Imagine you are standing on a bustling street corner waiting for the pedestrian crossing signal. A rushing commuter pushes by, navigating into the crosswalk. Maybe he's late for work. He watches for cars while crossing the street. Now, imagine this commuter has a mobile phone in his pocket, buzzing for attention. Will he wait until he exits the crosswalk before checking to see who texted? If he thought the message was important, he might not. Why? What compels him to look at his smartphone while also looking for oncoming traffic?
As mobile device usage continues to grow, most of us have experienced similar scenarios: Users frequently interact with mobile devices while doing other things, dividing their attention between the real world and the mediated world. We may infer that these users are highly engaged within a mobile experience or unfulfilled by their real-world experience, or maybe they're addicted to their social network or their texting app. But what if their frequent usage was actually instinctual, compelled by a form of vigilant attention toward their mobile devices? Could smartphones—with their nagging, beeping, and omnipresence—be creating unprecedented vigilance scenarios? If so, how prevalent are these scenarios and how do they affect designing for mobile use?
Vigilance is a state of watchfulness, the compulsion for which maps directly to the perceived consequence of missing out on possible observations. Vigilant user behavior, while not widely discussed, is uniquely taxing because it consumes a user's attentional resources by monitoring for signals to which they need to react . Research into vigilance in HCI—which started with an examination of the performance of WWII radar operators—has classically focused on situated usage with specific tasks that involve sustained visual search for target signals, in order to understand the physiological impact of a vigilant state when interacting with machines [1, 2, 3]. This focus implies that vigilant scenarios exist only within a small subset of human-computer interactions: sustained, situated tasks performed by trained professionals or operators.
But really, vigilance is much broader. The instinctual basis for entering into a vigilant state is the preservation of self or society . This extreme watchfulness can involve multiple senses and can coincide with other activities. For example, a grazing animal will instinctively balance two tasks necessary for survival: intermittently eating and watching vigilantly for approaching predators. What if our commuter was late to a meeting and his boss had been sending angry texts? Perhaps he was intermittently watching for oncoming traffic and watching for those texts to keep his job. Noticing both of those things may be critical to his sense of self-preservation.
The true measure of vigilance is not that usage is sustained, but that attention is sustained. In our interactions with computers, just as with our interactions in the physical world, vigilance is not necessarily limited to specific and situated tasks—everyday mobile users may sustain vigilant attention over multiple intermittent sessions of use.  This everyday vigilance is actually quite distinct from the classic situational form in that it sustains indefinitely, inter-session (Figure 1). Our commuter is an extreme everyday example, but let's consider something more common: teenage social networking behaviors.
Vigilance requires a paradigm shift. If you are an interaction design practitioner, then sooner or later these principles will conflict with the business objectives of your employer.
For many teens, social network usage is associated with self-preservation because their interactions within that system are a basis for socialization and, partially, the formation and preservation of their identity, their sense of self. In developing these social interactions, a teenager must break her natural methods for real-time communication into the discrete unidirectional communiqués that these technologies enable: things like messages, status updates, and photo uploads. These communiqués are signals that the teen uses for social encoding, impression management, and identity performance . In so doing, the teen extrapolates socialization from her natural and immediate embodied experience and pushes it into exogenous and persistent media that she must then watch over. The resultant compulsion to be watchful forces the teenager to be vigilant (Figure 2).
Because mobile devices are used primarily for communication, and because they are an extension of the body, they may compel vigilance for the simple reason that they give the user something to watch over. In becoming such powerful, embodied socialization tools, have mobile devices given rise to a new type of vigilant watchfulness? Perhaps by examining individual sessions of use we can start to understand the prevalence of vigilance in everyday contexts.
Because of the diversity of users and their relationship to computing devices and software, it is difficult to describe any general usage scenario as categorically vigilant or not. For a user who is deeply engaged in a social network, monitoring that network might become an important aspect of socialization—important enough to trigger vigilant attention. However, a casual user of that same social network may feel no need to monitor things—their usage may be entirely non-vigilant. In order to observe the prevalence of vigilance among day-to-day computer usage scenarios, we established a set of criteria against which an individual observed session of use can be evaluated as vigilant or not:
- High task importance. The importance of completing the associated task must be perceived by the user as being higher than average. This follows the simple logic that self-preservation is important; if a task is associated with self-preservation, then it must also be important.
- Endogenous or exogenous cueing. The usage session must have been invoked by an exogenous or endogenous cue. In other words, the user was either responding to some type of alert or notification, or had experienced a desire to check for some form of update or response. This differentiates vigilance tasks from planned tasks (like work or homework), directed inquiries (like Web searches), and boredom activities (like playing games or surfing the Internet). Exogenous cues are those that are specifically designed to capture the user's attention, like a smartphone buzzing when you receive a text message. Endogenous cues are invoked by the user's mind and manifest as the compulsion to direct attention to the device; for example, checking for a response to an email you sent, after having been away from the computer.
- Dire consequences for failure to complete task. The user should project that failure to complete the associated task would have dire consequences. This follows the logic that if you do something out of self-preservation, then you believe that you would be harmed in some way if you failed to do the task. This criterion helps establish that the task is associated with a scenario of sufficient gravity to compel vigilance.
Using these criteria, a small pilot study was performed within which two user groups were recruited and their device usage monitored for vigilance. One user group contained situated users and the other group contained mobile users. Eight participants were selected for each group—16 in total. Participants were asked to maintain an online daily journal, where they described a specific session of use that occurred on that day. Over the course of two weeks, each participant logged 14 unique journal entries. A total of 224 journal entries were recorded.
Users were not directly asked if they were being vigilant or not. Rather, they provided details about their usage sessions: what they were doing, the importance of these tasks, what prompted their usage session, and if failure to perform the tasks would have created any problems. A specific subset of the responses was associated with the vigilance criteria. If a participant's journal entry on any particular day contained all of the responses that mapped into the vigilance criteria, then that session of use was flagged as a vigilant one.
For the mobile group, 43.8 percent of the recorded sessions were vigilant. For the situated group, 14.3 percent of the recorded sessions were vigilant. Thus, mobile users were about three times more likely to be vigilant than situated users (Figure 3).
Mobile users had a higher standard deviation of 2.6 vigilant sessions per user compared with 1.5 for the situated users, indicating greater variance in the vigilance of mobile users. This could be because mobile devices support a greater diversity of usage scenarios, whereas situated usage was more often correlated with planned work. There may also be more variation among mobile users in the perceived importance of monitoring the device.
For a portion of the study, users were also directly asked if they felt a need to be watchful over their devices each day. On 78 percent of the days observed, mobile users said yes, compared with 44 percent for situated users. The ratio of daily watchfulness to vigilant sessions was 1.8:1 for mobile users, and 3:1 for situated users. As a measure of quality control for the study, vigilant sessions were expected to correlate to days when the user described a need to be watchful. This was found to be true: 95.5 percent of the vigilant sessions occurred on watchful days. The prevalence of watchfulness may also suggest that vigilant behavior occurred outside the sessions detailed in journal entries.
As an initial pilot, this study had several limitations, including a small sample of 16 individuals. Additional studies should be performed with a larger sample to substantiate the findings, ideally considering age and other demographics. This research also relied upon self-reporting rather than instrumentation or direct observation. The application of more quantitative methodologies could help mitigate participant bias, subjectivity, and errors.
Still, these early findings are compelling. A new form of vigilance has likely emerged in our evolving technological landscape. This new vigilance invokes and spans intermittent sessions of use, predominantly on mobile devices. Almost 50 percent of self-reported mobile sessions captured in this pilot were vigilant in nature, suggesting that vigilance is extremely prevalent among everyday mobile users.
So why does this matter? First, because any product or service that is designed for and delivered to mobile devices is being inserted into a usage session that just as likely as not was invoked by vigilant behavior. How should interactions be designed to integrate into that? Second, for any product or service that is designed for mobile use, we must consider whether it will directly compel vigilant usage. In either case, vigilance is an important dimension of how we must think about designing for mobile. Consider these interaction design principles for supporting vigilance:
- Vigilance first. Vigilance compelled the user to start their task. As such, the initial interface should be optimized for vigilant use. Users should be presented only enough information to know if they need to take further action or not. All actions should be enabled via efficient, intuitive paths within the application flow. The interface should eliminate distractors.
- Disengagement. Alerts should be minimized and vigilance should be supported via the shortest possible session of use. When the vigilant task is completed, the session should terminate. Shorter sessions create less interruption and distraction from the real-world experiences the user is balancing.
- Habituation. Interactions should be standardized and not novel. Users should be able to easily form habitual usage patterns that require less cognitive focus to perform. This also means that users may become desensitized to signals and alerts over time—variance can be introduced to combat that.
Vigilance likely isn't the only scenario that everyday consumer experiences must support. Vigilance emerges after a product or service is used enough that the user begins to feel the need to be watchful. Outside any moment of vigilance, the user may be more casual and exploratory, or even immersed or deeply engaged.
These principles call for reductive measures to better support vigilance, yielding products that pack all of their value into a delivery that's overall less engaging. Yet most consumer applications are designed in an effort to deepen user engagement. Vigilance requires a paradigm shift. If you are an interaction design practitioner, then sooner or later these principles will conflict with the business objectives of your employer. Designers learn to value the attributes of an engaging user experience—immersion, focused attention, and lengthy usage sessions—often above all others, because user attention is extremely valuable to software companies. Online user sessions are often monetized with advertisements, and the more time users spend within the product, the more opportunities there are to display ads. Many organizations track the length of user sessions or the average number of user interactions in order to raise the amount of money they can charge for ad impressions. Sites with engaged users can charge more than sites with disengaged users.
Even if your product doesn't compel vigilant use itself, there are many other mobile products that already do. Thus, even if you are only leveraging mobile devices as a channel to engage your customers, you should assess and model interruptions and user distractedness. Vigilance is still a behavior you must consider.
In spite of all challenges, it is important to design for vigilance in a world full of mobile users. A decade ago, before the proliferation of mobile connectedness, it was safe for an interaction designer to assume that vigilant watchfulness was limited to specific situated tasks performed by trained professionals or operators, and that the strain and distraction of sustained vigilance was confined to a user, in a chair, at a terminal. In today's world, that is an irresponsible assumption for a designer to make.
3. Parasuraman, R. and Caggiano, G. Neural and genetic assays of mental workload. In Quantifying Human Information Processing. D. McBride and D. Schmorrow, eds. Rowman and Littlefield, Lanham, MD, 2005, 123–155.
7. Boyd, D. Why youth (heart) social network sites: The role of networked publics in teenage social life. In MacArthur Foundation Series on Digital Learning—Youth, Identity, and Digital Media Volume. D. Buckingham, ed. MIT Press, Cambridge, MA, 2007, 119–142.
M. Giles Phillips is a designer and researcher examining the intersection of physical and mediated experience. He is the founder and executive director of Subforum, a research and design institute, and serves as the chief product designer at Constant Contact, where he is responsible for Web and mobile product design. firstname.lastname@example.org
Copyright held by author. Publication rights licensed to ACM.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2014 ACM, Inc.