Factors Influencing Sales Performance in Live Streaming E-commerce: An Analysis

Configuration Analysis of Factors Influencing the Configuration Analysis of Factors Influencing the Sales Performance of Live Sales Performance of Live Streaming E based on the fsQCA Method. based on the fsQCA Method. Streaming E- -commerce commerce Chen, J.H., Duan, L.Z., Feng, Y.Y., & Yang,C.C.

01 Background and Overview of the Research

Background and Overview of the Research 01 E-commerce live streaming Launched on Kuaishou and Douyin since 2016 Well-known as Li Jiaqi and Oriental Selection etc. Rapid development during COVID-19 transmission Estimated 5 trillion yuan economy by 2023

Background and Overview of the Research 01 Impulsive consumption Pressure from other consumers Psychological contracts, etc. Makes it difficult for companies and live streaming practitioners test various influencing factors.

Background and Overview of the Research 01 Questionnaire surveys VS Objective data multi-causal perspective based on configurational analysis

Background and Overview of the Research 01 "Oriental Selection" anchor Elaboration Likelihood Model (ELM) theory seven research variables. fsQCA method

Background and Overview of the Research 01 fsQCA ELM theory Fuzzy-Set Qualitative Comparative Analysis Boolean algebra Multiple conditions Internet consumer information processing and decision-making processes. the central route and peripheral route motivation and ability to process information. Variable calibration Necessary Condition Analysis In the context of live streaming e-commerce, Configurational Analysis related to the product the central route perception of the anchor peripheral route

Background and Overview of the Research Central route 01 Peripheral route Duration of the Product Live Broadcast Type of Product the duration of live streaming explanation experiential products search products Number of Likes Price of Product the total number of likes received selling price (during the live streaming explanation ) Number of Gifts the number of gifts sent Number of Comments about the Product the number of comments on the product (during the live streaming explanation ) (during the live streaming explanation ) Number of Comments about the Anchor the total number of comments on the anchor (during the live streaming explanation )

02 Research Process

Research process-Data collection 02 Approximately 50 hours of live streaming data During each product demonstration was used as a sample. Web software collection: comment data, like data, and gift data. Manual recording collection: product type, product price, Duration of the Product Live Broadcast, and product sales revenue. The Bert algorithm comments on the productcomments on the anchor. total 500 samples of product live streaming data removed the top 50 and bottom 50 samples

Research process-Comment data classification(BERT algorithm) 02 80% training set and 20% test set 2000 comments on the product and 2000 comments on the anchor were manually labeled ACC ACC=0.955 =0.955 F F1=0.951 1=0.951

Research process-Variable calibration 02 the 95th percentile value, 50th percentile value, and 5th percentile value (Ragin C. C,2009) as the completely membership point, cross point, and completely non-membership point

Research process-Necessary condition analysis 02 whether the result depends on a single variable consistency and coverage Consistency> 0.9 the conditional variable is a necessary condition.

Research process-Configurational analysis 02 The core of the QCA method Explore the impact of different combinations of antecedent conditions on the result. The consistency threshold : 0.8, and the frequency threshold : 2 The PRI threshold for exploring high-performance configurations: 0.75 The PRI threshold for exploring low-performance configurations 0.5. Simple solution and intermediate solution

Research process-Configurational analysis 02

03 Result Analysis

Result Analysis - Necessary condition cnalysis results 03 No variable can be a necessary condition for the result variable. >0.8 these two indicators have a relatively high impact on increasing sales of the product. Proposition 1: The live-streaming e-commerce scene is more conducive to the growth of sales of experience products.

Result Analysis - High sales configuration analysis results 03 Proposition 2: In the live streaming e-commerce context, for experience products to achieve high sales, the central antecedent condition of the number of comments on the product and the peripheral antecedent condition of the number of likes must both be present. On this basis, the peripheral antecedent conditions of the number of comments from consumers about the anchor, number of gifts, and duration of the live broadcast can be mutually substituted.

Result Analysis - High sales configuration analysis results 03 Proposition 3: When purchasing search products in live streaming e-commerce, consumers tend to buy low-priced products, willingness to interact with the anchor will decrease. and their Proposition 4: When products in the live-streaming e-commerce context, they need longer product marketing time from the anchor and more comments from about the anchor, while low-priced products require more opinions from other consumers about the product. consumers purchase high-priced

Result Analysis - Low sales configuration analysis results 03 A high comments number about low- priced experience products numbers The duration of the live broadcast for high-priced experience extended products is A high comments number about the anchor for experience products numbers

04 Conclusion and Discussion

Conclusion 04 Product selection: the main focus should be on experience products(food ect.) for search products, attract consumers through promotional low prices Live streaming marketing process: spend more time on high-priced products for experience products and high-priced products: interaction for experience products: conversation skills with consumers for search products and low-priced products, reduce marketing time Comment guidance: explaining search products, suppress the discussion of product attributes explaining high-priced products, lead discussions around the anchor

Discussion 04 Limitations and Future Research Prospects The study may not fully explain the effects of popularity on sales. Future: distinguish more live streaming categories to explore this issue The study did not clarify the relationship between gift-giving and purchasing behavior. Future: it is necessary to understand the psychological conditions The study did not distinguish between positive and negative attention. Future: refine bullet screen classification to explore how bullet screens affects their willingness to purchase more comprehensively

THANKS THANKS THANK YOU FOR WATCHING