;With vigorous development of the mobile network comes the evolving of smart devices from fixed-line ones to portable ones. The development of related application programs keeps growing furiously. Apps greatly change people’s lifestyle and life attitude. An app icon plays a crucial role when it comes to making an app stand out from millions of other similar ones to make that app more preferable for users to download. An icon is, after all, the very first visual stimulation that users receive when they download an app. An icon can be taken as the first impression of an app; it is indelible and has a profound influence as well.
Targeting users who have used handheld smart devices and downloaded any apps, this study focuses on the relevance between app design attributes’ ranking and icon traits in different categories of apps. Through organizing and studying related literature and on the base of the theory of Feature Discrimination Theory, this study presents five sorts of icon design attributes（activity, balance, complexity, depth and organic） and focuses on twelve app categories（arcade game, brain game, casual game, entertainment, tools, communication, productivity, personalization, music and audio, social, media and video, and travel and local） to perform research and provide analyses on the icon traits such as identification, influence of feelings, subjective familiarity, dominance and significance.
Through experimental design methodology, this study discovers that in all icon design attributes, activity is the most attractive one. This attribute can be taken into consideration when designing icons. Complexity and balance, however, are relatively not popular. This result shows that users have lower interests in the icons whose meanings are hard to understand or whose appearances are tedious. Furthermore, this study finds that all icon traits would influence the ranking of icon design attributes. Only identification would hardly influence the ranking in a long period of time. The ranking of some app categories, however, would not be influenced by identification in a short period of time, and those categories that would be influenced are the ones with higher download rates. This finding thus shows that iconic traits are profoundly influential to users.