An Analysis of User Experience of Smartphone based on Product Smartness utilizing Social Media Data

Abstract

With rapid technology advancement and an expanding product domain, the definition of smart products has slightly varied (Rijsdijk & Hultink, 2009; Zaeh, Reinhart, Ostgathe, Geiger, & Lau, 2010). Also, from previous studies (Freudenthal & Mook, 2003; Rijsdijk & Hultink, 2003, 2009; Park & Lee, 2014), the relationship between product smartness and consumer appreciations or values can be identified. However, it is unclear to understand implicit needs of the consumers through conducting questionnaire based survey method. This method does not often provide sufficient information on the underlying meaning of the data, and strong evidences of causation to an answer (Gable, 1994). Hence, it could be more effective to collect unrefined and numerous user experiences, which are freely expressed in their own words, for better observation of natural user behaviors. Therefore, we tried to observe user experiences utilizing social media data, which can infer people’s opinions, both at an individual level as well as in aggregate, regarding potentially any subject or event (Schonfeld, 2009), to identify perceived product smartness. Since a smartphone is one of the most successful smart products, it could be represent the characteristics of smart products better than other products. Thus, ‘smart phone’ and ‘mobile phone’ are selected as search keywords. Through literature reviews, the dimensions and attributes related to product smartness from various previous studies were collected. Then, the collected dimensions of product smartness were re-categorized into five main dimensions as follow: ‘Autonomy’, ‘Adaptability’, ‘Multi-functionality’, ‘Connectivity’, and ‘Personalization’. The overall procedure of analyzing the relationship between perceived product smartness and collected user experiences of smart products from external data source (Twitter) is as follow. First, user experience of smart products was collected through mining Twitter data using software tool (SOCIAL metrics). SOCIAL metrics ( http://socialmetrics.co.kr ), which is developed by DaumSoft, can help for analyzing big data. It enables to collect Twitter data and show the frequency of keywords related to user’s search keyword. Second, data pre-processing was conducted. In the search results, the tweets which are not related to user experiences of smart products are eliminated. Third, collected user experiences were categorized according to the conceptual model of product smartness. Then, identifying the relationship between each dimension of product smartness and users’ positive/negative experiences was performed by manually. Finally, the reason of users’ positive or negative emotions on experiences of smart products was identified. A total of 19,288 tweets including ‘smartphone’ were collected from 2014.06.01 ~ 2014.08.31. Among them, a total of 699 tweets are actually related to user experiences of smartphones. The collected tweets were categorized according to the dimension of product smartness and the reason of user’s emotion. According to the results, there were many positive experiences for all of dimensions, but there were negative experiences only for multi-functionality and connectivity. Some results were supported by existing studies. The reason for positive experience on autonomy corresponded with the result of other study that productive daily life is a critical means for users to develop sense of confidence (Jung, 2014). Negative experience of autonomous was not shown in the results, but actually autonomous product does not always increase satisfaction of product. According to Rijsdijk and Hultink (2003), high complexity in using products would decrease satisfaction of products. Providing an autonomous product with indicators that inform the user about what the product is doing may reduce risk perceptions (Rijsdijk & Hultink, 2009). The study suggested that a mining technique can be used to gather and analyze user experience effectively and quantitatively without bias. It is expected that the proposed method could be helpful for understanding user’s implicit needs on the products.