This work aims to explore the characteristics of marketer-generated content (MGC) affecting customer engagement on social media (SM). Data included 1,054 tweets posted by Lazada Thailand in 6-month period. Contextual, informational, and brand characteristics as well as vivid and timing characteristics were investigated. Content analysis for informational characteristics (the service marketing mix: 7Ps) and luxury attributes was conducted by ChatGPT-3.5. Descriptive statistics, regression analyses, and word frequency analysis were used to describe data and evaluate hypotheses. Results supported the influence of contextual (mentions, tweet length), vivid (media type), informational (people and process), brand (brand mentioned), and timing (weekends and working time) in both positive and negative ways. Top luxury attribute keywords in the e-marketplace’s tweets are presented. The findings offer guidelines for researchers, including the 7Ps to study e-marketplaces and for marketers in employing proper MGC characteristics to effectively communicate to customers, and reveal luxury attributes in tweets.