Article Preview
TopIntroduction
In recent years, the market for wrist-worn smart devices has seen a drastic increase. Not only fitness trackers, such as the Fitbit, but especially more advanced devices such as the Samsung Gear series or the Apple Watch contributed to this increase in interest by consumers. Even though we have seen various advancements that allow for more complex in- and output modalities, they still suffer from a rather short battery life and are still limited in terms of computational power and display capabilities. Meanwhile, the software side has undergone drastic advancements lately, with an increase in third-party applications, but it is still questionable how users will benefit from this development. For example, Pizza, Brown, McMillan, and Lampinen (2016) found in their video analysis of 1009 uses of the Apple Watch that the most common one (~50%) was to simply glance at the watch face. This of course raises the question, what are the main benefits of these devices compared to their analog counterparts?
Besides displaying the time, notifications are another primary use case (Schirra & Bentley, 2015). Especially for quick glances, the wrist has proven to be a well-suited location, that allows for easy perception of changes and sufficiently fast reaction to them (Ashbrook, Clawson, Lyons, Starner, & Patel, 2008). Shirazi and Henze’s (2015) analysis of notification techniques across mobile devices showed that users base their decision for whether a certain notification should be displayed not only on the importance of the notification, but also on the output device. Their work implies that the user not only needs a system to filter and control the notification flow of their mobile devices, but also a way of choosing suitable modalities for their notifications.
While a variety of research has focused on novel interaction techniques for smartwatches (e.g. Kerber, Krüger, & Löchtefeld, 2014; McIntosh et al., 2016), most current devices are still based on conventional touch input or utilize additional mechanical inputs (Kerber, Kiefer, & Löchtefeld, 2016). But those input techniques come with the drawback of requiring both hands to interact with the device (Kerber, Schardt, & Löchtefeld, 2015; Kerber et al., 2016) While gestural input has become a standard on Android Wear, and may seem like a valid alternative especially as it only requires one hand to control the device, the risk of false positives creates a need for a gesture delimiter (e.g. Kerber, Schardt, & Löchtefeld, 2015). Compared to these advances only very little work has focused on how to convey and manage notifications on such devices (e.g. Ashbrook, Clawson, Lyons, Starner, & Patel, 2008; Gouveia, Pereira, Karapanos, Munson, & Hassenzahl, 2016; Hwang, Song, & Gim, 2015). Typically, notifications are just shown in a linear way – one by one – and switching between them requires a dedicated user interaction. Additionally, currently only very limited possibilities to cluster and filter notifications effectively exist for smartwatches.
In this paper, we try to approach the problem from a slightly different angle. Instead of focusing on novel interaction techniques to deal with (maybe even unnecessary) notifications, we contribute by investigating novel filtering and visualization techniques for smartwatch notifications. Based on our custom-built, energy-efficient smartwatch prototype we explore how the combination of a low-resolution display and an ambient illumination frame, consisting of 12 RGB LEDs around it, can be utilized to communicate notifications more effectively. Additionally, we developed an application that allows users to effectively filter and prioritize the notifications that will reach the smartwatch in the first place. Based on the results of two user studies that explored the two aspects of this prototype we present a set of guidelines that can be transferred easily to current mass-market devices.