【作者】 戴和忠；

【导师】 陈劲；

【作者基本信息】 浙江大学 ， 管理科学与工程， 2014， 博士

【摘要】 数字内容商品是信息消费的重要领域。2013年国务院在《关于促进信息消费扩大内需的若干意见》中,明确提出要丰富信息消费内容,大力发展数字出版等新兴的数字内容产业。但是,由于网络消费环境的影响,以及典型的经验品属性等特点,数字内容商品消费面临着信息过载、质量不确定性等挑战,导致选择困难、购买延迟、对选择不满意,以及持续使用意愿低等问题。网络推荐和在线评论是推进数字内容商品消费的重要手段。其中,网络推荐是帮助消费者解决信息过载、筛选和发现目标、提升决策质量的有效工具,但有学者对其作用存在质疑。在线评论是网络消费者主动发现质量信息、实现间接体验,以及减少不确定性的重要信息资源,但其对产品销量的影响也存在多种观点。更为重要的是,现有研究忽略了网络推荐和在线评论之间的相互作用,也很少关注在用户观察学习过程两者同其它因素之间的相互影响。而对于数字内容商品而言,一般有一个体验和持续消费的过程,上述影响和作用会变得更加复杂。在此情况下,本文针对数字内容商品,研究网络推荐和在线评论等对产品体验和消费的整合作用,研究其中的影响因素和作用机理,从而能更好地指导网络推荐管理和在线口碑营销的开展。本文以中国移动手机阅读为例,从数字图书的访问、体验和持续消费的全过程出发,研究网络推荐和在线评论对体验和消费的影响。包括以下内容：一是对网络推荐、在线评论、观察学习及信息级联与网络消费之间的关系进行文献综述,同时研究梳理了持续体验和消费的相关理论和文献。二是针对文献中关于数字内容商品相关研究较少的问题,通过基于手机阅读的案例分析,从网站和图书两个层面整理归纳了数字内容商品在线体验和消费的基本过程和用户流失的漏斗模型,分析归纳了网络推荐和在线评论对其影响。三是在文献和案例的基础上,结合网络推荐和在线评论等影响,整合构建了“发现和访问--观察和选择--持续体验和消费”三阶段研究模型,提出了13个方面21条研究假设。四是针对第一阶段(发现和访问阶段),开展网络推荐影响的实证研究。分析了排行榜、智能推荐和网站人工推荐对数字图书详情页访问量的影响。五是针对第二阶段(观察和选择阶段),开展网络推荐、在线评论和观察学习对阅读选择影响的实验研究,并分析了其交互调节效应,研究了上述影响在不同生命周期阶段的动态变化。六是针对第三阶段(持续体验和消费阶段),开展在线评论及观察学习对持续体验消费影响的实证研究。结合思维定式理论和期望确认理论,实证分析了在不同章节消费期望和品质感知影响的动态变化。通过理论研究和实证分析相结合,本文的研究结果及贡献如下：首先,针对数字内容商品消费的特殊性,提出了三阶段整合模型。包括“发现和访问”、“观察和选择”、“持续体验和消费”三个阶段,并分别以详情页访问用户数、阅读率和用户留存率为因变量,环环相扣,同时将网络推荐、在线评论、观察学习、消费期望和品质感知等纳入整合模型进行系统的研究。实证研究也发现了三个阶段的相关性,如网络推荐类型在第一阶段会正向影响图书详情页的访问用户,在第二阶段对阅读率也有直接影响以及调节作用,而综合两阶段影响,网络推荐对阅读用户数有更显著的影响。其次,研究了网络推荐三种类型对图书访问量的影响,发现排行榜的信息级联效应和帕累托分布的存在,并发现智能推荐强度对详情页访问量有正向显著影响,但是到一定程度后边际效益递减。进一步采用A/B Test现场实验研究发现,网站人工推荐中推荐信息来源有显著影响,用户推荐大于专家推荐大于无来源推荐,但受到用户阅读深度的调节,浅度阅读用户更倾向于专家推荐,而深度用户更倾向于用户推荐。针对TOP5图书推荐,提示畅销信息,比不提示信息时,其访问量有显著提升,且此时推荐来源的影响差异不显著,说明畅销排行信息有高诊断力和信息级联的作用,但需要信息展现和用户的显著感知。第三,研究了网络推荐、在线评论和观察学习的整合影响,发现在线评论的评分和差评率对阅读率有显著影响。评论数量一开始有影响,但增加了观察学习变量,尤其是增加了其交互作用项后,影响变得不显著,说明观察学习对在线评论有部分替代作用。发现详情页的累积点击量作为观察学习变量对阅读率有正向影响,在线评论和观察学习之间存在一种互补机制。同时还发现,网络推荐不仅直接影响阅读率,还有调节作用。另外,图书上架时间越长,上述影响总体越强。最后还发现,内外部口碑效应(不包括在线评论)和产品扩散效应对阅读率有显著影响,需要控制。第四,研究了消费期望和感知品质对持续体验消费影响。创新测量了在线评论和观察学习带来的消费期望因子,并发现,随着阅读章节的增加,感知品质对留存率的正向影响逐步变小但仍然显著,而消费期望的正向影响会逐渐降低且变得不显著。图书前几章节对是否留存至关重要,另外免费转收费有显著负向影响。本研究在数据采集和研究方法上做了一定的探索和创新。通过大数据方法采集手机阅读平台上“图书-用户-图书”的深度面板数据,如用户访问路径、详情页访问用户数、用户在一本图书每章的留存等,构建了超越以往相关研究的大数据基础。同时,除了基于面板数据进行多元回归分析外,在网站人工推荐的研究上创新性地采用A/B Test现场实验方法,在持续体验消费研究上探索构建了分段多元回归比较的研究方法。更多还原

【Abstract】 Digital content product is an important field of information consumption. In the Several Opinions on Promoting Information Consumption to Expand Domestic Demand (2013), the State Council has explicitly put forward to enrich the content of information consumption, and to develop the emerging of digital content industry such as digital publishing. But because It is deeply affected by the network consumption environment, and it has the typical nature of experience goods, the consumption of digital content products is faced with challenges such as information overload, quality uncertainty. This leads to difficulty to choose, delayed purchase, dissatisfied for the choice, and lower willingness for continuous use.Online recommendation and online reviews are important means to propel the consumption of the digital content goods. Online recommendation is an effective tools to help consumers solve the information overload, filter and find the target products and improve the quality of decisions, but scholars remain suspicious of its effect. Online reviews are important source of information for online consumers to find out information about the quality of the product, to realize indirect experience and to reduce uncertainty, but there are many inconsistent opinions about its effect on sales. What’s more important, the existing researches ignores the interaction between online recommendation online reviews, and pays little attention to the mutual effect among these two and other factors when consumers are in the process of observational learning.In addition, there is a process of experience and continuous consumption, thus the influence and effect as mentioned above will become more complicated. In this case, this paper studies into the integration effect of online recommendation and online reviews to the experience and consumption of digital content products, explores the influential factors and mechanism, and aims to gives better guidance for the management of the online recommendation and online word-of-mouth marketing.This paper takes China Mobile Reading as research object; it starts from the whole process of access, experience and continuous consumption of the digital books, studies into the recommend research network and online comments on experience and the impact of online recommendation and online reviews on the basic process of experience and consumption of digital content products. It Includes the following contents:(1) A literature review for the relationship of online recommendation, online reviews, online experience and online marketing.(2) As the lack of existing related research, this paper, on the base of a case study of China Mobile Reading, concludes the basic process of experience and consumption of digital content products from the two aspects of website and books, analyzes and concludes the impact of online recommendation and online reviews.(3) Based on the theoretical and case analysis, our study combines the influences of online recommendation and online reviews, constructs a three-stage research model which includes "discovery and access--observe and select--continuous experience and consumption", and then proposes21hypotheses.(4) For the first stage (discovery and access stage), we empirically examine the effects of online recommendations. We analyse the effect of ranking、intelligent recommendation and artificial recommendation on the visits of digital book details page.(5) For the second stage (observation and selection stage), we empirically assess the effects of online reviews and observational learning. Our study analyses the effect of online reviews, online recommendation and observations learning on reading select, and the changes of the effects in different stages of the life cycle.(6) For the third stage (continuous experience and consumption stage),we empirically research the effect of online reviews and observational learning on continuous experience consumption.Based on the combination of theoretical and empirical analysis, we draw the following important conclusions:Firstly, Our study propose a three-stage integrated model for the particularity of digital content products. Three phases include "discovery and access,""observe and select","continuous experience and consumption", and we use the number of users to access the details page, reading rate and retention rate of users as the dependent variable. Meanwhile, online recommendation, online reviews observational learning, consumer expectation and perceived experience are involved in the integrated model system for analysis.Secondly, our research examines the effect of three types of recommendations on book visits through empirical ways. The study has found that there exists a cascade effect of ranking information and Pareto distribution, and the strength of intelligent recommendation has a significant positive impact on the details page visits, but to a certain extent diminishing marginal benefit. Further study using AB field test has found that sources of information in the artificial recommendation have a significant impact, and the impact of the experts recommendation is more than the impact of the users recommendation while the impact of the users recommendation is more than the recommendation with no source recommendation. For the TOPS ranking books. there are significant more book visitings when providing selling information labels than the same content but without these information.In this case the impact of the recommendation source was not significant, which fully explains that the top selling information has high diagnostic power and information cascade effect.Thirdly, we research the integrated impact of the online recommendation、online reviews and observational learning, finding that online reviews and poor rates have a significant effect on the reading rate. The quantity of rates is influential in the beginning, but becomes insignificant after observational learning variables are added, which indicates that observational learning has the partly substitute effect. Research finds that, as an observational learning variable, the accumulated PV (Page View) has positive influence on reading rate, and there is a complementary mechanism between online reviews and observational learning. In the meantime, we also find that, online recommendation can not only influence the starting rate of reading, but also have moderating effect, which enhances as the time goes from the books’putting on sales. Fourthly, the essay has done some research about the influence of consumption expectation and perceived quality on continuous experience consumption. We innovatively measured the consumption expectation factor showing the influence of online reviews and observational learning on next stage behaviors was constructed. Research finds that, as the number of pages read increase, the positive effect of perceived quality decreases but is still significant, while the positive effect of consumption expectation weakens and becomes non-significant in the end. One book’ s first few chapters are vital to decide whether the consumer will stay.Lastly, the essay has done some innovation on data collection and research methods. Using big data method, we collect a large sample of deep panel data on the Mobile Reading platform, like user’s visiting route, detail pages’quantity of visiting users, staying rate in every chapter of a book and so on. With these we construct the big-data foundation which surpasses previous researches. In the meantime, this essay does multiple regression analysis based mainly on panel data, and controls the inner difference of books using fixed effects method. In the research of artificial recommendation on a website, the essay uses the method of A/B Test field experiment innovatively to improve the inner and external effect. In the research of continuous experience consumption stage, the essay constructs a creative method of segmented multiple regression.更多还原