Dissertations / Theses on the topic 'Customer review quality'

碩士<br>東吳大學<br>企業管理學系<br>106<br>The customer satisfaction of accommodation service depends on guests’ expectations and actual experiences. Prior customer reviews, ratings, price, and promise from host are all related about travel experience. Airbnb is an online platform that connects people looking to rent their homes with people who are looking for accommodations, the customer reviews and expectation are different from hotel service. This study analyzes customers ratings and comments from Airbnb, combining prior studies and theories to explain service quality and customer satisfaction with reviews on Airbnb. We crawled all the customer reviews of unsold listings on Airbnb at 2016/10/07 in Taipei. The raw data are 9028 reviews from 253 listings, 5469 reviews in Chinese from 178 listings after data cleaning. After text mining from customer reviews, we grouped tokens by features and there are 36 keyword factors. We extracted 35 keyword factors to compare with repurchase intention keyword factors, price and region factors for analysis to know which would be crucial to customers satisfaction. The result showed there are 16 keyword factors related to customer reviews, 6 of them could be crucial to make a impact on customer rating, 5 of them would increase the probability of occurrence of repurchase intention keyword factors. In the future, we would like to analysis service quality with structured dataset in a systematic way. This data-based customer reviews analysis would be an important review for hosts on Airbnb in Taipei, government officials and travel agencies to know more about customers and hosts on Airbnb.

碩士<br>輔仁大學<br>資訊工程學系碩士班<br>106<br>Wireless networks are ubiquitous, and the role of wireless routers is becoming more and more important. Most wireless routers are manufactured by outsourcing foundries, and the quality of wireless routers depends on the specialty of the foundry. If product quality is not well controlled during mass production stage, there will be an extra burden of heavy work. Not only the high cost of the foundry, but also the customer satisfaction with the outsourcing foundry decline. If the product is returned by the consumer, loss is not just the sale revenue, even the brand image will be influenced. This study collects the customer reviews of wireless router products from the Amazon website, and tags the data using five parameters: reply, function, error status, connection problem, and keyword. The customer reviews is trained and tested using the decision tree to know the classification result and severity levels. Severity levels 1 to 5 are assigned to the engineering department to enhance the testing quality and methods, and Severity level 6 to 7 are assigned to the marketing department are assigned to the market research department for customer behavior analysis. This study also tags the customer reviews using wireless network category, year, and price, and classifies with the decision tree and logistic regression. Compare the two methods’ Precision and Recall values. Through the results of the experiment, the decision tree will be more suitable for the classification of customer reviews.

At the moment, the largest amount of tissue paper products are consumed in East Asia (which includes China and Japan), Europe and North America, and between 2010 and 2023, consumption of tissue paper products is expected to grow by 3% with the largest growth rate in China. Meanwhile, toilet paper, kitchen towels and facial tissues, combined make up 85.2% of all tissue paper products consumed in 2018. While different factors can be named as responsible for this growth in consumption, it is quite difficult to isolate the product-related factors that contribute to the growth of consumption of tissue paper in different countries due to the fact that tissue products have limited uniform specifications between different suppliers. This means, tissue suppliers and tissue machine manufacturers, like Valmet, are left with the challenge of understanding customers’ motivations behind demand of different tissue paper products. This is the motivation behind this study, which is to investigate what qualities of toilet paper, facial tissues and kitchen towels, are customers in the US, China and Japan satisfied or dissatisfied with. This is compared with measurements of the qualities of these products, like softness, strength, thickness, to determine the ranges of values of the products’ qualities that customers are satisfied with. Online reviews of 34 different brands of tissue paper products are collected ethically from online retailer websites (Amazon US, Amazon Japan, Walmart, JD.com), and analysed using the thematic analysis methodology to identify the list of qualities of tissue products that customers are satisfied or dissatisfied with. Then, the Pareto rule was used to identify the top 20% most important qualities to customers. These 34 brands were also purchased from these online retailers and their different qualities were measured. The results and conclusion from the study revealed different insights, among which is that softness is a must-be quality for toilet papers in all countries studied, while water absorption is a must-be quality for kitchen towels. The study also revealed ranges of the numerical values of these qualities that are acceptable to end consumers. Valmet intends to use the conclusions from the study in engagements with tissue paper suppliers in the three countries regarding product strategies and choice of technology.

碩士<br>開南大學<br>資訊學院碩士班<br>105<br>Advancing technologies have changed people's lives. A survey by Market Intelligence & Consulting Institute indicate that Taiwan internet users rely more and more heavily on online shopping and are willing to spend more on virtual channel. The population of online shoppers increases every year and has spawned to be called the “Otaku Economy”. The booming online shopping market and the low cost of switching between different online stores has led fierce competition among shopping website platforms. Thus, seizing business opportunities to maintain the upper hand in the market has become critical issue for every shopping website. This study chooses SHOP.COM as research subject to discuss the influence of system quality, information quality, service quality and perceived usefulness on customers’ satisfaction based on Information System Success Model and Technology Acceptance Model. The study reviews past literature to develop conceptual framework and then discuss the impact of website quality on customer satisfaction. The study intend to improve the quality of online shopping platform, to increase the costumer’s purchase intention, and also to form the basis for future business decision-making.

碩士<br>逢甲大學<br>國際經營與貿易學系<br>107<br>Information and Communication Technology (ICT) has become the major trend in the competitive market, more and more users prefer to purchase goods through e-commerce platforms. After the customers shop on the e-commerce platform, the shopping experience would be shared on the e-commerce platform in the form of comments. These reviews contain massive information that is useful to merchants. Logistics is an indispensable part of the entire online shopping process. Poor service of distribution personnel, damage of goods, errors in package delivery, and high freight costs may occur in the process of logistics. The existence of these factors affects the customers satisfaction during online shopping. Under this circumstance, understanding the impact of logistics service quality on customers' post-purchase behavior can help e-commerce companies improving customer loyalty, reducing marketing costs, and strengthening enterprises competitiveness. Based on text mining technique, web crawler is adopted to obtain comment text in this paper, and apply data mining algorithm to obtain the factors affecting customer experience satisfaction included in the comment text. This research can quickly and timely obtain the factors affecting customer experience satisfaction from customer reviews, provide effective information support for business operations, and assist decision makers in making decisions. We found that Amazon customers have higher satisfaction with logistics loyalty, return service, and logistics service quality. In terms of logistics customer service, customer satisfaction is lower, and Amazon's logistics customer service needs to be improved.

碩士<br>國立中興大學<br>行銷學系所<br>103<br>With the development of Big Data, the marketing research and the technology development have new potential opportunities. The growth of Web2.0 is one of the reasons why data volumes keep surging, and contributes to the discussion of online customer research. Customers evaluate the product or service often depends on their satisfaction. Compare to the traditional way of measuring satisfaction, customer reviews are close to the customer real feelings. Past studies have shown that there are some relationship and influence between positive and negative emotions and satisfaction, but the relative to their influence has no consistent conclusions. However, according to expectation-disconfirmation theory can explain it. Sentiment analysis was applied to various fields of customer reviews, and had learned about customer’s thoughts and emotions. This study used polarity dataset v2.0 and AFINN to build the database and trained the model through Naive Bayes. In this study, we selected customer reviews total of 21 TripAdvisor’ hotels. This customer reviews totally contain 20996 sentences wrote by 3054 users. Finally, we used Pivot Analysis to calculate the sentiment scores and regression analysis.to get results. The results can be verified expectation-disconfirmation theory. When the hotel is high (low) quality, the degree of the negative (positive) reviews’ effect on satisfaction is higher than the positive (negative) reviews’. That is the moderating effect of service quality on the relationship between the positive and negative customer reviews and customer satisfaction.

碩士<br>國立臺灣大學<br>資訊管理學研究所<br>102<br>Online customer reviews are considered as an important role plays on customer purchase decision making process. Previous researches show that the opinions of online reviews have influence on the perceived product quality, the purchase intention, and finally the sales. However, previous studies usually study quantitative information and qualitative information of online customer reviews separately, even if there are usually both the quantitative information, the rating, and the qualitative information, the comment in each online customer review. Based on the results of data mining, this study finds that there are discrepancies between the opinion of comment and the opinion of rating in some online reviews. Therefore, this study aims at finding the effects of quantitative information and qualitative information on perceived product quality in single online customer review for experience goods. The results show that the detailed qualitative information has greater influence on the perceived product quality than the quantitative information. In addition, this study also studies the effect of discrepancy between two online reviews on one product’s perceived product quality. The results indicate that customers really are affected by the qualitative information, the discrepancy will neutralize the perceived product quality and the negative information takes more weight on product quality judgment. The results follow the negativity effect discussed by prior researches. Finally, this study links the online reviews to perceived product quality, then the purchase intention. This study proposes a new idea to explain the effect of online customer reviews on customer purchase decision making process. Customers are not blind to receive the quantitative information, the discrepancy between rating and comment will be discovered and the quantitative information plays more important role on perceived product quality and purchase intention.

碩士<br>國立成功大學<br>企業管理學系碩博士班<br>100<br>This research focuses on service quality gaps between customers and front-line salesmen by taking Yuanta Securities, the largest securities organization in Taiwan, as a case study. Knowing that service quality is one of the most influential factor for customers to choose their ideal securities corporation, if we could discover service quality gaps and eliminate these gaps, it would greatly help securities companies to satisfy their customers, attract more potential customers, gain profitability, and eventually enhance competitive advantage. The objectives of this research are shown as below: (1) To revise SERVQUAL scale in order to fit features of securities industry. (2) To reveal the perception of service quality gaps between Yuanta customers and salesmen. (3) To offer viable suggestions to Yuanta in terms of improving service quality. Four hypotheses are established based on PZB model and three kinds of questionnaires are delivered to Yuanta customers, salesmen, and managers to test these hypotheses which are list as below: H1: Significant difference exists between “Yuanta perception of customer expectations” and “Yuanta consume expected service H2: Significant difference exists between “Yuanta perception of customer expectations” and “Yuanta translation of perception into service quality specification H3: Significant difference exists between “Yuanta translations of perception into service quality specification” and “Yuanta service delivery” H4: Significant difference exists between “Yuanta customer perceived service” and “Yuanta consume expected service. Results of this research are listed as below: (1) A revised SERVQUAL scale, which fits the features of securities industry, is established with five dimensions and 24 items. (2) Perceived service quality gap does exist, especially in telling customer investment risk clearly, assisting customer proactively, having professional knowledge and ability, and respecting customer equity. (3) Yuanta Securities should promptly improve: training salesmen to complete customer requests in time, not to postpone customer tasks, and tell customer investment risk clearly.

JEL: M31 Marketing JEL: L81 Retail and Wholesale Trade; e-Commerce<br>Companies in the fashion industry are increasingly establishing their presence online, in the form of online platforms, such as websites. However, some brands are more successful than others in selling fashion clothing online. In this perspective, the two main objectives of the dissertation are: (i) to adapt and extend the UTAUT model in the context of online shopping for fashion clothing; (ii) to compare the adapted model in two regions of Portugal and Italy. By testing the adapted model in two regions of Portugal and Italy, the study gives insights into online consumer behaviour when using fashion websites. Results are based on a survey of 312 Internet users who are familiar with fashion websites. Findings reveal that perceived website quality is the strongest predictor of online behaviour, having a positive impact on customer satisfaction, website trust and word-of-mouth. Besides, the present study explores whether online shoppers are or are not influenced by other customers’ reviews and finds evidence that online customer reviews do not impact consumers’ perceptions of online shopping usefulness.<br>As empresas na indústria da moda estão cada vez mais a estabelecer a sua presença online, sob a forma de plataformas online, tais como websites. No entanto, algumas marcas são melhor sucedidas do que outros na venda de roupas de moda. Nesta perspetiva, os objetivos principais da dissertação são: adaptar e estender o modelo UTAUT ao contexto de compras online para a roupa da moda; comparar o modelo adaptado em duas regiões de Portugal e de Itália. Ao testar o modelo adaptado em duas regiões de Portugal e Itália, o estudo permite obter informação sobre o comportamento do consumidor online ao usar sites de moda. Os resultados são baseados em uma pesquisa com 312 usuários de Internet que estão familiarizados com websites de moda. Os resultados mostram que a qualidade percebida do website é o mais forte determinante do comportamento online, tendo um efeito positivo na satisfação do cliente, na confiança e no passa- palavra. Além disso, o presente estudo investiga se os compradores online são ou não influenciados pelos comentários de outros clientes e encontra evidência que os comentários de clientes não têm efeito significativo na perceção dos consumidores online.

The hospitality and tourism industry was boosted by the help of hotel review sites, which consists in an increasing demand on the part of tourists. We extracted more than thirty thousand reviews from Tripadvisor to understand the variations in customers' perceptions of high/low end and chain/independent hotels and on which aspects this variation is most evident. We used sentiment analysis to assign a score to the aspects of each review. We compared machine learning algorithms, namely, random forest, decision tree and decision tree with adaBoost, to predict the overall score. Then, we used the Gini index to understand the aspects that most influence the overall score. Finally, we compared the reviews with temporal windows overtime with Jaccard index to characterize the dynamics of customer satisfaction focusing on three aspects: "Service", "Location" and "Sleep". Correlating the responses of the hotel to the users' reviews, we wanted to demonstrate the impact in the customers' perception of the hotel quality. The best performances were achieved by the decision trees which indicated that "Service" is the most influential aspect for satisfaction, while "Location" and "Sleep" were the aspects considered less important. By identifying the moments of drastic changes, we verified that "Service" is also the most related to the overall score. These analyses allow hotel management to track the trends of tourists' assessment in each category. Generally speaking, a focus on the "Service" should be done. However, an analysis, for a particular hotel, of the dynamics of the overall score to compare with its category would be advantageous.<br>A indústria da hospitalidade e turismo foi impulsionada pela ajuda de sites de avaliações de hotéis, que leva a uma exigencia crescente por parte dos turistas. Extraímos mais de trinta mil avaliações do Tripadvisor para entender as variações nas percepções dos clientes de hotéis de alta/baixa gama e cadeia/independentes e quais os aspectos essa variação é mais evidente. Usámos sentiment analysis para atribuir uma pontuação aos aspectos de cada avaliação. Comparámos algoritmos de aprendizagem automática, nomeadamente, "random forest", "decision tree" e "decision tree with adaBoost", para prever a pontuação geral. Depois, usámos o índice de Gini para entender os aspectos que mais influenciam a pontuação geral. Por fim, comparámos avaliações com as janelas temporais ao longo do tempo com o índice de Jaccard para caracterizar a dinâmica de satisfação do cliente com foco em três aspectos: "Service", "Location" e "Sleep". Ao correlacionar as respostas do hotel com as avaliações, queriamos demonstrar o impacto na percepção dos clientes sobre a qualidade dos hoteis. Os melhores desempenhos foram alcançados pelo decision tree que indicou que "Service" é o aspecto mais influente para satisfação, enquanto que "Location" e "Sleep" foram os aspectos considerados menos importantes. Ao identificar os momentos de mudanças drásticas, constatámos que "Service" também é o mais relacionado à pontuação geral. Estas análises permitem que a gestão dos hoteis acompanhe as tendências da avaliação dos turistas em cada categoria. De um modo geral, um foco no serviço deve ser feito. No entanto, uma análise, para um hotel particular, da dinâmica da pontuação geral para comparar com sua categoria seria vantajosa.