INTRODUCTION
Mobile apps are very popular and have been the focus
ofnumerous studies in recent years [19], [24]. A fundamentalchange introduced
by mobile apps is the way that they arereleased to users, which is through app
stores. App storesallow users to directly provide feedback on the mobile
appsthrough user reviews. Although these user reviews were meantto simply
provide feedback about the apps, they proved to bemuch more useful. https://codeshoppy.com/shop/product/water-complaint/ For
example, studies have shown that theycan be used to understand user problems so
that developerscan avoid low ratings, which can have a major impact on theapp’s
user base and revenues [12], [14], [25].
More recently, wearable devices have
been introduced,which complement handheld devices. Wearable devices i.e.,smart
watches and fitness trackers, are becoming increasinglypopular and are expected
to reach 101 million devices by2020 [6]. Wearable devices provide developers
with accessto unique sensors that can be used to enhance the user experi-ence
[2]. As such, developers began to develop apps that arespecifically designed to
run on these wearable devices, calledwearable apps.
Wearable apps are different
than handheldapps that run on mobile phones since they
1) often are
verylightweight (resources wise),
2) meant to run on very smallscreens,
3) have
access to a different set of sensors, and
4)heavily depend on the mobile device
to perform the majority ofthe heavy processing.
However, wearable devices have
uniquecharacteristics that pose challenges when compared to otherplatforms or
devices [26]. To the best of our knowledge, veryfew studies have focused on
wearable apps to date.Therefore, similar to the prior studies on (handheld)
mobileapp reviews [13], [15], [17], [28], we also investigate usercomplaints
but our study focuses on complaints from users ofwearable apps. To perform our
study, we manually classify589 reviews belonging to 6 wearable apps. The
reviews weretagged by the first two authors of the paper and groupedinto 15
different categories. For each category, we measuredthe frequency of the
complaints. Our findings indicate thatfunctional errors, lack of functionality,
and cost are the threemost frequent complaints.
RELATEDWORK
The work that is most related to our
study falls into twomain categories: work that leveraged mobile user reviews
andwork focusing on wearable apps.A. Work Leveraging Mobile User ReviewsOne of
the first studies to leverage mobile app reviews wasby Harmanet al.[14]. In their
paper, the authors studiedthe correlation of user reviews with key performance
metricssuch as the number of downloads. They found that there is astrong
correlation between app ratings and its rank based onthe number of downloads,
suggesting that developers shouldpay close attention to their user
ratings.Other studies mined user reviews to better understand thecontents of
these user reviews. Khalidet al.[17] studied low-rated user reviews from 20
free iOS apps in order to helpdevelopers understand their nature. They exposed
12 types ofcomplaints and found that feature requests, functional errorsand,
crashing apps were the most frequent reasons for negativereviews. Haet al.[13]
manually analyzed the user reviews of59 Android apps to examine the impact of privacy
and ethicalissues.
There are also a plethora of other works on mobile apps,that
leverage users reviews for their techniques. Due to spacelimitations, we only
discuss the most relevant studies in thissection, however, we refer the reader
to a recent survey byMartin et al. [19] for a more comprehensive list of
studies onmobile apps.B. Work Focusing on Wearable AppsVery few studies have
focused on the study of wearableapps, but many different paths are beginning to
get explored inthe domain. Recently in [29], Zhanget al.presented a
formalsemantics to statically model the notification mechanism ofAndroid Wear,
and contributed with the development of twodomain-specific tools, one for test
case execution and anotherfor automated test generation. Ahola [5] exposed 3
issuesand limitations in the Android Wear platform found duringwearable app
development that are better wear Internet con-nectivity, virtual button support
for watch faces, and softwareconfigurable language support for voice input.
From a differentperspective, Lyons [18] did a study on the user perceptions
offunctionality and design of smartwatches, including androidwearable devices.
Based on user feedback and contrast totraditional watches, possible features
for future wearable appare suggested. Min et al. [22] explored the battery
usage ofwearable apps and performed an online survey to get directfeedback and
concerns from users. They found that most usersdo not complain about the
battery usage of their wearabledevices.Chauhanet al.[6] did a previous
categorization of smartwatch apps from Samsung, Apple, and Android Wear.
Theyused data from Android Wear Centre and GoKO [1], [4] assources to get the
wearable app identifiers for crawling theirinformation; we applied the same
approach to initialize ourcrawling phase. To the best of our knowledge, this is
the firstwork focusing on the study of wearable app user complaints.
