라이브커머스 성과 결정요인 분석 및 예측: 실제 방송 데이터를 중심으로

초록

Purpose The study explores factors influencing performance in the food category of live commerce, a rapidly growing channel emphasizing real-time interaction and unique shopping experiences. This study contributes to the literature by quantitatively analyzing live commerce success factors using actual transaction data. It offers practical insights for companies to predict outcomes and strategize effectively before entering the live commerce space. Design/methodology/approach Using actual broadcast data collected from various live commerce platforms, the study examines key factors affecting performance metrics such as sales, conversion rates, and views. A machine learning model was applied to predict outcomes like revenue and viewership with high accuracy, enabling deeper insights into the dynamics of live commerce. Findings Results indicate that a higher number of products positively influences overall performance, while pre-broadcast views do not significantly affect conversion rates. High-priced products drive revenue growth but tend to decrease sales volume and conversion rates. Analysis of broadcast timing reveals that weekend and late-night sessions attract higher views, whereas weekday morning and afternoon broadcasts yield better conversion rates. The machine learning model demonstrated predictive accuracy with an error margin of 8–9% for sales and views.

키워드