| Previous studies have shown that online user reviews has important strategic value for product development.On the one hand,existing reviews will affect potential users’ down-load or use intentions of the product.On the other hand,reviews contain rich user demand information,it can encourage developers to generate better product versions.Therefore,companies need to pay full attention to user reviews in the process of APP development,especially version updates.Most of the previous literature emphasized that user reviews are very important,but in the actual product update process,few people pay attention to how developers use the demand information in the reviews to achieve better version updates.In other words,what role user reviews play in product updates remains to be studied.Based on the above background,The first research question of this article is:How does user review participate in the platform software upgrade?After solving this problem,the response pattern of the development team to user needs can be obtained,and the pressure of all aspects of the development team in the update may cause the response pattern to change.Then the second problem studied in this article is:What factors affect the response pattern of platform software to user needs and how these factors affect the response pattern?In order to study the role of massive user reviews in APP software update and re-design,firstly,this paper builds a set of text semantic matching methods based on deep learning,which can be used to extract the correspondence between product update log text and user review text.Secondly,from the two aspects of response degree and response cat-egory,we explore the response pattern of APP developers to user needs,propose response rate indicators to measure the response degree,and divide different matching result data into different data sets,perform topic mining separately to find the law in the response category.Finally,experiments are conducted using massive APP software upgrade log data and user review data collected from the real APP market to prove the effectiveness of the above methods.The research results show that the app development team’s adoption of user review content is less than 20%.Further analysis found that the content adopted by developers mainly focused on APP software functions,and the part of user reviews that pointed to corporate marketing activities,due to the limited role of the R&D team in corporate operations,was rarely taken into consideration.In terms of the research on the influencing factors of the response pattern of plat-form software,this article first puts forward three factors that may affect the response pattern of platform software to user demand:platform software’s own update preference,pressure from users and pressure from market competition.secondly,select Two types of platforms,social e-commerce and non-social e-commerce,are used as the research ob-jects.Based on theories of user value and user reviews,the relationship between each influencing factor and the response pattern of the platform is analyzed and hypotheses are proposed.Finally,a multiple linear regression model is used to test the hypothesis.Re-search has found that when platform software makes update decisions,it will first satisfy its own update preferences.The impact of pressure from users and market competition on the response mode will vary depending on the type of platform.In addition,this study also found that different types of platforms have different update preferences for upstream and downstream users.Social e-commerce platforms are more inclined to respond to the needs of downstream consumers,and non-social e-commerce platforms are more inclined to respond to the needs of upstream businesses.The research in this article has the following contributions.In terms of technology,this article proposes a machine learning architecture that analyzes product upgrades that users participate in based on the correspondence between product update logs and user reviews,this architecture can effectively solve the problem of "update log-user reviews"relationship identification and developers’ response pattern mining to user needs,and can be extended to update problems in other fields.In terms of management,this article pro-poses a model of the influencing factors of the developer’s response to user needs,which can be used to analyze the extent to which the development team is affected by pressure from different sources during software upgrades.On the one hand,this content fills the gap in existing research.On the other hand,the research conclusions can provide guid-ance for platform developers’ update decisions and achieve a more reasonable allocation of R&D resources. |
PDF Full Text Request