. (Image: Pure Digital Technologies.)

In practice, designers do not always have the option (or the desire) to reduce product features, and need techniques to support usable products. Therefore, it is valuable to address the pros and cons in the design of each user touch point from the perspective of interaction design principles. For example, a common design decision is whether interface control points should be dedicated (single function) or modal (multiple functions). Dedicated controls are easier to label and learn, but can result in a larger number of options that can confuse or intimidate the consumer, and can add clutter to a minimal aesthetic. Adding touch points can also require more actuator components that leads to higher costs.

By contrast, modal controls provide access to multiple functions through the same touch points. For instance, many clock radios have buttons that change functionality based on the mode of use. The plus and minus buttons may be used for tuning frequencies in the radio mode, and also to change the hours and minutes in the time-setting mode. As a rule of thumb, the modal/multiple-function user interface is more complex to learn and use, but is widely applied in situations where secondary tasks are done relatively infrequently, such as changing the time of day setting on an alarm clock.

Heuristics like frequency of use are essential for effective interface design, but they can only take one so far, as new functions and new interface capabilities arise. So how does one determine the appropriate design of user touch points in order to create a clear user experience? A more appropriate question would be, how does one know which design, among several approaches, would be clearest to the majority of users? In practice, achieving the best user experience typically requires developing several user interface concept designs or prototypes of equal functional complexity by considering various control arrangements or interactions. (Note, this assumes that expertise, or at least guidance, in interaction design was applied in the formulation of the various concepts.)

Grundig Music Boy radio illustrates how interaction design principles can guide decisions about control selection. In this case, the plus and minus buttons are modal controls for both tuning and time settings. (Image: Grundig Intermedia.)

There are two basic approaches to evaluating and comparing the clarity of alternative user interface approaches — modeling and testing. Modeling uses a set of criteria to objectively quantify the characteristics of a hypothetical user experience with the product interface. In cognitive modeling, a set of discrete actions are defined that represent how a user will interact with the interface — viewing a displayed list of options, making a decision, and reaching and pressing the preferred option. Each of these behaviors is assigned a time-based value. For example, a user experience might consist of the following sequence:

1. Reach and press a button: 1 second.
2. Wait for system to update display: 0.25 seconds.
3. Review a list of five displayed choices: 5.5 seconds.
4. Make a decision among five choices: 1.5 seconds.
5. Reach, grasp, and rotate a knob through 180 degrees: 3 seconds.

Note that these particular times are hypothetical values for explanatory purposes. Actual time values would be determined through guidelines and measurements, or even better, by averaging observed user performance with similar systems. Values can vary widely based on the factors such as the user’s level of experience with such an interface, so models need to be appropriately qualified.

With these behavioral times, one can sum the discrete steps a user executes when completing particular tasks, and compare the total time across different interfaces. For example, completing a particular task with one version of an interface might require the following sequence of steps: c, d, a (8 seconds). An alternative might require a different set of actions: a, e, b, c, d, a (12.25 seconds). In this example, the first interface was less complex with respect to both task time and numbers of steps. (There are specialized software applications that assist with this type of interface modeling, such as CogTool from Carnegie Mellon University.)

There are some obvious trade-offs to the modeling approach. On the positive side, it provides a consistent method for quantifying and comparing the complexity of two interfaces with respect to the time required to complete tasks. The approach is helpful for quickly evaluating concepts or prototypes against legacy and competitor products, particularly for complex, multi-step interactions, without involving end-users. Its primary limitations are that it oversimplifies the situation and does not accommodate any of the variability inherent in human experience.

Consider also that one interface may require a user to take more time and steps than another, but still be perceived as more usable due to other factors such as the tone or visual presentation of the information. Examples include recently developed gesture-based interfaces that employ intuitive user interactions, such as those found on the iPhone. It may take longer to scroll through an alphabetical list of options with your fingers compared to typing a keypad shortcut, but the ease of interaction and mild learning curve impact the perceived ease of use, as opposed to simply a time-based metric.

Fortunately, usability testing provides a much more robust and valid way for determining the complexity of the user experience. During usability testing, representative end-users conduct tasks using existing products, prototypes, or concepts. The level of realism in a prototype depends on the tasks and measures being considered. For example, choosing the correct item from a printed list of menu choices is largely comparable to doing the same on an electronic display if the focus is on the labeling and sequence of the choices. Using a paper list would not, however, be a good substitute for evaluating the readability of an electronic display in a low-light environment.

In usability testing, typical metrics include task success or failure rate, the time to complete a task, and the types and frequencies of errors. In addition, qualitative participant ratings and feedback, along with researcher observations, provide depth and meaning to better understand the relative strengths and weaknesses of each alternative design. I recommend reviewing the recently published international standard on the Ease of Operation of Everyday Products (ISO 20282), which provides guidance and metrics for the usability testing of appliances.

In summary, employing user-centered design in the era of increasingly complicated products requires effective complexity management. Complexity should be considered in its various forms as it relates to the product development process and the end-user’s experience. Designers should:

• Weigh the relative merits of maintaining or adding functions, particularly those that are rarely used by customers, against reducing complexity at the start.
• Follow interaction design guidelines and practices to address clarity across the product’s user touch points.
• Consider cognitive modeling as a preliminary tool for evaluating differences and opportunities across internal designs and competitive products, especially for lengthy workflows.
• Conduct iterative design and usability testing to determine which interface approach provides the greatest clarity in objective and subjective terms.

Admittedly, this will add effort and complexity to the design process, but will provide the data needed to make decisions that lead to usable products.

For more information, email: rtannen@bresslergroup.com