Recent developments in machine learning (ML), especially transformer-based discriminative and generative deep learning, transform the marketing landscape. So, e.g., marketers predict with high accuracy sentiment scores from online customer review (OCR) comments in natural language and gain valuable insights whether, when, and how apps, products, or services should be improved. However, oftentimes, OCR comments contain additional interesting information that goes beyond sentiment indications. In this work, we propose a new approach to predict – based on the well-known Technology Acceptance Model (TAM) – extended TAM construct scores from OCRs and compare the accuracy of this prediction with various ML models for this purpose. The comparison is based on a dataset with n = 5,356 OCR comments for the Ikea app, labeled by three human experts (n = 3), and 18 ML models. Following this we conduct a case study on the Ikea dataset and show how to use these TAM construct scores in conjunction with topic modeling to identify various usability issues of the Ikea app. Additionally, we propose an approach that leverages TAM constructs to identify OCRs with complex and rich content that would not be identifiable with sentiment alone.

Renting fashion using clothing box subscription is a growing trend in the textile industry. The element of surprise varies according to the box type chosen by the customer: the self-assembled or the curated surprise box. Our study focuses on the effects of consumer characteristics, box type and other attributes on the intention to subscribe. We collected data from 364 German respondents and used choice-based conjoint analysis to estimate these effects. A between-subject design helps to compare the self-assembled versus the curated surprise box type. Price has the highest relative importance in the curated surprise box subscription model. In both subscription models, consumers preferred to rent four fashion pieces rather than two. Sustainable fashion labels increase the willingness to pay in the case of the self-assembled box model. Most consumers still prefer new fashion items over second-hand or upcycled ones, leaving considerable room for circular communication strategies.

User-generated content (UGC) is generally understood as an expression of opinion in many forms (e.g., complaints, online customer reviews, posts, testimonials) and data types (e.g., text, image, audio, video, or a combination thereof) that has been created and made available by users of websites, platforms, and apps on the Internet. In the digital age, huge amounts of UGC are available. Since UGC often reflects evaluations of brands, products, services, and technologies, many consumers rely on UGC to support and secure their purchasing and/or usage decisions. But UGC also has significant value for marketing managers. UGC allows them to easily gain insights into consumer attitudes, preferences, and behaviors. In this article, we review the literature on UGC-based decision support from this managerial perspective and look closely at relevant methods. In particular, we discuss how to collect and analyze various types of UGC from websites, platforms, and apps. Traditional data analysis and machine learning based on feature extraction methods as well as discriminative and generative deep learning methods are discussed. Selected use cases across various marketing management decision areas (such as customer/market selection, brand management, product/service quality management, new product/service development) are summarized. We provide researchers and practitioners with a comprehensive understanding of the current state of UGC data collection and analysis and help them to leverage this powerful resource effectively. Moreover, we shed light on potential applications in managerial decision support and identify research questions for further exploration.

Online customer reviews (OCRs) are user-generated, semi-formal evaluations of products, services, or technologies. They usually consist of a timestamp, a star rating, and, in many cases, a comment that reflects perceived strengths and weaknesses. OCRs are easily accessible in large numbers on the Internet – for example, through app stores, electronic marketplaces, online shops, and review websites. This paper presents new transfer models to predict technology acceptance and its determinants from OCRs. We train, test, and validate these prediction models using large OCR samples and corresponding observed construct ratings by human experts and generative artificial intelligence chatbots as well as estimated ratings from a traditional customer survey. From a management perspective, the new approach enhances former technology acceptance measurement since we use OCRs as a basis for prediction and discuss the evolution of acceptance over time.

Selecting or adjusting attribute-levels (e.g. components, equipments, flavors, ingredients, prices, tastes) for multiple new and/or status quo products is an important task for a focal firm in a dynamic market. Usually, the goal is to maximize expected overall buyers’ welfare based on consumers’ partworths or expected revenue, market share, and profit under given assumptions. However, in general, these so-called product-line design problems cannot be solved exactly in acceptable computing time. Therefore, heuristics have been proposed: Two-stage heuristics select promising candidates for single products and evaluate sets of them as product-lines. One-stage heuristics directly search for multiple attribute-level combinations. In this paper, Ant Colony Optimization, Genetic Algorithms, Particle Swarm Optimization, Simulated Annealing and, firstly, Cluster-based Genetic Algorithm and Max-Min Ant Systems are applied to 78 small- to large-size product-line design problem instances. In contrast to former comparisons, data is generated according to a large sample of commercial conjoint analysis applications ( n = 2,089). The results are promising: The firstly applied heuristics outperform the established ones.

Understanding customer needs is key for fashion retailers to stay competitive and innovative. Surprisingly, however, extant literature mainly explores customer needs in terms of a garment and its attributes rather than viewing shopping as a problem-solving process to meet customer needs. Moreover, these studies fail to address how customers meet their needs in-store (ISFR) and online fashion retailing (OFR). To fill this research gap, we empirically investigate customers' personal and social needs and how they can be met through the jobs-to-be-done theory. Findings reveal that, beyond the purchase of a garment, customer needs can be fulfilled through different ways, such as smart technology or a person's high interaction with social others in ISFR and the online shop experience or a social linkage without social interaction in OFR. Additionally, our findings offer potential service innovations for fashion retailing managers.

This paper updates a 1988 review of marketing data analysis by dual scaling. Since then, the number of applications has grown considerably. However, the spread is still low compared to, for example, conjoint analysis. On the other side, recent progress in data collection, methodology, and related dual scaling software packages creates new opportunities. The ability to analyse complex and varied data (answers to open questions, associations, cross-tabulations, discrete choices, preferences, ratings) could be a decisive advantage and is demonstrated by a new large-scale marketing application. A sample of online shop customers ( $$n = 4411$$ ) was asked to rank-order sustainable improvement options. Dual scaling helps managers and deciders to focus on preferred improvements.

Selecting adequate attribute-levels (e.g., components, ingredients, materials, prices, qualities) for new and/or existing products is an important task for marketeers. The goal is to maximize a focal firm’s overall revenue or profit. The typical knowledge base consists of customers’ attribute-level partworths, marginal contributions, and descriptions of own and competing status quo products. However, since these so-called product-line design problems are known to be NP-hard, they often cannot be solved exactly. Instead, heuristics have to be applied. In this paper, we give an overview on proposed solution methods. Moreover, we apply two recent propositions—Cluster-Based Genetic Algorithms (CGA) and Tabu Search (TS)—to a sample of 460 small (up to about 10 $$^6$$ possible solutions)-to-large-size problems (more than 10 $$^{12}$$ possible solutions). The results are promising: Especially CGA solves small-size problems accurately and in acceptable computing time (within seconds), the latter even when applied to medium- and large-size problems.

This paper investigates the success of sustainable product innovation in the textile, clothing, and leather industry. Basing on the typology of Medeiros et al. (2014) a literature review was conducted to update the critical success factors. To investigate the importance of the factors, this research made use of a mixed methodology. 1634 digital news magazine articles and corporate communications were collected in a database describing the development and introduction of 176 new products from 158 companies using low-waste, circular, or "fair" production processes and new materials such as organic, recycled, bio-based synthetic fabrics. From the product and innovation managers responsible, a sample of 33 respondents participated in a survey, where their products had to be evaluated using operationalizations of three success aspects and fourteen success factors. Additionally, five sustainable textile product innovation experts characterized all 176 products in a similar manner, based on the information in the database. Relying on an indirect benchmarking approach both samples produce similar conclusions. The most important success factors are customer expectation fulfillment (in first place) followed by compliance with laws and regulations, green creativity, knowledge about factors that drive sustainable buying, investments in R&D infrastructure, and competitor monitoring. While the importance of market acceptance and the understanding of customers' perspectives is highlighted, the research makes a contribution to the literature on success factors of sustainable textile innovations.

Purpose To examine whether the country of origin (COO) effect actually exists in an e-commerce context, the authors intend to contribute to the ongoing debate by measuring the COO effect through a series of connected studies. Design/methodology/approach Drawing on cue utilization theory, the authors emphasize the urge to investigate the COO effect in multiple cue settings in order to reveal a more realistic picture of its actual effect size. In contrast to most prior research, which often does not analyze COO using methodological plurality and neglects important contextual factors, the authors employed a four-staged research design in an attempt to trigger and measure the COO’s implicit effect size in today’s pervasive context of online shopping. The importance of brands (inhering the COO) is decompositionally calculated relative to other extrinsic cues by applying a Hierarchical Bayes estimation, with the COO impact being extracted subsequently. Findings The results deepen concerns that the COO effect actually does not exist, particularly in the more contemporary context of online shopping. Specifically, preferences for previously favored German products faded when controlling for brand attitude for both high-involvement ( p = 0.003) and low-involvement products ( p = 0.024). Research limitations/implications The study focused on consumers of Generation Y, as they represent one of the most important segments in online shopping. Findings might be replicated for other consumer generations. The study focused on Chinese consumers, as the Chinese e-commerce market represents the world’s largest one. Future studies might investigate other markets. Practical implications As brands, rather than a COO effect, impacted consumer preferences, companies selling their products to Chinese consumers online need to establish a reputation for quality early on. Chinese companies should emphasize their COO to make use of the ethnocentrism detected. Companies profit from the Best-Worst Scaling investigation revealing which product categories Chinese consumers most preferably buy online from German companies. Originality/value To the best of the authors’ knowledge, this study is the first to capture the importance of COO in the contemporary context of ubiquitous online shopping. Moreover, a more realistic and less biased way of measuring the importance of COO is enabled by building upon three pre-connected studies. The findings allow to develop a generalization for both high- and low-involvement products.

The COVID-19 pandemic brought about an increase in online shopping because of government-imposed restrictions and consumer anxiety over the potential health risk associated with in-store shopping. By end of 2021, many health concerns had been alleviated through efforts such as vaccinations and reductions in hospitalizations in certain countries. Some governments started to relax their restrictions and consumers started to return to in-store shopping, creating the possibility that the volume of online shopping would decrease once stores reopened. However, consumers may continue to shop online more than they did prior to the pandemic because of their experience during the lockdown. This study seeks to understand the factors that explain the potential of online shopping continuance. A novel model is constructed by extending ES-QUAL, and adding hedonic motivation, social shopping and health susceptibility as mediators. Empirical data is collected from Canada, Germany and the US. We find that convenience and efficiency, as well as security for some females, are important factors contributing to online shopping's perceived usefulness and, ultimately, intentions to continue shopping online. In addition, creating an enjoyable online shopping experience adds to these continuance intentions.

Modern call centers require precise forecasts of call and e-mail arrivals to optimize staffing decisions and to ensure high customer satisfaction through short waiting times and the availability of qualified agents. In the dynamic environment of multi-channel customer contact, organizational decision-makers often rely on robust but simplistic forecasting methods. Although forecasting literature indicates that incorporating additional information into time series predictions adds value by improving model performance, extant research in the call center domain barely considers the potential of sophisticated multivariate models. Hence, with an extended dynamic harmonic regression (DHR) approach, this study proposes a new reliable method for call center arrivals’ forecasting that is able to capture the dynamics of a time series and to include contextual information in form of predictor variables. The study evaluates the predictive potential of the approach on the call and e-mail arrival series of a leading German online retailer comprising 174 weeks of data. The analysis involves time series cross-validation with an expanding rolling window over 52 weeks and comprises established time series as well as machine learning models as benchmarks. The multivariate DHR model outperforms the compared models with regard to forecast accuracy for a broad spectrum of lead times. This study further gives contextual insights into the selection and optimal implementation of marketing-relevant predictor variables such as catalog releases, mail as well as postal reminders, or billing cycles.

Although there is a shift in consumers' consumption behavior towards more sustainable patterns across a variety of different contexts, sustainable apparel has still not become a mainstream trend despite the textile industry's excessive usage of valuable resources. Albeit extant research found different potential barriers elucidating why consumers hesitate to purchase such apparel, it remains unclear whether sustainability really matters to consumers in a clothing context and further, which aspects are of relevance during consumers' purchase decision. We thus conducted two studies with four best-worst scaling experiments in which 4,350 online shoppers assessed the importance of both conventional and sustainable apparel attributes, as well as sustainable apparel attributes only, and the willingness to pay for sustainable product attributes. We further inquired the importance of conventional as well as sustainable online shop attributes. Our findings indicate that conventional apparel attributes such as fit and comfort, price-performance ratio, and quality are of higher relevance to consumers than sustainable attributes. The most important sustainable apparel attributes are the garment's durability, fair wages and working conditions, as well as an environmentally friendly production process. Consumers also indicated to prefer the latter three attributes to a 20% discount. Moreover, consumers demand less as well as sustainable packaging, free returns, and discount campaigns. Our findings reveal a gender gap regarding green consumerism with female respondents assessing most sustainable attributes as more important than male respondents do.

In the present study, we investigate, from a consumer perspective, the importance of different types of marketing actions frequently used by online apparel retailers to serve different beneficiaries during the COVID-19 pandemic. The results of a Maximum Difference Scaling experiment among German consumers recruited from an online panel (n = 503) reveal that marketing actions with other-benefit components, such as corporate social responsibility initiatives, have the potential to outperform traditional sales promotion methods, such as price discounts. By deploying latent class analysis, two consumer segments can be distinguished according to their preferences: those valuing marketing actions with other-benefits and those preferring marketing campaigns with mere self-benefits. Finally, combinations of marketing actions with maximum reach are identified, from which recommendations for action by retailers are offered.