An empirical comparison of machine learning methods for text-based sentiment analysis of online consumer reviews

The amount of digital text-based consumer review data has increased dramatically and there exist many machine learning approaches for automated text-based sentiment analysis. Marketing researchers have employed various methods for analyzing text reviews but lack a comprehensive comparison of their performance to guide method selection in future applications. We focus on the fundamental relationship between a consumer’s overall empirical evaluation, and the text-based explanation of their evaluation. We study the empirical tradeoff between predictive and diagnostic abilities, in applying various methods to estimate this fundamental relationship. We incorporate methods previously employed in the marketing literature, and methods that are so far less common in the marketing literature. For generalizability, we analyze 25,241 products in nine product categories, and 260,489 reviews across five review platforms. We find that neural network-based machine learning methods, in particular pre-trained versions, offer the most accurate predictions, while topic models such as Latent Dirichlet Allocation offer deeper diagnostics. However, neural network models are not suited for diagnostic purposes and topic models are ill equipped for making predictions. Consequently, future selection of methods to process text reviews is likely to be based on analysts’ goals of prediction versus diagnostics.     

Extracting brand information from social networks: Integrating image,text, and social tagging data

Images are an essential feature of many social networking services, such as Facebook, Instagram, and Twitter. Through brand-related images, consumers communicate about brands with each other and link the brand with rich contextual and consumption experiences. However, previous articles in marketing research have concentrated on deriving brand information from textual user-generated content and have largely not considered brand-related images. The analysis of brand-related images yields at least two challenges. First, the content displayed in images is heterogeneous, and second, images rarely show what users think and feel in or about the situations displayed. To meet these challenges, this article presents a two-step approach that involves collecting, labeling, clustering, aggregating, mapping, and analyzing brand-related user-generated content. The collected data are brand-related images, caption texts, and social tags posted on Instagram. Clustering images labeled via Google Cloud Vision API enabled to identify heterogeneous contents (e.g. products) and contexts (e.g. situations) that consumers create content about. Aggregating and mapping the textual information for the resulting image clusters in the form of associative networks empowers marketers to derive meaningful insights by inferring what consumers think and feel about their brand regarding different contents and contexts.     

Comparing consumer-produced product reviews across multiple websites with sentiment classification

ABSTRACT

Online consumer reviews have been extensively studied. However, existing literature analyzing online consumer review data mostly relies on a single data source, resulting in potentially biased analytics conclusions. Many websites encourage consumers to post reviews of their purchased products, so that new consumers can evaluate these reviews for the same product across different websites to help them make purchasing decisions. Confusions often arise in this process, because there often exist substantial discrepancies in customer reviews across different retailers on the same product. Clarifying such confusions can help consumers reduce concerns to make up their mind for their purchases, therefore benefiting both consumers and retailers. Through text analytics and sentiment analysis, we comparatively examine the underlying patterns of online consumer reviews of three large retailers including Sears, Home Depot, and Best Buy for a same product. Afterward, we combine online consumer reviews from these large retailers and conduct an overall text analytics and sentiment analysis. The overall results are further compared with the results from individual retailers. The findings show that the sentiment of the online consumer reviews could vary substantially so relying on a single data source to make purchase decision is not a wise idea. Based on the results, we further devise a framework to comparatively examine and integrate multiple data sources for social media analytics of online consumer reviews. This study offers important managerial implications and identifies several new research directions for social media analytics.     

Making use of quantitative content analysis: Insights from academia and business practice

This article features a moderated conversation between the author and five leading researchers and practitioners in the field of quantitative content analysis and, more specifically, computer-aided textual analysis (CATA). The interview subjects discuss the factors contributing to CATA’s rise as a research method as well as the benefits and pitfalls of using CATA. The conversation also covers best-practice advice that companies may employ to increase the accuracy of CATA results. These include using different programs or algorithms for the same problem, known as algorithm error.     

Social media analytics for enterprises: Typology,methods, and processes

This article provides an overview of social media analytics for managers that seek to utilize the practice for social media intelligence. Currently, managers are challenged to analyze an abundance of social media data but lack a framework within which to do so. Toward this end, this article presents a simple typology of social media analytics for enterprises. It also discusses various analytics methods for social media data. Then, this article discusses management processes of social media analytics for enterprises. An illustration of social media analytics is provided with real-world consumer review data. Finally, four challenges are discussed.     

Seeing the wood for the trees: How machine learning can help firms in identifying relevant electronic word-of-mouth in social media

The increasing volume of firm-related conversations on social media has made it considerably more difficult for marketers to track and analyse electronic word-of-mouth (eWOM) about brands, products or services. Firms often use sentiment analysis to identify relevant eWOM that requires a response to consequently engage in webcare. In this paper, we show that sentiment analysis of any kind might not be ideal for this purpose, because it relies on the questionable assumption that only negative eWOM is response-worthy and it is not able to infer meaning from text. We propose and test an approach based on supervised machine learning that first decides whether eWOM is relevant for the brand to respond, and then—based on a categorization of seven different types of eWOM (e.g., question, complaint)—classifies three customer satisfaction dimensions. Using a dataset of approximately 60,000 Facebook comments and 11,000 tweets about 16 different brands in eight different industries, we test and compare the efficacy of various sentiment analysis, dictionary-based and machine learning techniques to detect relevant eWOM. In doing so, this study identifies response-worthy eWOM based on the content instead of its expressed sentiment. The results indicate that these machine learning techniques achieve considerably higher accuracy in detecting relevant eWOM on social media compared to any kind of sentiment analysis. Moreover, it is shown that industry-specific classifiers can further improve this process and that algorithms are applicable across different social networks.     

Automated inference of product attributes and their importance from user-generated content: Can we replace traditional market research?

User-generated content, particularly online product reviews by customers, provide marketers with rich data of customer evaluations of product attributes. This study proposes, benchmarks, and validates a new approach for inferring attribute-level evaluations from user-generated content. Moreover, little is known about whether and when insights from product reviews gained in such a way are consistent with traditional research methods, such as conjoint analysis and satisfaction driver analysis. To provide first insights into this question, the authors apply their approach to a dataset with almost one million product reviews from 52 product categories and run conjoint and satisfaction driver analyses for these categories. Results indicate that the consistency between methods largely varies across product categories. Initial exploratory analyses suggest that consistency might be higher for categories characterized by low experience qualities, high hedonic value, and high customer willingness to post online reviews—but further work will be necessary to validate these findings.     

Two hearts in three-quarter time: How to waltz the social media/viral marketing dance

The concept of viral marketing has been discussed in the literature for over 15 years, since Jeffrey Rayport first introduced the term in 1996. However, the more widespread use of social media has recently pushed this idea to a whole new level. We provide insight into the relationship between social media and viral marketing, and illustrate the six steps executives should take in order to dance the social media/viral marketing waltz. We define viral marketing as electronic word-of-mouth whereby some form of marketing message related to a company, brand, or product is transmitted in an exponentially growing way—often through the use of social media applications. We consider the three conditions that need to be fulfilled to create a viral marketing epidemic (i.e., giving the right message to the right messengers in the right environment) and present four different groups of social media viral marketing campaigns (nightmares, strokes-of-luck, homemade issues, and triumphs). We conclude with five points of caution that managers should heed when trying to launch their own viral marketing campaign.     

Social media? Get serious! Understanding the functional building blocks of social media

Traditionally, consumers used the Internet to simply expend content: they read it, they watched it, and they used it to buy products and services. Increasingly, however, consumers are utilizing platforms—such as content sharing sites, blogs, social networking, and wikis—to create, modify, share, and discuss Internet content. This represents the social media phenomenon, which can now significantly impact a firm's reputation, sales, and even survival. Yet, many executives eschew or ignore this form of media because they don’t understand what it is, the various forms it can take, and how to engage with it and learn. In response, we present a framework that defines social media by using seven functional building blocks: identity, conversations, sharing, presence, relationships, reputation, and groups. As different social media activities are defined by the extent to which they focus on some or all of these blocks, we explain the implications that each block can have for how firms should engage with social media. To conclude, we present a number of recommendations regarding how firms should develop strategies for monitoring, understanding, and responding to different social media activities.     

Business performance and social media: Love or hate?

The social media space has become a common place for communication, networking, and content sharing. Many companies seek marketing and business opportunities via these platforms. However, the link between resources generated from these sites and business performance remains largely unexploited. Both managers and financial advisors can profit from the lessons learned in this study. We conceptualize four channels by which social media impacts financial, operational, and corporate social performance: social capital, customers’ revealed preferences, social marketing, and social corporate networking. An empirical test of our framework shows that ‘followers’ and ‘likes’ positively influence a firm's share value, but only after a critical mass of followers is attained. Our estimates suggest that Twitter is a more powerful tool to enhance business performance than Facebook.     

The early bird catches the news: Nine things you should know about micro-blogging

Micro-blogs (e.g., Twitter, Jaiku, Plurk, Tumblr) are starting to become an established category within the general group of social media. Yet, while they rapidly gain interest among consumers and companies alike, there is no evidence to explain why anybody should be interested in an application that is limited to the exchange of short, 140-character text messages. To this end, our article intends to provide some insight. First, we demonstrate that the success of micro-blogs is due to the specific set of characteristics they possess: the creation of ambient awareness; a unique form of push-push-pull communication; and the ability to serve as a platform for virtual exhibitionism and voyeurism. We then discuss how applications such as Twitter can generate value for companies along all three stages of the marketing process: pre-purchase (i.e., marketing research); purchase (i.e., marketing communications); and post-purchase (i.e., customer services). Finally, we present a set of rules—The Three Rs of Micro-Blogging: Relevance; Respect; Return—which companies should consider when relying on this type of application.     

Collaborative intelligence: How human and artificial intelligence create value along the B2B sales funnel

The B2B sales process is undergoing substantial transformations fueled by advances in information and communications technology, specifically in artificial intelligence (AI). The premise of AI is to turn vast amounts of data into information for superior knowledge creation and knowledge management in B2B sales. In doing so, AI can significantly alter the traditional human-centric sales process. In this article, we describe how AI affects the B2B sales funnel. For each stage of the funnel, we describe key sales tasks, explain the specific contributions AI can bring, and clarify the role humans play. We also outline managerial considerations to maximize the contributions from AI and people in the context of B2B sales.     

雷达辐射源分类识别方法综述

阐述了雷达辐射源识别研究的必要性,详细介绍了现有雷达辐射源基本分类识别方法,系统梳理了现有分类识别技术的优缺点,合理分析了多、大且杂的电磁环境下雷达辐射源识别研究趋势,提出将深度学习与雷达辐射源识别全面结合,实现雷达辐射源识别智能化。     

Augmented reality: Designing immersive experiences that maximize consumer engagement

Innovative marketers can now leverage augmented reality to craft immersive brand experiences, create more interactive advertising, and enable consumers to experience products and spaces in novel ways. Augmented reality (AR) is the practice of displaying digital information over people's real-time view of objects, people, or spaces in the physical world. While AR can play a valuable role in integrated marketing programs, little is known about the practice and how to execute effective AR programs in the marketplace. We address this gap by presenting a framework that describes the active and passive ingredients of augmented reality. We then describe the basic design decisions that marketers need to make when planning an augmented reality campaign. In addition, we explain how understanding and addressing the dynamics between various active and passive AR ingredients can help marketers to optimize their AR campaigns and enhance various types of consumer engagement: user-brand engagement, user-user engagement, and user-bystander engagement. Through our framework and analysis, we develop eight actionable recommendations—described with the acronym ENTANGLE—marketing managers can use to design immersive AR experiences that maximize consumer engagement.     

Big data on a smaller scale: A social media analytics assignment

It is truly important for students to understand how to monitor online marketing buzz. This assignment, social media analytics, utilizes the content analysis research method to build student's in-depth understanding on how to evaluate and interpret user-generated content (UGC) to create social media campaigns. The authors adapted Resnik and Stern's (1977) coding scheme for UGC. Through experiential learning, students immerse themselves in data and analyze UGC. The assignment scored high in knowledge acquisition as a pedagogical tool. Finally, the authors provide an updated social media analytics coding scheme, guidelines for instructors, student rubric information, and student learning outcomes.