Impact of Livestream Shopping on Online Shopping Behaviour

Livestream shopping has rapidly emerged as a dominant online retail channel, reshaping the e-commerce landscape and consumer purchasing behaviour. While this evolution has drawn increasing academic and industry attention, research examining the actual behavioural outcomes of marketing strategies—especially through big data and psychologically grounded experimentation—remains limited. This thesis addresses this gap by investigating how livestream marketing strategies influence consumer behaviour across different stages of the marketing funnel, from attention and interest to decision and action. It further explores the psychological mechanisms and motivational drivers that shape consumers' responses under time-sensitive and socially interactive contexts. To this end, three empirical studies were conducted, each aligned with a different stage of the marketing funnel and capturing distinct dimensions of consumer motivation and engagement. The first study, “Time’s Up! The Effects of Marketing Time Pressure on Sales and User Stickiness in Livestream Shopping,” examines how time-pressure strategies influence immediate consumer responses (i.e., the action stage). Drawing on over 1,800 hours of livestream video and more than 99,000 minute-level observations on sales and viewer retention, this study utilises machine learning techniques to identify scarcity-based persuasion strategies, including limited-time scarcity, limited-quantity scarcity, and special promotional conditions. The livestream video data were first transcribed into timestamped text, which was then merged with platform analytics capturing both transactional outcomes (purchase behaviour) and relational outcomes (viewer retention) at the minute level. Latent Dirichlet Allocation (LDA) modelling was applied to the panelised transcript data to detect periods of heightened time pressure. Subsequently, econometric modelling using a random effects specification was employed to examine the effects of time pressure, along with other influential factors such as seller attributes and product characteristics, on both sales and retention outcomes. Findings reveal that while time pressure significantly boosts sales conversion, it reduces user stickiness, indicating a trade-off between short-term gains and long-term customer relationship. The effectiveness of these tactics also varies based on seller characteristics, with mega and macro influencers outperforming micro sellers. This study highlights the instrumental role of urgency-based strategies in driving transactional outcomes at the lower funnel. Building on these results, the second study, “See Now, Buy Now: The Impact of Time Pressure on Purchase Acceleration in Livestream Shopping,” delves into the underlying psychological mechanisms that govern accelerated decision-making under time pressure. Anchored in Construal Level Theory (CLT), this experimental study explores how temporal distance perceptions shape consumers' transition through the evaluation and decision stages of the funnel. The research design employed an AI-generated video to simulate a livestream shopping session under two experimental conditions. The time pressure condition featured cues such as countdown timers and urgency-oriented language to induce a heightened sense of urgency, while the control condition included reassurances from the seller encouraging participants to take their time in making purchase decisions. Two product types were featured: a utilitarian item (a basic white T-shirt for everyday wear) and a hedonic item (a ticket to a favourite artist’s concert). Following the viewing, participants completed a questionnaire designed to assess the effects of time pressure on purchase acceleration, as well as the influence of potential moderating variables. Time pressure was found to reduce perceived temporal distance, heightening the salience of immediate rewards and prompting quicker decisions. This effect is further moderated by social dynamics: consumer perceptions of community building and social competition amplify purchase acceleration, as does susceptibility to FOMO (Fear of Missing Out). Additionally, motivational differences across product types reveal that utilitarian products benefit more from time pressure than hedonic ones. This study offers a nuanced understanding of how psychological distance and social cues interact to influence mid-funnel behaviours. Extending the inquiry to the top of the funnel, the third study, “Fur Real: The Power of Pet Influencers in Shaping Customer Engagement in Livestream Shopping,” investigates emerging marketing strategies that capture consumer attention and foster engagement. Grounded in Regulatory Focus Theory, the study explores how different types of pet influencers (dogs vs. cats) trigger distinct motivational orientations—promotion vs. prevention focus—that drive varied viewer engagement behaviours. Three studies were employed to examine pet influencers’ impact on viewer engagement behaviours in livestream shopping at multiple dimensions (i.e. virtual gifting, likes and comments) from both seller and viewer perspective. To analyse 14,567 minutes of unstructured livestream video data, both the videos and crawled viewer comments were transcribed into per-second text segments and combined with corresponding engagement behaviours. The first study presents descriptive insights into pet influencer livestream sessions. Using the GPT-4o large language model with chain-of-thought prompting, the results reveal that cat influencer streams are characterised by relaxed tones and neutral emotional expressions, often focusing on product information. In contrast, dog influencer sessions exhibit higher levels of positive emotions and pet attachment. The second and third studies apply econometric modelling to examine the effects of pet type (dog versus cat) and discussion content, including aspects, emotions, and sentiments, on viewer engagement behaviours. These analyses draw on 70,299 observations of seller speech and 761,495 observations of individual viewer comments, demonstrating how interactions in pet influencers’ livestream session impact on virtual gifting, likes and comments. Results show that dog influencers drive immediate action through virtual gifting and product-related discussions, while cat influencers foster passive, emotionally driven engagement through likes and comments. These findings demonstrate how symbolic and emotional cues at the awareness and interest stages of the funnel can shape consumer interaction patterns and engagement intensity. Collectively, these three studies form a coherent narrative that spans the full marketing funnel—from engagement to evaluation to purchase—while unpacking the roles of time pressure, psychological distance, social dynamics, and motivational triggers in livestream shopping. This thesis adopts a holistic and methodologically innovative approach, combining AI-enabled data analytics, experimental design, and real-world business data. It contributes to the advancement of consumer research by extending Construal Level Theory and Regulatory Focus Theory into livestream commerce contexts and offers actionable insights for practitioners on how to optimise urgency tactics, tailor promotional strategies to product type and consumer profiles, and leverage emerging influencer formats to drive conversions and engagement. In doing so, it equips brands with data-driven strategies to navigate the evolving livestream shopping as emerging e-commerce channel and remain competitive in digital marketplace.<p></p>