Research on The Impact of Visual Elements on Sales in E-commerce Context: Empirical Evidence of Pictures and Virtual Reality


Student thesis: Doctoral Thesis

View graph of relations


Related Research Unit(s)


Awarding Institution
Award date7 Sep 2021


The rapid development of Internet and digital technology has spawned many e-commerce platforms. The digital economy plays an increasingly important role in GDP. For instance, China’s online retail volume has exceeded 100 billion CNY in 2019.

On e-commerce platforms, online presentation is the critical channel for merchants to demonstrate their attributes. Based on the information format, online materials can be generally classified into two types: 1) text elements and 2) visual elements. Text elements show literal information, such as the price, discount rate, consumer reviews, and merchant responses. Prior literature has discussed much about the impact of text elements on users’ behaviors and merchants’ sales. Besides, visual elements display graphical information, such as images, videos, and virtual reality. Even though a few works have investigated the effects of visual elements on market performance, most of them focus on pictures’ presence or not. While visual elements belong to unstructured data, which can transfer diverse semantic meanings within a single piece. The cone of experience theory reveals that people can only remember 10% of the textual information, 20% of the audio information, but 30% of graphical information. Thus, visual information is a more informative format than text and audio, which can affect humans’ perception and understanding of the external environment. However, the impacts of visual elements on online platforms are still under investigation.

This thesis conducts empirical analyses on two kinds of visual elements, “Picture” (a commonly used visual format) and “Virtual Reality” (an emerging visual format). I concentrate on e-commerce platforms as the research environment, collect large-scale empirical data, employ advanced econometric tools, and examine their effects on product sales.

To be specific, each merchant displays a representative photo on consumers’ search results pages, denoting the merchant’s cover picture. Merchants’ cover pictures are essential marketer-generated visual content and may affect consumers’ behaviors and final purchase decisions. Although prior studies have discussed the impact of pictures in the e-commerce context, most of them concentrate on a single type of photos.

However, the information depicted by a photo is flexible. Different merchants may display different categories of pictures, either brand, product, or environment pictures. Hence, the first project aims to understand: Can different cover pictures generate distinct sales? If yes, which aspects can explain the sales variation? Particularly, I focus on the cover picture’s content and composition. Picture content denotes the focal objects and elements depicted, reflecting the photo’s informational effect. Picture composition denotes the arrangement of objects and elements, which influences the photo’s visually compelling, reflecting the aesthetic effect.

Results of the first study contributes to the literature of marketer-generated visual content and by revealing the brand category of cover pictures as a dominant strategy, which can generate better sales than the product and environment categories. This study provides theoretical insights to visual marketing in the online marketplace by explaining the sales distinction from aesthetics and information channels. On one side, the brand category has an intrinsic advantage in composition, where brand pictures’ aesthetic effect can partly explain the sales variation. On the other side, the magnitude of brand awareness can also affect merchants’ sales performance, demonstrating cover pictures’ informational effect. Besides, the impact of picture content and composition are more influential for new merchants.

Moreover, VR can generate a 3D virtual environment for users to explore rich visual information and interact with the interface in real-time. The second project focuses on a recent application of non-immersive VR on real estate platforms, which demonstrates property details through a screen. Prior works mainly employ lab experiments in investigating the influence of immersive VR on consumers’ behavior and show conflicting results. However, immersive VR usually needs the support of expensive accompanied devices (e.g., head-mounted sets), thus only employing in limited areas, such as medicine, electricity, and aerospace. Few in-depth empirical works discuss the influence of non-immersive VR applications on market performance. In 2018, Beike, a leading online real estate platform in China, introduces VR as a novel online house tour function. This is the very first large-scale application of non-immersive VR in the e-commerce context. Besides, properties are typical high involvement products with high price, importance, and purchase risk. The effects of non-immersive VR on the selling process of high involvement products are still unclear, calling for further investigations.

Therefore, the second project examines the impact of VR on properties’ selling process. Specifically, I collect a dataset of property transaction records, employ a set of econometric tools to examine the impact of VR on properties’ selling time, demonstrating the effect of VR on consumer’s decision-making process, showing the impact of consumer’s prior using experience on VR, and uncovering the relationship between VR and pictures on properties’ selling process.

Results of the second study contribute to the literature of VR, product involvement, and e-commerce by demonstrating the effectiveness of VR on selling high involvement products. Overall, VR can accelerate properties’ selling process. This study also generates theoretical insights to the technology adoption, offering a unified viewpoint — consumers’ prior using experience to explain conflicting effects in prior literature. Specifically, VR may extend the selling time in a short period when consumers have no prior experience. Over time, consumers get familiar with it. The acceleration effect of VR will occur and gradually increase. Later, the acceleration effect will become stable and sustain in the long run. This study reveals underlying mechanism of VR in consumers’ decision-making process. In the information search stage, the presence of a “VR badge” increases the number of web visits to a property’s detail page. In the alternative evaluation stage, VR may increase the number of attentions and consumers’ conversion rate, demonstrating the effectiveness of VR in delivering property information and discerning consumers’ product quality uncertainty. Moreover, the study enhances the understanding of the relationship between visual materials. VR presents a substitution effect over pictures. For a property supporting both VR and pictures, the seller does not need to show many pictures.

This thesis extends the literature of marketer-generated visual content, web design, e-commerce, VR, and high involvement products. The results highlight the brand category of cover pictures as the dominant visual presentation strategy and demonstrate the effectiveness of VR in selling high involvement products in the e-commerce context. The thesis advances visual marketing and brand management theories. The findings reveal that the sales variation between the brand category and other alternatives relates with the aesthetic effect and informational effect. Besides, the findings also generate theoretical insights for visual perception theory in the online marketplace and explain existing conflicting results of VR with the technology adoption theory. Moreover, this thesis can provide managerial implications for platform owners and merchants in making use of cover pictures and VR for online product presentations.

    Research areas

  • e-commerce, visual elements, sales, pictures, content, composition, virtual reality, consumers’ decision-making process